├── .travis.yml ├── LICENSE ├── Makefile ├── README.md ├── WebAssemby ├── README.md └── face_detection_browser_usage.html ├── samples ├── README.md ├── acgt.png ├── batch_img_loading.c ├── blob_detection.c ├── canny.jpg ├── canny_edge_detection.c ├── cnn_coco.c ├── cnn_face_detection.c ├── cnn_faces.jpg ├── cnn_object_detection.c ├── cnn_voc.c ├── crop_image.c ├── dilate_image.c ├── erode_image.c ├── fingerprint │ ├── fp001_0.pgm │ ├── fp001_1.pgm │ ├── fp002_0.pgm │ ├── fp002_1.pgm │ ├── fp003_0.pgm │ ├── fp003_1.pgm │ ├── fp004_0.pgm │ ├── fp004_1.pgm │ ├── fp005_0.pgm │ ├── fp005_1.pgm │ ├── fp006_0.pgm │ ├── fp006_1.pgm │ ├── fp007_0.pgm │ ├── fp007_1.pgm │ ├── fp008_0.pgm │ ├── fp008_1.pgm │ ├── fp_part.pgm │ └── fp_whole.pgm ├── flower.jpg ├── gaussian_blur.c ├── grayscale_image.c ├── hilditch_thin.c ├── hough_lines_detection.c ├── license_plate_detection.c ├── minutiae.c ├── otsu.jpg ├── otsu_image.c ├── out_cnn.png ├── plate.jpg ├── realnet_face_detection.c ├── realnet_face_detection_embedded.c ├── realnet_faces.jpg ├── realnet_train_model.c ├── resize_image.c ├── rnn_text_gen.c ├── rotate_image.c ├── sepia_filter.c ├── sobel.jpg ├── sobel_operator_img.c ├── test.png ├── text.jpg └── train.txt ├── sod.c ├── sod.h ├── sod_img_reader.h └── sod_img_writer.h /.travis.yml: -------------------------------------------------------------------------------- 1 | language: c 2 | compiler: clang 3 | script: make -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | /* 2 | * SOD - An Embedded Computer Vision & Machine Learning Library. 3 | * Copyright (C) 2018 - 2020 PixLab| Symisc Systems. https://sod.pixlab.io 4 | * Version 1.1.8 5 | * 6 | * Symisc Systems employs a dual licensing model that offers customers 7 | * a choice of either our open source license (GPLv3) or a commercial 8 | * license. 9 | * 10 | * For information on licensing, redistribution of the SOD library, and for a DISCLAIMER OF ALL WARRANTIES 11 | * please visit: 12 | * https://pixlab.io/sod 13 | * or contact: 14 | * licensing@symisc.net 15 | * support@pixlab.io 16 | */ 17 | --------------------------------------------------------------------------------------- 18 | GNU GENERAL PUBLIC LICENSE 19 | Version 3, 29 June 2007 20 | 21 | Copyright (C) 2007 Free Software Foundation, Inc. 22 | Everyone is permitted to copy and distribute verbatim copies 23 | of this license document, but changing it is not allowed. 24 | 25 | Preamble 26 | 27 | The GNU General Public License is a free, copyleft license for 28 | software and other kinds of works. 29 | 30 | The licenses for most software and other practical works are designed 31 | to take away your freedom to share and change the works. By contrast, 32 | the GNU General Public License is intended to guarantee your freedom to 33 | share and change all versions of a program--to make sure it remains free 34 | software for all its users. We, the Free Software Foundation, use the 35 | GNU General Public License for most of our software; it applies also to 36 | any other work released this way by its authors. You can apply it to 37 | your programs, too. 38 | 39 | When we speak of free software, we are referring to freedom, not 40 | price. Our General Public Licenses are designed to make sure that you 41 | have the freedom to distribute copies of free software (and charge for 42 | them if you wish), that you receive source code or can get it if you 43 | want it, that you can change the software or use pieces of it in new 44 | free programs, and that you know you can do these things. 45 | 46 | To protect your rights, we need to prevent others from denying you 47 | these rights or asking you to surrender the rights. Therefore, you have 48 | certain responsibilities if you distribute copies of the software, or if 49 | you modify it: responsibilities to respect the freedom of others. 50 | 51 | For example, if you distribute copies of such a program, whether 52 | gratis or for a fee, you must pass on to the recipients the same 53 | freedoms that you received. You must make sure that they, too, receive 54 | or can get the source code. And you must show them these terms so they 55 | know their rights. 56 | 57 | Developers that use the GNU GPL protect your rights with two steps: 58 | (1) assert copyright on the software, and (2) offer you this License 59 | giving you legal permission to copy, distribute and/or modify it. 60 | 61 | For the developers' and authors' protection, the GPL clearly explains 62 | that there is no warranty for this free software. For both users' and 63 | authors' sake, the GPL requires that modified versions be marked as 64 | changed, so that their problems will not be attributed erroneously to 65 | authors of previous versions. 66 | 67 | Some devices are designed to deny users access to install or run 68 | modified versions of the software inside them, although the manufacturer 69 | can do so. This is fundamentally incompatible with the aim of 70 | protecting users' freedom to change the software. The systematic 71 | pattern of such abuse occurs in the area of products for individuals to 72 | use, which is precisely where it is most unacceptable. Therefore, we 73 | have designed this version of the GPL to prohibit the practice for those 74 | products. If such problems arise substantially in other domains, we 75 | stand ready to extend this provision to those domains in future versions 76 | of the GPL, as needed to protect the freedom of users. 77 | 78 | Finally, every program is threatened constantly by software patents. 79 | States should not allow patents to restrict development and use of 80 | software on general-purpose computers, but in those that do, we wish to 81 | avoid the special danger that patents applied to a free program could 82 | make it effectively proprietary. To prevent this, the GPL assures that 83 | patents cannot be used to render the program non-free. 84 | 85 | The precise terms and conditions for copying, distribution and 86 | modification follow. 87 | 88 | TERMS AND CONDITIONS 89 | 90 | 0. Definitions. 91 | 92 | "This License" refers to version 3 of the GNU General Public License. 93 | 94 | "Copyright" also means copyright-like laws that apply to other kinds of 95 | works, such as semiconductor masks. 96 | 97 | "The Program" refers to any copyrightable work licensed under this 98 | License. Each licensee is addressed as "you". "Licensees" and 99 | "recipients" may be individuals or organizations. 100 | 101 | To "modify" a work means to copy from or adapt all or part of the work 102 | in a fashion requiring copyright permission, other than the making of an 103 | exact copy. The resulting work is called a "modified version" of the 104 | earlier work or a work "based on" the earlier work. 105 | 106 | A "covered work" means either the unmodified Program or a work based 107 | on the Program. 108 | 109 | To "propagate" a work means to do anything with it that, without 110 | permission, would make you directly or secondarily liable for 111 | infringement under applicable copyright law, except executing it on a 112 | computer or modifying a private copy. Propagation includes copying, 113 | distribution (with or without modification), making available to the 114 | public, and in some countries other activities as well. 115 | 116 | To "convey" a work means any kind of propagation that enables other 117 | parties to make or receive copies. Mere interaction with a user through 118 | a computer network, with no transfer of a copy, is not conveying. 119 | 120 | An interactive user interface displays "Appropriate Legal Notices" 121 | to the extent that it includes a convenient and prominently visible 122 | feature that (1) displays an appropriate copyright notice, and (2) 123 | tells the user that there is no warranty for the work (except to the 124 | extent that warranties are provided), that licensees may convey the 125 | work under this License, and how to view a copy of this License. If 126 | the interface presents a list of user commands or options, such as a 127 | menu, a prominent item in the list meets this criterion. 128 | 129 | 1. Source Code. 130 | 131 | The "source code" for a work means the preferred form of the work 132 | for making modifications to it. "Object code" means any non-source 133 | form of a work. 134 | 135 | A "Standard Interface" means an interface that either is an official 136 | standard defined by a recognized standards body, or, in the case of 137 | interfaces specified for a particular programming language, one that 138 | is widely used among developers working in that language. 139 | 140 | The "System Libraries" of an executable work include anything, other 141 | than the work as a whole, that (a) is included in the normal form of 142 | packaging a Major Component, but which is not part of that Major 143 | Component, and (b) serves only to enable use of the work with that 144 | Major Component, or to implement a Standard Interface for which an 145 | implementation is available to the public in source code form. A 146 | "Major Component", in this context, means a major essential component 147 | (kernel, window system, and so on) of the specific operating system 148 | (if any) on which the executable work runs, or a compiler used to 149 | produce the work, or an object code interpreter used to run it. 150 | 151 | The "Corresponding Source" for a work in object code form means all 152 | the source code needed to generate, install, and (for an executable 153 | work) run the object code and to modify the work, including scripts to 154 | control those activities. However, it does not include the work's 155 | System Libraries, or general-purpose tools or generally available free 156 | programs which are used unmodified in performing those activities but 157 | which are not part of the work. For example, Corresponding Source 158 | includes interface definition files associated with source files for 159 | the work, and the source code for shared libraries and dynamically 160 | linked subprograms that the work is specifically designed to require, 161 | such as by intimate data communication or control flow between those 162 | subprograms and other parts of the work. 163 | 164 | The Corresponding Source need not include anything that users 165 | can regenerate automatically from other parts of the Corresponding 166 | Source. 167 | 168 | The Corresponding Source for a work in source code form is that 169 | same work. 170 | 171 | 2. Basic Permissions. 172 | 173 | All rights granted under this License are granted for the term of 174 | copyright on the Program, and are irrevocable provided the stated 175 | conditions are met. This License explicitly affirms your unlimited 176 | permission to run the unmodified Program. The output from running a 177 | covered work is covered by this License only if the output, given its 178 | content, constitutes a covered work. This License acknowledges your 179 | rights of fair use or other equivalent, as provided by copyright law. 180 | 181 | You may make, run and propagate covered works that you do not 182 | convey, without conditions so long as your license otherwise remains 183 | in force. You may convey covered works to others for the sole purpose 184 | of having them make modifications exclusively for you, or provide you 185 | with facilities for running those works, provided that you comply with 186 | the terms of this License in conveying all material for which you do 187 | not control copyright. Those thus making or running the covered works 188 | for you must do so exclusively on your behalf, under your direction 189 | and control, on terms that prohibit them from making any copies of 190 | your copyrighted material outside their relationship with you. 191 | 192 | Conveying under any other circumstances is permitted solely under 193 | the conditions stated below. Sublicensing is not allowed; section 10 194 | makes it unnecessary. 195 | 196 | 3. Protecting Users' Legal Rights From Anti-Circumvention Law. 197 | 198 | No covered work shall be deemed part of an effective technological 199 | measure under any applicable law fulfilling obligations under article 200 | 11 of the WIPO copyright treaty adopted on 20 December 1996, or 201 | similar laws prohibiting or restricting circumvention of such 202 | measures. 203 | 204 | When you convey a covered work, you waive any legal power to forbid 205 | circumvention of technological measures to the extent such circumvention 206 | is effected by exercising rights under this License with respect to 207 | the covered work, and you disclaim any intention to limit operation or 208 | modification of the work as a means of enforcing, against the work's 209 | users, your or third parties' legal rights to forbid circumvention of 210 | technological measures. 211 | 212 | 4. Conveying Verbatim Copies. 213 | 214 | You may convey verbatim copies of the Program's source code as you 215 | receive it, in any medium, provided that you conspicuously and 216 | appropriately publish on each copy an appropriate copyright notice; 217 | keep intact all notices stating that this License and any 218 | non-permissive terms added in accord with section 7 apply to the code; 219 | keep intact all notices of the absence of any warranty; and give all 220 | recipients a copy of this License along with the Program. 221 | 222 | You may charge any price or no price for each copy that you convey, 223 | and you may offer support or warranty protection for a fee. 224 | 225 | 5. Conveying Modified Source Versions. 226 | 227 | You may convey a work based on the Program, or the modifications to 228 | produce it from the Program, in the form of source code under the 229 | terms of section 4, provided that you also meet all of these conditions: 230 | 231 | a) The work must carry prominent notices stating that you modified 232 | it, and giving a relevant date. 233 | 234 | b) The work must carry prominent notices stating that it is 235 | released under this License and any conditions added under section 236 | 7. This requirement modifies the requirement in section 4 to 237 | "keep intact all notices". 238 | 239 | c) You must license the entire work, as a whole, under this 240 | License to anyone who comes into possession of a copy. This 241 | License will therefore apply, along with any applicable section 7 242 | additional terms, to the whole of the work, and all its parts, 243 | regardless of how they are packaged. This License gives no 244 | permission to license the work in any other way, but it does not 245 | invalidate such permission if you have separately received it. 246 | 247 | d) If the work has interactive user interfaces, each must display 248 | Appropriate Legal Notices; however, if the Program has interactive 249 | interfaces that do not display Appropriate Legal Notices, your 250 | work need not make them do so. 251 | 252 | A compilation of a covered work with other separate and independent 253 | works, which are not by their nature extensions of the covered work, 254 | and which are not combined with it such as to form a larger program, 255 | in or on a volume of a storage or distribution medium, is called an 256 | "aggregate" if the compilation and its resulting copyright are not 257 | used to limit the access or legal rights of the compilation's users 258 | beyond what the individual works permit. Inclusion of a covered work 259 | in an aggregate does not cause this License to apply to the other 260 | parts of the aggregate. 261 | 262 | 6. Conveying Non-Source Forms. 263 | 264 | You may convey a covered work in object code form under the terms 265 | of sections 4 and 5, provided that you also convey the 266 | machine-readable Corresponding Source under the terms of this License, 267 | in one of these ways: 268 | 269 | a) Convey the object code in, or embodied in, a physical product 270 | (including a physical distribution medium), accompanied by the 271 | Corresponding Source fixed on a durable physical medium 272 | customarily used for software interchange. 273 | 274 | b) Convey the object code in, or embodied in, a physical product 275 | (including a physical distribution medium), accompanied by a 276 | written offer, valid for at least three years and valid for as 277 | long as you offer spare parts or customer support for that product 278 | model, to give anyone who possesses the object code either (1) a 279 | copy of the Corresponding Source for all the software in the 280 | product that is covered by this License, on a durable physical 281 | medium customarily used for software interchange, for a price no 282 | more than your reasonable cost of physically performing this 283 | conveying of source, or (2) access to copy the 284 | Corresponding Source from a network server at no charge. 285 | 286 | c) Convey individual copies of the object code with a copy of the 287 | written offer to provide the Corresponding Source. This 288 | alternative is allowed only occasionally and noncommercially, and 289 | only if you received the object code with such an offer, in accord 290 | with subsection 6b. 291 | 292 | d) Convey the object code by offering access from a designated 293 | place (gratis or for a charge), and offer equivalent access to the 294 | Corresponding Source in the same way through the same place at no 295 | further charge. You need not require recipients to copy the 296 | Corresponding Source along with the object code. If the place to 297 | copy the object code is a network server, the Corresponding Source 298 | may be on a different server (operated by you or a third party) 299 | that supports equivalent copying facilities, provided you maintain 300 | clear directions next to the object code saying where to find the 301 | Corresponding Source. Regardless of what server hosts the 302 | Corresponding Source, you remain obligated to ensure that it is 303 | available for as long as needed to satisfy these requirements. 304 | 305 | e) Convey the object code using peer-to-peer transmission, provided 306 | you inform other peers where the object code and Corresponding 307 | Source of the work are being offered to the general public at no 308 | charge under subsection 6d. 309 | 310 | A separable portion of the object code, whose source code is excluded 311 | from the Corresponding Source as a System Library, need not be 312 | included in conveying the object code work. 313 | 314 | A "User Product" is either (1) a "consumer product", which means any 315 | tangible personal property which is normally used for personal, family, 316 | or household purposes, or (2) anything designed or sold for incorporation 317 | into a dwelling. In determining whether a product is a consumer product, 318 | doubtful cases shall be resolved in favor of coverage. For a particular 319 | product received by a particular user, "normally used" refers to a 320 | typical or common use of that class of product, regardless of the status 321 | of the particular user or of the way in which the particular user 322 | actually uses, or expects or is expected to use, the product. A product 323 | is a consumer product regardless of whether the product has substantial 324 | commercial, industrial or non-consumer uses, unless such uses represent 325 | the only significant mode of use of the product. 326 | 327 | "Installation Information" for a User Product means any methods, 328 | procedures, authorization keys, or other information required to install 329 | and execute modified versions of a covered work in that User Product from 330 | a modified version of its Corresponding Source. The information must 331 | suffice to ensure that the continued functioning of the modified object 332 | code is in no case prevented or interfered with solely because 333 | modification has been made. 334 | 335 | If you convey an object code work under this section in, or with, or 336 | specifically for use in, a User Product, and the conveying occurs as 337 | part of a transaction in which the right of possession and use of the 338 | User Product is transferred to the recipient in perpetuity or for a 339 | fixed term (regardless of how the transaction is characterized), the 340 | Corresponding Source conveyed under this section must be accompanied 341 | by the Installation Information. But this requirement does not apply 342 | if neither you nor any third party retains the ability to install 343 | modified object code on the User Product (for example, the work has 344 | been installed in ROM). 345 | 346 | The requirement to provide Installation Information does not include a 347 | requirement to continue to provide support service, warranty, or updates 348 | for a work that has been modified or installed by the recipient, or for 349 | the User Product in which it has been modified or installed. Access to a 350 | network may be denied when the modification itself materially and 351 | adversely affects the operation of the network or violates the rules and 352 | protocols for communication across the network. 353 | 354 | Corresponding Source conveyed, and Installation Information provided, 355 | in accord with this section must be in a format that is publicly 356 | documented (and with an implementation available to the public in 357 | source code form), and must require no special password or key for 358 | unpacking, reading or copying. 359 | 360 | 7. Additional Terms. 361 | 362 | "Additional permissions" are terms that supplement the terms of this 363 | License by making exceptions from one or more of its conditions. 364 | Additional permissions that are applicable to the entire Program shall 365 | be treated as though they were included in this License, to the extent 366 | that they are valid under applicable law. If additional permissions 367 | apply only to part of the Program, that part may be used separately 368 | under those permissions, but the entire Program remains governed by 369 | this License without regard to the additional permissions. 370 | 371 | When you convey a copy of a covered work, you may at your option 372 | remove any additional permissions from that copy, or from any part of 373 | it. (Additional permissions may be written to require their own 374 | removal in certain cases when you modify the work.) You may place 375 | additional permissions on material, added by you to a covered work, 376 | for which you have or can give appropriate copyright permission. 377 | 378 | Notwithstanding any other provision of this License, for material you 379 | add to a covered work, you may (if authorized by the copyright holders of 380 | that material) supplement the terms of this License with terms: 381 | 382 | a) Disclaiming warranty or limiting liability differently from the 383 | terms of sections 15 and 16 of this License; or 384 | 385 | b) Requiring preservation of specified reasonable legal notices or 386 | author attributions in that material or in the Appropriate Legal 387 | Notices displayed by works containing it; or 388 | 389 | c) Prohibiting misrepresentation of the origin of that material, or 390 | requiring that modified versions of such material be marked in 391 | reasonable ways as different from the original version; or 392 | 393 | d) Limiting the use for publicity purposes of names of licensors or 394 | authors of the material; or 395 | 396 | e) Declining to grant rights under trademark law for use of some 397 | trade names, trademarks, or service marks; or 398 | 399 | f) Requiring indemnification of licensors and authors of that 400 | material by anyone who conveys the material (or modified versions of 401 | it) with contractual assumptions of liability to the recipient, for 402 | any liability that these contractual assumptions directly impose on 403 | those licensors and authors. 404 | 405 | All other non-permissive additional terms are considered "further 406 | restrictions" within the meaning of section 10. If the Program as you 407 | received it, or any part of it, contains a notice stating that it is 408 | governed by this License along with a term that is a further 409 | restriction, you may remove that term. If a license document contains 410 | a further restriction but permits relicensing or conveying under this 411 | License, you may add to a covered work material governed by the terms 412 | of that license document, provided that the further restriction does 413 | not survive such relicensing or conveying. 414 | 415 | If you add terms to a covered work in accord with this section, you 416 | must place, in the relevant source files, a statement of the 417 | additional terms that apply to those files, or a notice indicating 418 | where to find the applicable terms. 419 | 420 | Additional terms, permissive or non-permissive, may be stated in the 421 | form of a separately written license, or stated as exceptions; 422 | the above requirements apply either way. 423 | 424 | 8. Termination. 425 | 426 | You may not propagate or modify a covered work except as expressly 427 | provided under this License. Any attempt otherwise to propagate or 428 | modify it is void, and will automatically terminate your rights under 429 | this License (including any patent licenses granted under the third 430 | paragraph of section 11). 431 | 432 | However, if you cease all violation of this License, then your 433 | license from a particular copyright holder is reinstated (a) 434 | provisionally, unless and until the copyright holder explicitly and 435 | finally terminates your license, and (b) permanently, if the copyright 436 | holder fails to notify you of the violation by some reasonable means 437 | prior to 60 days after the cessation. 438 | 439 | Moreover, your license from a particular copyright holder is 440 | reinstated permanently if the copyright holder notifies you of the 441 | violation by some reasonable means, this is the first time you have 442 | received notice of violation of this License (for any work) from that 443 | copyright holder, and you cure the violation prior to 30 days after 444 | your receipt of the notice. 445 | 446 | Termination of your rights under this section does not terminate the 447 | licenses of parties who have received copies or rights from you under 448 | this License. If your rights have been terminated and not permanently 449 | reinstated, you do not qualify to receive new licenses for the same 450 | material under section 10. 451 | 452 | 9. Acceptance Not Required for Having Copies. 453 | 454 | You are not required to accept this License in order to receive or 455 | run a copy of the Program. Ancillary propagation of a covered work 456 | occurring solely as a consequence of using peer-to-peer transmission 457 | to receive a copy likewise does not require acceptance. However, 458 | nothing other than this License grants you permission to propagate or 459 | modify any covered work. These actions infringe copyright if you do 460 | not accept this License. Therefore, by modifying or propagating a 461 | covered work, you indicate your acceptance of this License to do so. 462 | 463 | 10. Automatic Licensing of Downstream Recipients. 464 | 465 | Each time you convey a covered work, the recipient automatically 466 | receives a license from the original licensors, to run, modify and 467 | propagate that work, subject to this License. You are not responsible 468 | for enforcing compliance by third parties with this License. 469 | 470 | An "entity transaction" is a transaction transferring control of an 471 | organization, or substantially all assets of one, or subdividing an 472 | organization, or merging organizations. If propagation of a covered 473 | work results from an entity transaction, each party to that 474 | transaction who receives a copy of the work also receives whatever 475 | licenses to the work the party's predecessor in interest had or could 476 | give under the previous paragraph, plus a right to possession of the 477 | Corresponding Source of the work from the predecessor in interest, if 478 | the predecessor has it or can get it with reasonable efforts. 479 | 480 | You may not impose any further restrictions on the exercise of the 481 | rights granted or affirmed under this License. For example, you may 482 | not impose a license fee, royalty, or other charge for exercise of 483 | rights granted under this License, and you may not initiate litigation 484 | (including a cross-claim or counterclaim in a lawsuit) alleging that 485 | any patent claim is infringed by making, using, selling, offering for 486 | sale, or importing the Program or any portion of it. 487 | 488 | 11. Patents. 489 | 490 | A "contributor" is a copyright holder who authorizes use under this 491 | License of the Program or a work on which the Program is based. The 492 | work thus licensed is called the contributor's "contributor version". 493 | 494 | A contributor's "essential patent claims" are all patent claims 495 | owned or controlled by the contributor, whether already acquired or 496 | hereafter acquired, that would be infringed by some manner, permitted 497 | by this License, of making, using, or selling its contributor version, 498 | but do not include claims that would be infringed only as a 499 | consequence of further modification of the contributor version. For 500 | purposes of this definition, "control" includes the right to grant 501 | patent sublicenses in a manner consistent with the requirements of 502 | this License. 503 | 504 | Each contributor grants you a non-exclusive, worldwide, royalty-free 505 | patent license under the contributor's essential patent claims, to 506 | make, use, sell, offer for sale, import and otherwise run, modify and 507 | propagate the contents of its contributor version. 508 | 509 | In the following three paragraphs, a "patent license" is any express 510 | agreement or commitment, however denominated, not to enforce a patent 511 | (such as an express permission to practice a patent or covenant not to 512 | sue for patent infringement). To "grant" such a patent license to a 513 | party means to make such an agreement or commitment not to enforce a 514 | patent against the party. 515 | 516 | If you convey a covered work, knowingly relying on a patent license, 517 | and the Corresponding Source of the work is not available for anyone 518 | to copy, free of charge and under the terms of this License, through a 519 | publicly available network server or other readily accessible means, 520 | then you must either (1) cause the Corresponding Source to be so 521 | available, or (2) arrange to deprive yourself of the benefit of the 522 | patent license for this particular work, or (3) arrange, in a manner 523 | consistent with the requirements of this License, to extend the patent 524 | license to downstream recipients. "Knowingly relying" means you have 525 | actual knowledge that, but for the patent license, your conveying the 526 | covered work in a country, or your recipient's use of the covered work 527 | in a country, would infringe one or more identifiable patents in that 528 | country that you have reason to believe are valid. 529 | 530 | If, pursuant to or in connection with a single transaction or 531 | arrangement, you convey, or propagate by procuring conveyance of, a 532 | covered work, and grant a patent license to some of the parties 533 | receiving the covered work authorizing them to use, propagate, modify 534 | or convey a specific copy of the covered work, then the patent license 535 | you grant is automatically extended to all recipients of the covered 536 | work and works based on it. 537 | 538 | A patent license is "discriminatory" if it does not include within 539 | the scope of its coverage, prohibits the exercise of, or is 540 | conditioned on the non-exercise of one or more of the rights that are 541 | specifically granted under this License. You may not convey a covered 542 | work if you are a party to an arrangement with a third party that is 543 | in the business of distributing software, under which you make payment 544 | to the third party based on the extent of your activity of conveying 545 | the work, and under which the third party grants, to any of the 546 | parties who would receive the covered work from you, a discriminatory 547 | patent license (a) in connection with copies of the covered work 548 | conveyed by you (or copies made from those copies), or (b) primarily 549 | for and in connection with specific products or compilations that 550 | contain the covered work, unless you entered into that arrangement, 551 | or that patent license was granted, prior to 28 March 2007. 552 | 553 | Nothing in this License shall be construed as excluding or limiting 554 | any implied license or other defenses to infringement that may 555 | otherwise be available to you under applicable patent law. 556 | 557 | 12. No Surrender of Others' Freedom. 558 | 559 | If conditions are imposed on you (whether by court order, agreement or 560 | otherwise) that contradict the conditions of this License, they do not 561 | excuse you from the conditions of this License. If you cannot convey a 562 | covered work so as to satisfy simultaneously your obligations under this 563 | License and any other pertinent obligations, then as a consequence you may 564 | not convey it at all. For example, if you agree to terms that obligate you 565 | to collect a royalty for further conveying from those to whom you convey 566 | the Program, the only way you could satisfy both those terms and this 567 | License would be to refrain entirely from conveying the Program. 568 | 569 | 13. Use with the GNU Affero General Public License. 570 | 571 | Notwithstanding any other provision of this License, you have 572 | permission to link or combine any covered work with a work licensed 573 | under version 3 of the GNU Affero General Public License into a single 574 | combined work, and to convey the resulting work. The terms of this 575 | License will continue to apply to the part which is the covered work, 576 | but the special requirements of the GNU Affero General Public License, 577 | section 13, concerning interaction through a network will apply to the 578 | combination as such. 579 | 580 | 14. Revised Versions of this License. 581 | 582 | The Free Software Foundation may publish revised and/or new versions of 583 | the GNU General Public License from time to time. Such new versions will 584 | be similar in spirit to the present version, but may differ in detail to 585 | address new problems or concerns. 586 | 587 | Each version is given a distinguishing version number. If the 588 | Program specifies that a certain numbered version of the GNU General 589 | Public License "or any later version" applies to it, you have the 590 | option of following the terms and conditions either of that numbered 591 | version or of any later version published by the Free Software 592 | Foundation. If the Program does not specify a version number of the 593 | GNU General Public License, you may choose any version ever published 594 | by the Free Software Foundation. 595 | 596 | If the Program specifies that a proxy can decide which future 597 | versions of the GNU General Public License can be used, that proxy's 598 | public statement of acceptance of a version permanently authorizes you 599 | to choose that version for the Program. 600 | 601 | Later license versions may give you additional or different 602 | permissions. However, no additional obligations are imposed on any 603 | author or copyright holder as a result of your choosing to follow a 604 | later version. 605 | 606 | 15. Disclaimer of Warranty. 607 | 608 | THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY 609 | APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT 610 | HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY 611 | OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, 612 | THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 613 | PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM 614 | IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF 615 | ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 616 | 617 | 16. Limitation of Liability. 618 | 619 | IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING 620 | WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS 621 | THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY 622 | GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE 623 | USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF 624 | DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD 625 | PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), 626 | EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF 627 | SUCH DAMAGES. 628 | 629 | 17. Interpretation of Sections 15 and 16. 630 | 631 | If the disclaimer of warranty and limitation of liability provided 632 | above cannot be given local legal effect according to their terms, 633 | reviewing courts shall apply local law that most closely approximates 634 | an absolute waiver of all civil liability in connection with the 635 | Program, unless a warranty or assumption of liability accompanies a 636 | copy of the Program in return for a fee. 637 | 638 | END OF TERMS AND CONDITIONS 639 | 640 | How to Apply These Terms to Your New Programs 641 | 642 | If you develop a new program, and you want it to be of the greatest 643 | possible use to the public, the best way to achieve this is to make it 644 | free software which everyone can redistribute and change under these terms. 645 | 646 | To do so, attach the following notices to the program. It is safest 647 | to attach them to the start of each source file to most effectively 648 | state the exclusion of warranty; and each file should have at least 649 | the "copyright" line and a pointer to where the full notice is found. 650 | 651 | 652 | Copyright (C) 2018 2018 PixLab| Symisc Systems. https://sod.pixlab.io 653 | 654 | This program is free software: you can redistribute it and/or modify 655 | it under the terms of the GNU General Public License as published by 656 | the Free Software Foundation, either version 3 of the License, or 657 | (at your option) any later version. 658 | 659 | This program is distributed in the hope that it will be useful, 660 | but WITHOUT ANY WARRANTY; without even the implied warranty of 661 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 662 | GNU General Public License for more details. 663 | 664 | You should have received a copy of the GNU General Public License 665 | along with this program. If not, see . 666 | 667 | Also add information on how to contact you by electronic and paper mail. 668 | 669 | If the program does terminal interaction, make it output a short 670 | notice like this when it starts in an interactive mode: 671 | 672 | Symisc SOD Copyright (C) 2018 2018 PixLab| Symisc Systems. https://sod.pixlab.io 673 | 674 | This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'. 675 | This is free software, and you are welcome to redistribute it 676 | under certain conditions; type `show c' for details. 677 | 678 | The hypothetical commands `show w' and `show c' should show the appropriate 679 | parts of the General Public License. Of course, your program's commands 680 | might be different; for a GUI interface, you would use an "about box". 681 | 682 | You should also get your employer (if you work as a programmer) or school, 683 | if any, to sign a "copyright disclaimer" for the program, if necessary. 684 | For more information on this, and how to apply and follow the GNU GPL, see 685 | . 686 | 687 | The GNU General Public License does not permit incorporating your program 688 | into proprietary programs. If your program is a subroutine library, you 689 | may consider it more useful to permit linking proprietary applications with 690 | the library. If this is what you want to do, use the GNU Lesser General 691 | Public License instead of this License. But first, please read 692 | . 693 | -------------------------------------------------------------------------------- /Makefile: -------------------------------------------------------------------------------- 1 | # SOD does not generally require a Makefile to build. Just drop sod.c and its accompanying 2 | # header files on your source tree and you are done. 3 | CC = clang 4 | CFLAGS = -lm -Ofast -march=native -Wall -std=c99 5 | 6 | sod: sod.c 7 | $(CC) sod.c samples/cnn_face_detection.c -o sod_face_detect -I. $(CFLAGS) -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 |

SOD

An Embedded Computer Vision & Machine Learning Library
sod.pixlab.io

2 | 3 | [![API documentation](https://img.shields.io/badge/API%20documentation-Ready-green.svg)](https://sod.pixlab.io/api.html) 4 | [![dependency](https://img.shields.io/badge/dependency-none-ff96b4.svg)](https://pixlab.io/downloads) 5 | [![Getting Started](https://img.shields.io/badge/Getting%20Started-Now-f49242.svg)](https://sod.pixlab.io/intro.html) 6 | [![license](https://img.shields.io/badge/License-dual--licensed-blue.svg)](https://pixlab.io/downloads) 7 | [![Forum](https://img.shields.io/gitter/room/nwjs/nw.js.svg)](https://community.faceio.net/) 8 | [![Tiny Dreal](https://pixlab.io/images/logo.png)](https://pixlab.io/tiny-dream) 9 | 10 | ![Output](https://i.imgur.com/YIbb8wr.jpg) 11 | 12 | * [Introduction](#sod-embedded). 13 | * [Features](#notable-sod-features). 14 | * [Programming with SOD](#programming-interfaces). 15 | * [Useful Links](#other-useful-links). 16 | 17 | ## SOD Embedded 18 | 19 | ### Release 1.1.9 (July 2023) | [Changelog](https://sod.pixlab.io/changelog.html) | [Downloads](https://pixlab.io/downloads) 20 | 21 | SOD is an embedded, modern cross-platform computer vision and machine learning software library that exposes a set of APIs for deep-learning, advanced media analysis & processing including real-time, multi-class object detection and model training on embedded systems with limited computational resource and IoT devices. 22 | 23 | SOD was built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in open source as well commercial products. 24 | 25 | Designed for computational efficiency and with a strong focus on real-time applications. SOD includes a comprehensive set of both classic and state-of-the-art deep-neural networks with their pre-trained models. Built with SOD: 26 | * Convolutional Neural Networks (CNN) for multi-class (20 and 80) object detection & classification. 27 | * Recurrent Neural Networks (RNN) for text generation (i.e. Shakespeare, 4chan, Kant, Python code, etc.). 28 | * Decision trees for single class, real-time object detection. 29 | * A brand new architecture written specifically for SOD named RealNets. 30 | 31 | ![Multi-class object detection](https://i.imgur.com/Mq98uTv.png) 32 | 33 | Cross platform, dependency free, amalgamated (single C file) and heavily optimized. Real world use cases includes: 34 | * Detect & recognize objects (faces included) at Real-time. 35 | * License plate extraction. 36 | * Intrusion detection. 37 | * Mimic Snapchat filters. 38 | * Classify human actions. 39 | * Object identification. 40 | * Eye & Pupil tracking. 41 | * Facial & Body shape extraction. 42 | * Image/Frame segmentation. 43 | 44 | ## Notable SOD features 45 | 46 | * Built for real world and real-time applications. 47 | * State-of-the-art, CPU optimized deep-neural networks including the brand new, exclusive RealNets architecture. 48 | * Patent-free, advanced computer vision algorithms. 49 | * Support major image format. 50 | * Simple, clean and easy to use API. 51 | * Brings deep learning on limited computational resource, embedded systems and IoT devices. 52 | * Easy interpolatable with OpenCV or any other proprietary API. 53 | * Pre-trained models available for most architectures. 54 | * CPU capable, RealNets model training. 55 | * Production ready, cross-platform, high quality source code. 56 | * SOD is dependency free, written in C, compile and run unmodified on virtually any platform & architecture with a decent C compiler. 57 | * Amalgamated - All SOD source files are combined into a single C file (*sod.c*) for easy deployment. 58 | * Open-source, actively developed & maintained product. 59 | * Developer friendly support channels. 60 | 61 | ## Programming Interfaces 62 | 63 | The documentation works both as an API reference and a programming tutorial. It describes the internal structure of the library and guides one in creating applications with a few lines of code. Note that SOD is straightforward to learn, even for new programmer. 64 | 65 | Resources | Description 66 | ------------ | ------------- 67 | SOD in 5 minutes or less | A quick introduction to programming with the SOD Embedded C/C++ API with real-world code samples implemented in C. 68 | C/C++ API Reference Guide | This document describes each API function in details. This is the reference document you should rely on. 69 | C/C++ Code Samples | Real world code samples on how to embed, load models and start experimenting with SOD. 70 | License Plate Detection | Learn how to detect vehicles license plates without heavy Machine Learning techniques, just standard image processing routines already implemented in SOD. 71 | Porting our Face Detector to WebAssembly | Learn how we ported the SOD Realnets face detector into WebAssembly to achieve Real-time performance in the browser. 72 | 73 | ## Other useful links 74 | 75 | Resources | Description 76 | ------------ | ------------- 77 | Downloads | Get a copy of the last public release of SOD, pre-trained models, extensions and more. Start embedding and enjoy programming with. 78 | Copyright/Licensing | SOD is an open-source, dual-licensed product. Find out more about the licensing situation there. 79 | Online Support Channels | Having some trouble integrating SOD? Take a look at our numerous support channels. 80 | 81 | ![face detection using RealNets](https://i.imgur.com/ZLno8Lz.jpg) 82 | -------------------------------------------------------------------------------- /WebAssemby/README.md: -------------------------------------------------------------------------------- 1 | The article depicting the whole process of porting the SOD realnets face detector to WebAssembly is available to consult at: Porting a Face Detector Written in C to WebAssembly. 2 | 3 | This frontal face detector, **WebAssemby model** is pre-trained on the Genki-4K datatset for **Web oriented applications**. 4 | The model is production ready, **works at Real-Time on all modern browsers (mobile devices included)**. Usage instruction already included in the package. 5 | 6 | The model must be downloaded from https://pixlab.io/downloads. Once downloaded, just put it on the directory where the HTML file `usage.html` reside. 7 | 8 | When you deploy the Webassembly face model on your server, make sure 9 | your HTTP server (Apache, Nginx, etc.) return the appropriate MIME type 10 | for the `wasm` file extension. Under Apache, simply put the following 11 | directives on your .htaccess or Virtual host configuration: 12 | 13 | **AddType application/wasm .wasm** 14 | 15 | **AddOutputFilterByType DEFLATE application/wasm** 16 | 17 | 18 | For chrome users, you must test the model on an actual web server, whether served locally (i.e http://127.0.0.1) or remotely. 19 | This is due to the fact that chrome does not allow WebAssembly modules to be loaded directly from the file system (Edge and Firefox do not have such issue). 20 | -------------------------------------------------------------------------------- /WebAssemby/face_detection_browser_usage.html: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | WebAssembly Real-Time Face Detection 6 | 7 | 8 | 13 | 14 |
15 |

WebAssembly Real-Time Face Detection via SOD RealNets

16 |
17 | 18 | 19 |
20 | 31 | 32 | 92 | 93 | 94 | -------------------------------------------------------------------------------- /samples/README.md: -------------------------------------------------------------------------------- 1 | ## Programming Interfaces 2 | 3 | The documentation works both as an API reference and a programming tutorial. It describes the internal structure of the library and guides one in creating applications with a few lines of code. Note that SOD is straightforward to learn, even for new programmer. 4 | 5 | Resources | Description 6 | ------------ | ------------- 7 | SOD in 5 minutes or less | A quick introduction to programming with the SOD Embedded C/C++ API with real-world code samples implemented in C. 8 | C/C++ API Reference Guide | This document describes each API function in details. This is the reference document you should rely on. 9 | C/C++ Code Samples | Real world code samples on how to embed, load models and start experimenting with SOD. 10 | 11 | ## Other useful links 12 | 13 | Resources | Description 14 | ------------ | ------------- 15 | Downloads | Get a copy of the last public release of SOD, pre-trained models, extensions and more. Start embedding and enjoy programming with. 16 | Copyright/Licensing | SOD is an open-source, dual-licensed product. Find out more about the licensing situation there. 17 | Online Support Channels | Having some trouble integrating SOD? Take a look at our numerous support channels. 18 | -------------------------------------------------------------------------------- /samples/acgt.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/acgt.png -------------------------------------------------------------------------------- /samples/batch_img_loading.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c batch_img_loading.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Load all supported image format present in a given directory. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input directory (pass a path or load from the current working directory) */ 34 | const char *zDir = argc > 1 ? argv[1] : "./"; 35 | /* Our image array */ 36 | sod_img *aEntries; 37 | int nEntries; 38 | int i, rc; 39 | /* Bulk loading */ 40 | rc = sod_img_set_load_from_directory(zDir, &aEntries, &nEntries, 100 /* Maximum images to load, pass 0 to load every image */); 41 | if (rc != SOD_OK) { 42 | printf("IO error while loading images from '%s'\n", zDir); 43 | return -1; 44 | } 45 | /* Report */ 46 | printf("%d image(s) were loaded from '%s'..\n", nEntries, zDir); 47 | for (i = 0; i < nEntries; ++i) { 48 | printf("Width: %d, Height: %d, Color channels: %d\n", aEntries[i].w, aEntries[i].h, aEntries[i].c); 49 | /* Uncomment to apply some transformation if desired */ 50 | //sod_grayscale_image_3c(aEntries[i]); 51 | } 52 | /* Cleanup */ 53 | sod_img_set_release(aEntries, nEntries); 54 | return 0; 55 | } 56 | -------------------------------------------------------------------------------- /samples/blob_detection.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c blob_detection.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | static int filter_cb(int width, int height) 29 | { 30 | /* A filter callback invoked by the blob routine each time 31 | * a potential blob region is identified. 32 | * We use the `width` and `height` parameters supplied 33 | * to discard regions of non interest (i.e. too big or too small). 34 | */ 35 | if ((width > 300 && height > 300) || width < 35 || height < 35) { 36 | /* Ignore small or big boxes */ 37 | return 0; 38 | } 39 | return 1; /* Region accepted */ 40 | } 41 | /* 42 | * Perform connected component labeling on an input binary image 43 | * which is useful for finding blobs (i.e. bloc of texts). 44 | */ 45 | int main(int argc, char *argv[]) 46 | { 47 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 48 | const char *zInput = argc > 1 ? argv[1] : "./text.jpg"; 49 | /* Processed output image path */ 50 | const char *zOut = argc > 2 ? argv[2] : "./out_blobs.png"; 51 | /* Load the input image in the grayscale colorspace */ 52 | sod_img imgIn = sod_img_load_grayscale(zInput); 53 | if (imgIn.data == 0) { 54 | /* Invalid path, unsupported format, memory failure, etc. */ 55 | puts("Cannot load input image..exiting"); 56 | return 0; 57 | } 58 | /* A full color copy of the input image so we can draw rose rectangles on blobs. 59 | * Note that drawing on grayscale images is also possible via 60 | * sod_image_draw_box_grayscale() 61 | */ 62 | sod_img imgCopy = sod_img_load_color(zInput); 63 | /* 64 | * Binarize the input image before the dilation process. 65 | */ 66 | sod_img binImg = sod_binarize_image(imgIn, 0); 67 | /* Dilate the binary image, say 12 times */ 68 | sod_img dilImg = sod_dilate_image(binImg, 12); 69 | /* Perform the blob detection process on our dilated image */ 70 | sod_box *box = 0; 71 | int i, nbox; 72 | sod_image_find_blobs(dilImg, &box, &nbox, filter_cb /* Our filter callback to discard small and big regions*/); 73 | /* 74 | * Draw a rectangle on each extracted & validated blob region. 75 | */ 76 | for (i = 0; i < nbox; i++) { 77 | sod_image_draw_bbox_width(imgCopy, box[i], 5, 255., 0, 225.); /* rose box */ 78 | } 79 | /* Finally save the output image to the specified path */ 80 | sod_img_save_as_png(imgCopy, zOut); 81 | /* Cleanup */ 82 | sod_image_blob_boxes_release(box); 83 | sod_free_image(imgIn); 84 | sod_free_image(binImg); 85 | sod_free_image(dilImg); 86 | sod_free_image(imgCopy); 87 | return 0; 88 | } -------------------------------------------------------------------------------- /samples/canny.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/canny.jpg -------------------------------------------------------------------------------- /samples/canny_edge_detection.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c canny_edge_detection.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Perform canny edge detection on an input image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./canny.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_canny.png"; 37 | /* Load the input image in the grayscale colorspace */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_GRAYSCALE/* single channel colorspace (gray)*/); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* Perform canny edge detection. */ 45 | sod_img imgOut = sod_canny_edge_image(imgIn, 0 /* Set this to 1 if you want to reduce noise */); 46 | /* Finally save our processed image to the specified path */ 47 | sod_img_save_as_png(imgOut, zOut); 48 | /* Cleanup */ 49 | sod_free_image(imgIn); 50 | sod_free_image(imgOut); 51 | return 0; 52 | } -------------------------------------------------------------------------------- /samples/cnn_coco.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Convolutional/Recurrent Neural Networks (CNN/RNN) API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c cnn_coco.c -lm -Ofast -march=native -Wall -std=c99 -o sod_cnn_intro 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | int main(int argc, char *argv[]) 29 | { 30 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 31 | const char *zInput = argc > 1 ? argv[1] : "./test.png"; 32 | /* Draw detection boxes (i.e. rectangles) on this output image which 33 | * is a copy of the input plus the boxes. 34 | */ 35 | const char *zOut = argc > 2 ? argv[2] : "./out.png"; 36 | /* 37 | * The CNN handle that should perform the detection process */ 38 | sod_cnn *pNet; 39 | /* Load the input image */ 40 | sod_img imgIn = sod_img_load_from_file(zInput,SOD_IMG_COLOR/* Full colors*/); 41 | if (imgIn.data == 0) { 42 | /* Invalid path, unsupported format, memory failure, etc. */ 43 | puts("Cannot load input image..exiting"); 44 | return 0; 45 | } 46 | /* Make a copy so we can draw anything we want. */ 47 | sod_img imgOut = sod_copy_image(imgIn); 48 | int rc; 49 | const char *zErr; /* Error log if any */ 50 | /* 51 | * Create our CNN handle using the built-in fast 52 | * architecture trained on the MS COCO dataset 53 | * and is able to detect 80 classes of objects at 54 | * real-time on a modern CPU. 55 | */ 56 | rc = sod_cnn_create(&pNet, ":coco", "./tiny80.sod", &zErr); 57 | /* 58 | * ":coco" is the magic word for the built-in MS COCO (80 classes) 59 | * fast architecture. The list of built-in Magic words (pre-ready to use 60 | * configurations and their associated models) are documented here: 61 | * https://sod.pixlab.io/c_api/sod_cnn_create.html. 62 | * 63 | * "tiny80.sod" is the pre-trained model associated with the ":coco" architecture 64 | * and is available to download from https://pixlab.io/downloads 65 | */ 66 | if (rc != SOD_OK) { 67 | /* Display the error message and exit */ 68 | puts(zErr); 69 | return 0; 70 | } 71 | /* 72 | * A sod_box instance always store the coordinates for each detected object 73 | * returned by the CNN via sod_cnn_predict() as we'll see later. 74 | */ 75 | sod_box *box; 76 | int i, nbox; 77 | /* Prepare our input image for the detection process which 78 | * is resized to the network dimension (This op is always very fast) 79 | */ 80 | float * blob = sod_cnn_prepare_image(pNet, imgIn); 81 | if (!blob) { 82 | /* Very unlikely this happen: Invalid architecture, out-of-memory */ 83 | puts("Something went wrong while preparing image.."); 84 | return 0; 85 | } 86 | puts("Starting CNN object detection"); 87 | /* Detect.. */ 88 | sod_cnn_predict(pNet, blob, &box, &nbox); 89 | /* Report the detection result. */ 90 | printf("%d object(s) were detected..\n",nbox); 91 | for (i = 0; i < nbox; i++) { 92 | /* Report the coordinates, name and score of the current detected object */ 93 | printf("(%s) X:%d Y:%d Width:%d Height:%d score:%f%%\n", box[i].zName, box[i].x, box[i].y, box[i].w, box[i].h, box[i].score * 100); 94 | if( box[i].score < 0.3) continue; /* Discard low score detection, remove if you want to report all objects */ 95 | /* 96 | * Draw a rose (RGB: 255,0,255) rectangle of width 3 on the object coordinates. */ 97 | sod_image_draw_bbox_width(imgOut, box[i], 3, 255., 0, 225.); 98 | /* Of course, one could draw a circle via sod_image_draw_circle() or 99 | * crop the entire region via sod_crop_image() instead of drawing a rectangle. */ 100 | } 101 | /* Finally save our output image with the boxes drawn on it */ 102 | sod_img_save_as_png(imgOut, zOut); 103 | /* Cleanup */ 104 | sod_free_image(imgIn); 105 | sod_free_image(imgOut); 106 | /* Release all resources allocated to the CNN handle */ 107 | sod_cnn_destroy(pNet); 108 | return 0; 109 | } -------------------------------------------------------------------------------- /samples/cnn_face_detection.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Convolutional/Recurrent Neural Networks (CNN/RNN) API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c cnn_face_detection.c -lm -Ofast -march=native -Wall -std=c99 -o sod_cnn_intro 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* Real-Time multi-scale face detection using SOD CNN */ 29 | int main(int argc, char *argv[]) 30 | { 31 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 32 | const char *zInput = argc > 1 ? argv[1] : "./cnn_faces.png"; 33 | /* Draw detection boxes (i.e. rectangles) on this output image which 34 | * is a copy of the input plus the boxes. 35 | */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out.png"; 37 | /* 38 | * The CNN handle that should perform the detection process */ 39 | sod_cnn *pNet; 40 | /* Load the input image */ 41 | sod_img imgIn = sod_img_load_from_file(zInput,SOD_IMG_COLOR/* Full colors*/); 42 | if (imgIn.data == 0) { 43 | /* Invalid path, unsupported format, memory failure, etc. */ 44 | puts("Cannot load input image..exiting"); 45 | return 0; 46 | } 47 | /* Make a copy so we can draw anything we want. */ 48 | sod_img imgOut = sod_copy_image(imgIn); 49 | int rc; 50 | const char *zErr; /* Error log if any */ 51 | /* 52 | * Create our CNN handle using the built-in `face` 53 | * architecture trained to detect frontal, partial, 54 | * tiny & large faces at Real-time. 55 | */ 56 | rc = sod_cnn_create(&pNet, ":face", "./face_cnn.sod", &zErr); 57 | /* 58 | * ":face" is the magic word for the built-in face (single class) 59 | * architecture. The list of built-in Magic words (pre-ready to use 60 | * configurations and their associated models) are documented here: 61 | * https://sod.pixlab.io/c_api/sod_cnn_create.html. 62 | * 63 | * "face_cnn.sod" is the pre-trained model associated with the ":face" architecture 64 | * and is available to download from https://pixlab.io/downloads 65 | */ 66 | if (rc != SOD_OK) { 67 | /* Display the error message and exit */ 68 | puts(zErr); 69 | return 0; 70 | } 71 | /* 72 | * A sod_box instance always store the coordinates for each detected object 73 | * returned by the CNN via sod_cnn_predict() as we'll see later. 74 | */ 75 | sod_box *box; 76 | int i, nbox; 77 | /* Prepare our input image for the detection process which 78 | * is resized to the network dimension (This op is always very fast) 79 | */ 80 | float * blob = sod_cnn_prepare_image(pNet, imgIn); 81 | if (!blob) { 82 | /* Very unlikely this happen: Invalid architecture, out-of-memory */ 83 | puts("Something went wrong while preparing image.."); 84 | return 0; 85 | } 86 | puts("Starting CNN face detection"); 87 | /* Detect.. */ 88 | sod_cnn_predict(pNet, blob, &box, &nbox); 89 | /* Report the detection result. */ 90 | printf("%d face(s) were detected..\n",nbox); 91 | for (i = 0; i < nbox; i++) { 92 | /* Report the coordinates and score of the current detected face */ 93 | printf("(%s) X:%d Y:%d Width:%d Height:%d score:%f%%\n", box[i].zName, box[i].x, box[i].y, box[i].w, box[i].h, box[i].score * 100); 94 | if( box[i].score < 0.3) continue; /* Discard low score detection, remove if you want to report all objects */ 95 | /* 96 | * Draw a rose (RGB: 255,0,255) circle of width 3 on the object coordinates. */ 97 | sod_image_draw_circle_thickness(imgOut, box[i].x + (box[i].w / 2), box[i].y + (box[i].h / 2), box[i].w, 5, 255., 0, 225.); 98 | /* Of course, one could draw a box via sod_image_draw_bbox_width() or 99 | * crop the entire region via sod_crop_image() instead of drawing a circle. */ 100 | } 101 | /* Finally save our output image with the boxes drawn on it */ 102 | sod_img_save_as_png(imgOut, zOut); 103 | /* Cleanup */ 104 | sod_free_image(imgIn); 105 | sod_free_image(imgOut); 106 | /* Release all resources allocated to the CNN handle */ 107 | sod_cnn_destroy(pNet); 108 | return 0; 109 | } -------------------------------------------------------------------------------- /samples/cnn_faces.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/cnn_faces.jpg -------------------------------------------------------------------------------- /samples/cnn_object_detection.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Convolutional/Recurrent Neural Networks (CNN/RNN) API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c cnn_object_detection.c -lm -Ofast -march=native -Wall -std=c99 -o sod_cnn_intro 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | int main(int argc, char *argv[]) 29 | { 30 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 31 | const char *zInput = argc > 1 ? argv[1] : "./test.png"; 32 | /* Draw detection boxes (i.e. rectangles) on this output image which 33 | * is a copy of the input plus the boxes. 34 | */ 35 | const char *zOut = argc > 2 ? argv[2] : "./out.png"; 36 | /* 37 | * The CNN handle that should perform the detection process */ 38 | sod_cnn *pNet; 39 | /* Load the input image */ 40 | sod_img imgIn = sod_img_load_from_file(zInput,SOD_IMG_COLOR/* Full colors*/); 41 | if (imgIn.data == 0) { 42 | /* Invalid path, unsupported format, memory failure, etc. */ 43 | puts("Cannot load input image..exiting"); 44 | return 0; 45 | } 46 | /* Make a copy so we can draw anything we want. */ 47 | sod_img imgOut = sod_copy_image(imgIn); 48 | int rc; 49 | const char *zErr; /* Error log if any */ 50 | /* 51 | * Create our CNN handle using the built-in fast 52 | * architecture trained on the Pascal VOC dataset 53 | * and is able to detect 20 classes of objects at 54 | * real-time on a modern CPU. 55 | */ 56 | rc = sod_cnn_create(&pNet, ":fast", "./tiny20.sod", &zErr); 57 | /* 58 | * ":fast" is the magic word for the built-in Pascal VOC (20 classes) 59 | * fast architecture. The list of built-in Magic words (pre-ready to use 60 | * configurations and their associated models) are documented here: 61 | * https://sod.pixlab.io/c_api/sod_cnn_create.html. 62 | * 63 | * "tiny20.sod" is the pre-trained model associated with the ":fast" architecture 64 | * and is available to download from https://pixlab.io/downloads 65 | */ 66 | if (rc != SOD_OK) { 67 | /* Display the error message and exit */ 68 | puts(zErr); 69 | return 0; 70 | } 71 | /* 72 | * A sod_box instance always store the coordinates for each detected object 73 | * returned by the CNN via sod_cnn_predict() as we'll see later. 74 | */ 75 | sod_box *box; 76 | int i, nbox; 77 | /* Prepare our input image for the detection process which 78 | * is resized to the network dimension (This op is always very fast) 79 | */ 80 | float * blob = sod_cnn_prepare_image(pNet, imgIn); 81 | if (!blob) { 82 | /* Very unlikely this happen: Invalid architecture, out-of-memory */ 83 | puts("Something went wrong while preparing image.."); 84 | return 0; 85 | } 86 | puts("Starting CNN object detection"); 87 | /* Detect.. */ 88 | sod_cnn_predict(pNet, blob, &box, &nbox); 89 | /* Report the detection result. */ 90 | printf("%d object(s) were detected..\n",nbox); 91 | for (i = 0; i < nbox; i++) { 92 | /* Report the coordinates, name and score of the current detected object */ 93 | printf("(%s) X:%d Y:%d Width:%d Height:%d score:%f%%\n", box[i].zName, box[i].x, box[i].y, box[i].w, box[i].h, box[i].score * 100); 94 | if( box[i].score < 0.3) continue; /* Discard low score detection, remove if you want to report all objects */ 95 | /* 96 | * Draw a rose (RGB: 255,0,255) rectangle of width 3 on the object coordinates. */ 97 | sod_image_draw_bbox_width(imgOut, box[i], 3, 255., 0, 225.); 98 | /* Of course, one could draw a circle via sod_image_draw_circle() or 99 | * crop the entire region via sod_crop_image() instead of drawing a rectangle. */ 100 | } 101 | /* Finally save our output image with the boxes drawn on it */ 102 | sod_img_save_as_png(imgOut, zOut); 103 | /* Cleanup */ 104 | sod_free_image(imgIn); 105 | sod_free_image(imgOut); 106 | /* Release all resources allocated to the CNN handle */ 107 | sod_cnn_destroy(pNet); 108 | return 0; 109 | } -------------------------------------------------------------------------------- /samples/cnn_voc.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Convolutional/Recurrent Neural Networks (CNN/RNN) API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c cnn_voc.c -lm -Ofast -march=native -Wall -std=c99 -o sod_cnn_intro 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | int main(int argc, char *argv[]) 29 | { 30 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 31 | const char *zInput = argc > 1 ? argv[1] : "./test.png"; 32 | /* Draw detection boxes (i.e. rectangles) on this output image which 33 | * is a copy of the input plus the boxes. 34 | */ 35 | const char *zOut = argc > 2 ? argv[2] : "./out.png"; 36 | /* 37 | * The CNN handle that should perform the detection process */ 38 | sod_cnn *pNet; 39 | /* Load the input image */ 40 | sod_img imgIn = sod_img_load_from_file(zInput,SOD_IMG_COLOR/* Full colors*/); 41 | if (imgIn.data == 0) { 42 | /* Invalid path, unsupported format, memory failure, etc. */ 43 | puts("Cannot load input image..exiting"); 44 | return 0; 45 | } 46 | /* Make a copy so we can draw anything we want. */ 47 | sod_img imgOut = sod_copy_image(imgIn); 48 | int rc; 49 | const char *zErr; /* Error log if any */ 50 | /* 51 | * Create our CNN handle using the built-in fast 52 | * architecture trained on the Pascal VOC dataset 53 | * and is able to detect 20 classes of objects at 54 | * real-time on a modern CPU. 55 | */ 56 | rc = sod_cnn_create(&pNet, ":voc", "./tiny20.sod", &zErr); 57 | /* 58 | * ":voc" is the magic word for the built-in Pascal VOC (20 classes) 59 | * fast architecture. The list of built-in Magic words (pre-ready to use 60 | * configurations and their associated models) are documented here: 61 | * https://sod.pixlab.io/c_api/sod_cnn_create.html. 62 | * 63 | * "tiny20.sod" is the pre-trained model associated with the ":fast" architecture 64 | * and is available to download from https://pixlab.io/downloads 65 | */ 66 | if (rc != SOD_OK) { 67 | /* Display the error message and exit */ 68 | puts(zErr); 69 | return 0; 70 | } 71 | /* 72 | * A sod_box instance always store the coordinates for each detected object 73 | * returned by the CNN via sod_cnn_predict() as we'll see later. 74 | */ 75 | sod_box *box; 76 | int i, nbox; 77 | /* Prepare our input image for the detection process which 78 | * is resized to the network dimension (This op is always very fast) 79 | */ 80 | float * blob = sod_cnn_prepare_image(pNet, imgIn); 81 | if (!blob) { 82 | /* Very unlikely this happen: Invalid architecture, out-of-memory */ 83 | puts("Something went wrong while preparing image.."); 84 | return 0; 85 | } 86 | puts("Starting CNN object detection"); 87 | /* Detect.. */ 88 | sod_cnn_predict(pNet, blob, &box, &nbox); 89 | /* Report the detection result. */ 90 | printf("%d object(s) were detected..\n",nbox); 91 | for (i = 0; i < nbox; i++) { 92 | /* Report the coordinates, name and score of the current detected object */ 93 | printf("(%s) X:%d Y:%d Width:%d Height:%d score:%f%%\n", box[i].zName, box[i].x, box[i].y, box[i].w, box[i].h, box[i].score * 100); 94 | if( box[i].score < 0.3) continue; /* Discard low score detection, remove if you want to report all objects */ 95 | /* 96 | * Draw a rose (RGB: 255,0,255) rectangle of width 3 on the object coordinates. */ 97 | sod_image_draw_bbox_width(imgOut, box[i], 3, 255., 0, 225.); 98 | /* Of course, one could draw a circle via sod_image_draw_circle() or 99 | * crop the entire region via sod_crop_image() instead of drawing a rectangle. */ 100 | } 101 | /* Finally save our output image with the boxes drawn on it */ 102 | sod_img_save_as_png(imgOut, zOut); 103 | /* Cleanup */ 104 | sod_free_image(imgIn); 105 | sod_free_image(imgOut); 106 | /* Release all resources allocated to the CNN handle */ 107 | sod_cnn_destroy(pNet); 108 | return 0; 109 | } -------------------------------------------------------------------------------- /samples/crop_image.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c crop_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Extract a region (crop) from a given image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./flower.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_crop.png"; 37 | /* Load the input image in full color */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_COLOR /* full color channels */); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* Crop offset: Target region coordinates, height & width */ 45 | int x = imgIn.w / 2; 46 | int y = imgIn.h / 2; 47 | int width = 128; 48 | int height = 128; 49 | /* Crop */ 50 | sod_img crop = sod_crop_image(imgIn, x, y, width, height); 51 | /* Save the cropped region to the specified path */ 52 | sod_img_save_as_png(crop, zOut); 53 | /* Cleanup */ 54 | sod_free_image(imgIn); 55 | sod_free_image(crop); 56 | return 0; 57 | } -------------------------------------------------------------------------------- /samples/dilate_image.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c dilate_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Perform dilation on an input binary image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./text.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_dilated.png"; 37 | /* Load the input image in the grayscale colorspace */ 38 | sod_img imgIn = sod_img_load_grayscale(zInput); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* 45 | * Binarize the input image before the dilation process. 46 | */ 47 | sod_img binImg = sod_binarize_image(imgIn, 0); 48 | /* Finally, dilate the binary image, say 12 times */ 49 | sod_img dilImg = sod_dilate_image(binImg, 12); 50 | /* Save the dilated image to the specified path */ 51 | sod_img_save_as_png(dilImg, zOut); 52 | /* Cleanup */ 53 | sod_free_image(imgIn); 54 | sod_free_image(binImg); 55 | sod_free_image(dilImg); 56 | return 0; 57 | } -------------------------------------------------------------------------------- /samples/erode_image.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c erode_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Erode an input binary image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./text.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_dilated.png"; 37 | /* Load the input image in the grayscale colorspace */ 38 | sod_img imgIn = sod_img_load_grayscale(zInput); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* 45 | * Binarize the input image before the dilation process. 46 | */ 47 | sod_img binImg = sod_binarize_image(imgIn, 0); 48 | /* Finally, erode the binary image, say 5 times */ 49 | sod_img erodeImg = sod_erode_image(binImg, 5); 50 | /* Save the eroded image to the specified path */ 51 | sod_img_save_as_png(erodeImg, zOut); 52 | /* Cleanup */ 53 | sod_free_image(imgIn); 54 | sod_free_image(binImg); 55 | sod_free_image(erodeImg); 56 | return 0; 57 | } -------------------------------------------------------------------------------- /samples/fingerprint/fp001_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp001_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp001_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp001_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp002_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp002_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp002_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp002_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp003_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp003_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp003_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp003_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp004_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp004_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp004_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp004_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp005_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp005_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp005_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp005_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp006_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp006_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp006_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp006_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp007_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp007_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp007_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp007_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp008_0.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp008_0.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp008_1.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp008_1.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp_part.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp_part.pgm -------------------------------------------------------------------------------- /samples/fingerprint/fp_whole.pgm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/fingerprint/fp_whole.pgm -------------------------------------------------------------------------------- /samples/flower.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/flower.jpg -------------------------------------------------------------------------------- /samples/gaussian_blur.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c gaussian_blur_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Apply Gaussian Blur Filter to a given RGB/BGR image 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./flower.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_blur.png"; 37 | /* Load the input image in the RGB colorspace */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_COLOR); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* Apply Gaussian Blur on the loaded RGB image. */ 45 | sod_img imgOut = sod_gaussian_blur_image(imgIn, 5, 1.94); 46 | /* Finally save our blurred image to the specified path */ 47 | sod_img_save_as_png(imgOut, zOut); 48 | /* Cleanup */ 49 | sod_free_image(imgIn); 50 | sod_free_image(imgOut); 51 | return 0; 52 | } 53 | -------------------------------------------------------------------------------- /samples/grayscale_image.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c grayscale_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Grayscale a full color image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./test.png"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_gray.png"; 37 | /* Load the input image in full color */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_COLOR /* full color channels */); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* 45 | * Perform grayscale colorspace conversion. 46 | */ 47 | sod_img gray = sod_grayscale_image(imgIn); 48 | /* Save the grayscale image to the specified path */ 49 | sod_img_save_as_png(gray, zOut); 50 | /* Cleanup */ 51 | sod_free_image(imgIn); 52 | sod_free_image(gray); 53 | return 0; 54 | } -------------------------------------------------------------------------------- /samples/hilditch_thin.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c hilditch_thin.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Perform Hilditch thinning on an input image. 30 | * 31 | * Skeletonization is useful when we are interested not in the size of the pattern but rather 32 | * in the relative position of the strokes in the pattern (Character Recognition, 33 | * X, Y Chromosome Recognition, etc.). 34 | * The target image must be binary (i.e. images whose pixels have only two possible intensity 35 | * value mostly black or white). You can obtain a binary image via sod_canny_edge_image(), 36 | * sod_otsu_binarize_image(), sod_binarize_image() or sod_threshold_image(). 37 | */ 38 | int main(int argc, char *argv[]) 39 | { 40 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 41 | const char *zInput = argc > 1 ? argv[1] : "./acgt.png"; 42 | /* Processed output image path */ 43 | const char *zOut = argc > 2 ? argv[2] : "./out_hilditch.png"; 44 | /* Load the input image in the grayscale colorspace */ 45 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_GRAYSCALE/* single channel colorspace (gray)*/); 46 | if (imgIn.data == 0) { 47 | /* Invalid path, unsupported format, memory failure, etc. */ 48 | puts("Cannot load input image..exiting"); 49 | return 0; 50 | } 51 | /* Binarize the input image before the thinning process */ 52 | sod_img binImg = sod_threshold_image(imgIn, 0.5); 53 | /* Perform Hilditch thinning on this binary image. */ 54 | sod_img imgOut = sod_hilditch_thin_image(binImg); 55 | /* Finally save our processed image to the specified path */ 56 | sod_img_save_as_png(imgOut, zOut); 57 | /* Cleanup */ 58 | sod_free_image(imgIn); 59 | sod_free_image(binImg); 60 | sod_free_image(imgOut); 61 | return 0; 62 | } -------------------------------------------------------------------------------- /samples/hough_lines_detection.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c hough_lines_detection.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Perform hough line detection (Generally named Hough Transform) on an input image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./test.png"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_lines.png"; 37 | /* Load the input image in the grayscale colorspace */ 38 | sod_img imgIn = sod_img_load_grayscale(zInput); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* A full color copy of the input image so we can draw later rose lines on it. 45 | */ 46 | sod_img imgCopy = sod_img_load_color(zInput); 47 | /* 48 | * Perform Canny edge detection first which is a mandatory step */ 49 | sod_img cannyImg = sod_canny_edge_image(imgIn, 0); 50 | /* 51 | * Each detected line is represented by an instance of the `sod_pts` 52 | * structure returned as an array by the Hough interface where 53 | * each entry of this array (i and i + 1) hold the starting and 54 | * ending position (x_start, y_start, x_end, y_end) for each line. 55 | */ 56 | sod_pts * aLines; 57 | int i, nPts, nLines; 58 | /* Perform hough line detection on the canny edged image 59 | * Depending on the analyzed image/frame, you should experiment 60 | * with different thresholds for best results. 61 | */ 62 | aLines = sod_hough_lines_detect(cannyImg, 0 /* default threshold which may not good for all images */, &nPts); 63 | /* Report */ 64 | nLines = nPts / 2; 65 | printf("%d line(s) were detected\n", nLines); 66 | /* Draw a rose line for each entry on the full color image copy */ 67 | for (i = 0; i < nLines; i += 2) { 68 | sod_image_draw_line(imgCopy, aLines[i], aLines[i + 1], 255, 0, 255); 69 | } 70 | /* Finally save the output image to the specified path */ 71 | sod_img_save_as_png(imgCopy, zOut); 72 | /* Cleanup */ 73 | sod_hough_lines_release(aLines); 74 | sod_free_image(imgIn); 75 | sod_free_image(cannyImg); 76 | sod_free_image(imgCopy); 77 | return 0; 78 | } -------------------------------------------------------------------------------- /samples/license_plate_detection.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c license_plate_detection.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Frontal License Plate detection without deep-learning. Only image processing code. 30 | */ 31 | static int filter_cb(int width, int height) 32 | { 33 | /* A filter callback invoked by the blob routine each time 34 | * a potential blob region is identified. 35 | * We use the `width` and `height` parameters supplied 36 | * to discard regions of non interest (i.e. too big or too small). 37 | */ 38 | if ((width > 300 && height > 200) || width < 45 || height < 45) { 39 | /* Ignore small or big boxes (You should take in consideration 40 | * U.S plate size here and adjust accordingly). 41 | */ 42 | return 0; /* Discarded region */ 43 | } 44 | return 1; /* Accepted region */ 45 | } 46 | int main(int argc, char *argv[]) 47 | { 48 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 49 | const char *zInput = argc > 1 ? argv[1] : "./plate.jpg"; 50 | /* Processed output image path */ 51 | const char *zOut = argc > 2 ? argv[2] : "./out_plate.png"; 52 | /* Load the input image in the grayscale colorspace */ 53 | sod_img imgIn = sod_img_load_grayscale(zInput); 54 | if (imgIn.data == 0) { 55 | /* Invalid path, unsupported format, memory failure, etc. */ 56 | puts("Cannot load input image..exiting"); 57 | return 0; 58 | } 59 | /* A full color copy of the input image so we can draw rose boxes 60 | * marking the plate in question if any. 61 | */ 62 | sod_img imgCopy = sod_img_load_color(zInput); 63 | /* Obtain a binary image first */ 64 | sod_img binImg = sod_threshold_image(imgIn, 0.5); 65 | /* 66 | * Perform Canny edge detection next which is a mandatory step 67 | */ 68 | sod_img cannyImg = sod_canny_edge_image(binImg, 1/* Reduce noise */); 69 | /* 70 | * Dilate the image say 12 times but you should experiment 71 | * with different values for best results which depend 72 | * on the quality of the input image/frame. */ 73 | sod_img dilImg = sod_dilate_image(cannyImg, 12); 74 | /* Perform connected component labeling or blob detection 75 | * now on the binary, canny edged, Gaussian noise reduced and 76 | * finally dilated image using our filter callback that should 77 | * discard small or large rectangle areas. 78 | */ 79 | sod_box *box = 0; 80 | int i, nbox; 81 | sod_image_find_blobs(dilImg, &box, &nbox, filter_cb); 82 | /* Draw a box on each potential plate coordinates */ 83 | for (i = 0; i < nbox; i++) { 84 | sod_image_draw_bbox_width(imgCopy, box[i], 5, 255., 0, 225.); // rose box 85 | } 86 | sod_image_blob_boxes_release(box); 87 | /* Finally save the output image to the specified path */ 88 | sod_img_save_as_png(imgCopy, zOut); 89 | /* Cleanup */ 90 | sod_free_image(imgIn); 91 | sod_free_image(cannyImg); 92 | sod_free_image(binImg); 93 | sod_free_image(dilImg); 94 | sod_free_image(imgCopy); 95 | return 0; 96 | } -------------------------------------------------------------------------------- /samples/minutiae.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c minutiae.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * This program extracts ridges and bifurcations from a fingerprint image. 30 | * 31 | * Minutiae extraction is applied to skeletonized fingerprint image 32 | * 33 | * The target image must be binary (i.e. images whose pixels have only two possible intensity 34 | * value mostly black or white). You can obtain a binary image via sod_canny_edge_image(), 35 | * sod_otsu_binarize_image(), sod_binarize_image() or sod_threshold_image(). 36 | * 37 | * More information about this process can be found at https://sod.pixlab.io/c_api/sod_minutiae.html 38 | */ 39 | int main(int argc, char *argv[]) 40 | { 41 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) 42 | * 43 | * Minutiae extraction is applied to skeletonized fingerprint image. 44 | */ 45 | const char *zInput = argc > 1 ? argv[1] : "./fingerprint/fp_whole.pgm"; 46 | /* Processed output image path */ 47 | const char *zOut = argc > 2 ? argv[2] : "./out_fp.png"; 48 | /* Load the input image in the grayscale colorspace */ 49 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_GRAYSCALE/* single channel colorspace (gray)*/); 50 | if (imgIn.data == 0) { 51 | /* Invalid path, unsupported format, memory failure, etc. */ 52 | puts("Cannot load input image..exiting"); 53 | return 0; 54 | } 55 | /* Binarize the input image before the thinning process */ 56 | sod_img binImg = sod_threshold_image(imgIn, 0.5); 57 | 58 | /* Perform Hilditch thinning on this binary image. */ 59 | sod_img hildImg = sod_hilditch_thin_image(binImg); 60 | 61 | /* Perform extraction of minutiae candidates in fingerprint image */ 62 | int total, np1, np2; 63 | sod_img imgOut = sod_minutiae(hildImg, &total, &np1, &np2); 64 | printf("total number of BLACK points = %d\n", total); 65 | printf("number of ending points = %d\n", np1); 66 | printf("number of bifurcations = %d\n\n", np2); 67 | 68 | /* Finally save our processed image to the specified path */ 69 | sod_img_save_as_png(imgOut, zOut); 70 | /* Cleanup */ 71 | sod_free_image(imgIn); 72 | sod_free_image(binImg); 73 | sod_free_image(hildImg); 74 | sod_free_image(imgOut); 75 | return 0; 76 | } -------------------------------------------------------------------------------- /samples/otsu.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/otsu.jpg -------------------------------------------------------------------------------- /samples/otsu_image.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c otsu_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Binarize an input image via the Otsu's method. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./otsu.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_otsu.png"; 37 | /* Load the input image in the grayscale colorspace */ 38 | sod_img imgIn = sod_img_load_grayscale(zInput); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* 45 | * Binarize the input image via the Otsu's method. 46 | */ 47 | sod_img binImg = sod_otsu_binarize_image(imgIn); 48 | /* Save the binary image to the specified path */ 49 | sod_img_save_as_png(binImg, zOut); 50 | /* Cleanup */ 51 | sod_free_image(imgIn); 52 | sod_free_image(binImg); 53 | return 0; 54 | } -------------------------------------------------------------------------------- /samples/out_cnn.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/out_cnn.png -------------------------------------------------------------------------------- /samples/plate.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/plate.jpg -------------------------------------------------------------------------------- /samples/realnet_face_detection.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded RealNets API (Frontal Facial detection). 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c realnet_face_detection.c -lm -Ofast -march=native -Wall -std=c99 -o sod_realnet_face_detect 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | int main(int argc, char *argv[]) 29 | { 30 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 31 | const char *zFile = argc > 1 ? argv[1] : "./realnet_faces.jpg"; 32 | /* 33 | * By default, RealNets are designed to process video streams thanks 34 | * to their very fast processing speed. However, for the sake of simplicity 35 | * we'll stick with images for this programming intro to RealNets. 36 | */ 37 | sod_realnet *pNet; /* Realnet handle */ 38 | int i,rc; 39 | /* 40 | * Allocate a new RealNet handle */ 41 | rc = sod_realnet_create(&pNet); 42 | if (rc != SOD_OK) return rc; 43 | /* 44 | * Register and load a RealNet model. 45 | * You can train your own RealNet model on your CPU using the training interfaces [sod_realnet_train_start()] 46 | * or download pre-trained models like this one from https://pixlab.io/downloads 47 | */ 48 | rc = sod_realnet_load_model_from_disk(pNet, "./face.realnet.sod", 0); 49 | if (rc != SOD_OK) return rc; 50 | /* Load the target image in grayscale colorspace */ 51 | sod_img img = sod_img_load_grayscale(zFile); 52 | if (img.data == 0) { 53 | puts("Cannot load image"); 54 | return 0; 55 | } 56 | /* Load a full color copy of the target image so we draw rose boxes 57 | * Note that drawing on grayscale images is also supported. 58 | */ 59 | sod_img color = sod_img_load_color(zFile); 60 | /* 61 | * convert the grayscale image to blob. 62 | */ 63 | unsigned char *zBlob = sod_image_to_blob(img); 64 | /* 65 | * Bounding boxes array 66 | */ 67 | sod_box *aBoxes; 68 | int nbox; 69 | /* 70 | * Perform Real-Time detection on this blob 71 | */ 72 | rc = sod_realnet_detect(pNet, zBlob, img.w, img.h, &aBoxes, &nbox); 73 | if (rc != SOD_OK) return rc; 74 | /* Consume result */ 75 | printf("%d potential face(s) were detected..\n", nbox); 76 | for (i = 0; i < nbox; i++) { 77 | /* Ignore low score detection */ 78 | if (aBoxes[i].score < 5.0) continue; 79 | /* Report current object */ 80 | printf("(%s) x:%d y:%d w:%d h:%d prob:%f\n", aBoxes[i].zName, aBoxes[i].x, aBoxes[i].y, aBoxes[i].w, aBoxes[i].h, aBoxes[i].score); 81 | /* Draw a rose box on the target coordinates */ 82 | sod_image_draw_bbox_width(color, aBoxes[i], 3, 255., 0, 225.); 83 | //sod_image_draw_circle(color, aBoxes[i].x + (aBoxes[i].w / 2), aBoxes[i].y + (aBoxes[i].h / 2), aBoxes[i].w, 255., 0, 225.); 84 | } 85 | /* Save the detection result */ 86 | sod_img_save_as_png(color, argc > 2 ? argv[2] : "./out.png"); 87 | /* cleanup */ 88 | sod_free_image(img); 89 | sod_free_image(color); 90 | sod_image_free_blob(zBlob); 91 | sod_realnet_destroy(pNet); 92 | return 0; 93 | } -------------------------------------------------------------------------------- /samples/realnet_face_detection_embedded.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded RealNets API (Frontal Facial detection). 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c realnet_face_detection_embedded.c -lm -Ofast -march=native -Wall -std=c99 -o sod_realnet_face_detect 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * The following include is shipped with the built-in realnet embedded face detection model. 30 | * There is no need to bring an independent model in order to detect faces at real-time. 31 | * 32 | * All you need, is to register the model hex array defined in "sod_face_realnet.h" via 33 | * `sod_realnet_load_model_from_mem()` and use the same C/C++ as for detecting faces 34 | * via Realnets as shown below in this sample. 35 | * 36 | * This feature is available in the commercial version of the library. 37 | * You can obtain your commercial license from https://pixlab.io/downloads. 38 | * 39 | * Advantages includes: 40 | * 41 | * Multi-core CPU support for all platforms - Up to 3 ~ 10 times faster processing speed. 42 | * Built-in (C Code), high performance RealNets frontal face detector. 43 | * 75 days of integration & technical assistance. 44 | * Royalty-free commercial licenses without any GPL restrictions. 45 | * Application source code stays private. 46 | */ 47 | #include "sod_face_realnet.h" 48 | int main(int argc, char *argv[]) 49 | { 50 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 51 | const char *zFile = argc > 1 ? argv[1] : "./realnet_faces.jpg"; 52 | /* 53 | * By default, RealNets are designed to process video streams thanks 54 | * to their very fast processing speed. However, for the sake of simplicity 55 | * we'll stick with images for this programming intro to RealNets. 56 | */ 57 | sod_realnet *pNet; /* Realnet handle */ 58 | int i,rc; 59 | /* 60 | * Allocate a new RealNet handle */ 61 | rc = sod_realnet_create(&pNet); 62 | if (rc != SOD_OK) return rc; 63 | /* 64 | * Load our RealNet model from memory. No need to bring an external one like the previous sample. 65 | * 66 | * You can train your own RealNet model on your CPU using the training interfaces [sod_realnet_train_start()] 67 | * or download pre-trained models like this one from https://pixlab.io/downloads 68 | */ 69 | rc = sod_realnet_load_model_from_mem(pNet, face_model, face_model_length, 0); 70 | if (rc != SOD_OK) return rc; 71 | /* Load the target image in grayscale colorspace */ 72 | sod_img img = sod_img_load_grayscale(zFile); 73 | if (img.data == 0) { 74 | puts("Cannot load image"); 75 | return 0; 76 | } 77 | /* Load a full color copy of the target image so we draw rose boxes 78 | * Note that drawing on grayscale images is also supported. 79 | */ 80 | sod_img color = sod_img_load_color(zFile); 81 | /* 82 | * convert the grayscale image to blob. 83 | */ 84 | unsigned char *zBlob = sod_image_to_blob(img); 85 | /* 86 | * Bounding boxes array 87 | */ 88 | sod_box *aBoxes; 89 | int nbox; 90 | /* 91 | * Perform Real-Time detection on this blob 92 | */ 93 | rc = sod_realnet_detect(pNet, zBlob, img.w, img.h, &aBoxes, &nbox); 94 | if (rc != SOD_OK) return rc; 95 | /* Consume result */ 96 | printf("%d potential face(s) were detected..\n", nbox); 97 | for (i = 0; i < nbox; i++) { 98 | /* Ignore low score detection */ 99 | if (aBoxes[i].score < 5.0) continue; 100 | /* Report current object */ 101 | printf("(%s) x:%d y:%d w:%d h:%d prob:%f\n", aBoxes[i].zName, aBoxes[i].x, aBoxes[i].y, aBoxes[i].w, aBoxes[i].h, aBoxes[i].score); 102 | /* Draw a rose box on the target coordinates */ 103 | sod_image_draw_bbox_width(color, aBoxes[i], 3, 255., 0, 225.); 104 | //sod_image_draw_circle(color, aBoxes[i].x + (aBoxes[i].w / 2), aBoxes[i].y + (aBoxes[i].h / 2), aBoxes[i].w, 255., 0, 225.); 105 | } 106 | /* Save the detection result */ 107 | sod_img_save_as_png(color, argc > 2 ? argv[2] : "./out.png"); 108 | /* cleanup */ 109 | sod_free_image(img); 110 | sod_free_image(color); 111 | sod_image_free_blob(zBlob); 112 | sod_realnet_destroy(pNet); 113 | return 0; 114 | } 115 | -------------------------------------------------------------------------------- /samples/realnet_faces.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/realnet_faces.jpg -------------------------------------------------------------------------------- /samples/realnet_train_model.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded RealNets Model Training API. 3 | * Training must be enabled via the compile-time directive SOD_ENABLE_NET_TRAIN. 4 | * 5 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 6 | */ 7 | /* 8 | * Compile this file together with the SOD embedded source code to generate 9 | * the executable. For example: 10 | * 11 | * gcc sod.c realnet_train_model.c -D SOD_ENABLE_NET_TRAIN -lm -Ofast -march=native -Wall -std=c99 -o sod_realnet_train_intro 12 | * 13 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 14 | * header files on your source tree and you're done. If you have any trouble 15 | * integrating SOD in your project, please submit a support request at: 16 | * https://sod.pixlab.io/support.html 17 | */ 18 | /* 19 | * This simple program is a quick introduction on how to embed and start 20 | * experimenting with SOD without having to do a lot of tedious 21 | * reading and configuration. 22 | * 23 | * Make sure you have the latest release of SOD from: 24 | * https://pixlab.io/downloads 25 | * The SOD Embedded C/C++ documentation is available at: 26 | * https://sod.pixlab.io/api.html 27 | */ 28 | #include 29 | #include "sod.h" 30 | /* 31 | * Training log consumer callback that should be called 32 | * by the Realnet trainer to report training progress. 33 | */ 34 | void log_consumer_callback(const char *zText, size_t text_len, void *pUserdata) 35 | { 36 | /* Simply redirect to stdout */ 37 | puts(zText); 38 | } 39 | int main(int argc, char *argv[]) 40 | { 41 | /* Training instructions (i.e. where positive and negative samples 42 | * are located, tree minimal depth, max trees, model copyright notice and so on). 43 | * Pass a path or download one from https://pixlab.io/downloads 44 | */ 45 | const char *zTrainFile = argc > 1 ? argv[1] : "train.txt"; 46 | /* 47 | * Relanet trainer handle 48 | */ 49 | sod_realnet_trainer *pNet; 50 | int rc; 51 | /* Allocate a new Realnet Trainer handle */ 52 | rc = sod_realnet_train_init(&pNet); 53 | if (rc != SOD_OK) return rc; 54 | /* 55 | * Install our training progress log consumer callback. 56 | */ 57 | rc = sod_realnet_train_config(pNet, SOD_REALNET_TR_LOG_CALLBACK, log_consumer_callback, 0); 58 | if (rc != SOD_OK) return rc; 59 | /* 60 | * Where to store the output model. 61 | */ 62 | rc = sod_realnet_train_config(pNet, SOD_REALNET_TR_OUTPUT_MODEL, "./pedestrian_detetcor.realnet"); 63 | if (rc != SOD_OK) return rc; 64 | /* 65 | * Start the heavy training process on your CPU driven by 66 | * the Realnet instructions found on `zTrainFile`. 67 | */ 68 | rc = sod_realnet_train_start(pNet, zTrainFile); 69 | /* Wait some days...*/ 70 | sod_realnet_train_release(pNet); 71 | /* check the progress log and you should find 72 | * a working model on the path you specified earlier. 73 | */ 74 | return rc; 75 | } -------------------------------------------------------------------------------- /samples/resize_image.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c resize_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Resize an image (Minify) to half its original size. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./flower.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_rz.png"; 37 | /* Load the input image in full color */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_COLOR /* full color channels */); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* Resize to half its original size */ 45 | int newWidth = imgIn.w / 2; 46 | int newHeight = imgIn.h / 2; 47 | 48 | sod_img rz = sod_resize_image(imgIn, newWidth, newHeight); 49 | /* Save the resized image to the specified path */ 50 | sod_img_save_as_png(rz, zOut); 51 | /* Cleanup */ 52 | sod_free_image(imgIn); 53 | sod_free_image(rz); 54 | return 0; 55 | } -------------------------------------------------------------------------------- /samples/rnn_text_gen.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Recurrent Neural Networks (RNN) API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c rnn_text_gen.c -lm -Ofast -march=native -Wall -std=c99 -o sod_rnn 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* RNN text generation (i.e. Kant, Shakespeare, Tolstoy, Python code, 4 Chan, etc.) depending 29 | * on the pre-trained model. 30 | */ 31 | static void text_consumer_callback( 32 | const char *zText, /* Text ready to be consumed (always nil terminated) */ 33 | size_t text_len, /* zText[] length */ 34 | void *pUserdata /* Arbitrary user pointer passed verbatim by the RNN layer */ 35 | ) 36 | { 37 | /* 38 | * This is the callback that is called by the RNN 39 | * each time a generated text is ready 40 | * to be consumed. See below on how to install 41 | * this callback. 42 | */ 43 | puts(zText); /* Simply redirect the generated text to stdout */ 44 | } 45 | int main(int argc, char *argv[]) 46 | { 47 | /* 48 | * Path to the pre-trained RNN model where the generated text is based on. 49 | * You can download pre-trained RNN models on https://pixlab.io/downloads. 50 | */ 51 | const char *zRnnModel = argc > 1 ? argv[1] : "./tolstoy-rnn.sod"; 52 | /* 53 | * The RNN (although named sod_cnn) handle that should perform text generation for us */ 54 | sod_cnn *pNet; 55 | int rc; 56 | const char *zErr; /* Error log if any */ 57 | /* 58 | * Create our RNN handle using the built-in `rnn` 59 | * architecture and associate the desired pre-trained 60 | * model with it. 61 | */ 62 | rc = sod_cnn_create(&pNet, ":rnn", zRnnModel, &zErr); 63 | /* 64 | * ":rnn" is the magic word for the built-in RNN 65 | * architecture. The list of built-in Magic words (pre-ready to use 66 | * configurations and their associated models) are documented here: 67 | * https://sod.pixlab.io/c_api/sod_cnn_create.html. 68 | */ 69 | if (rc != SOD_OK) { 70 | /* Display the error message and exit */ 71 | puts(zErr); 72 | return 0; 73 | } 74 | /* Register the text consumer callback */ 75 | sod_cnn_config(pNet, SOD_RNN_CALLBACK, text_consumer_callback, 0 /* user data (NULL) */); 76 | /* Generate a text of length 500 characters */ 77 | sod_cnn_config(pNet, SOD_RNN_TEXT_LENGTH, 500); 78 | /* Seed for the text */ 79 | sod_cnn_config(pNet, SOD_RNN_SEED, "@>"); 80 | /* 81 | * Start the text generation process. The consumer callback 82 | * should redirect the generated text to STDOUT. 83 | */ 84 | sod_cnn_predict(pNet, 0, 0, 0); 85 | /* 86 | * At this stage, text_consumer_callback() should have been called 87 | * and the generated text already consumed. 88 | */ 89 | /* Clean-up */ 90 | sod_cnn_destroy(pNet); 91 | return 0; 92 | } 93 | -------------------------------------------------------------------------------- /samples/rotate_image.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c rotate_image.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Rotate an image 180 degree. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./test.png"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_rotate.png"; 37 | /* Load the input image in full color */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_COLOR /* full color channels */); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* 45 | * Perform the rotation process. 46 | */ 47 | sod_img rot = sod_rotate_image(imgIn, 180.0); 48 | /* Save the rotated image to the specified path */ 49 | sod_img_save_as_png(rot, zOut); 50 | /* Cleanup */ 51 | sod_free_image(imgIn); 52 | sod_free_image(rot); 53 | return 0; 54 | } -------------------------------------------------------------------------------- /samples/sepia_filter.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c sepia_filter.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Apply Sepia Tone (filter) to a given image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./flower.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./sepia.png"; 37 | /* Load the input image in the RGB colorspace */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_COLOR); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* Create the SEPIA tone. */ 45 | sod_image_sepia_filter(imgIn); 46 | /* Finally save our processed image to the specified path */ 47 | sod_img_save_as_png(imgIn, zOut); 48 | /* Cleanup */ 49 | sod_free_image(imgIn); 50 | return 0; 51 | } 52 | -------------------------------------------------------------------------------- /samples/sobel.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/sobel.jpg -------------------------------------------------------------------------------- /samples/sobel_operator_img.c: -------------------------------------------------------------------------------- 1 | /* 2 | * Programming introduction with the SOD Embedded Image Processing API. 3 | * Copyright (C) PixLab | Symisc Systems, https://sod.pixlab.io 4 | */ 5 | /* 6 | * Compile this file together with the SOD embedded source code to generate 7 | * the executable. For example: 8 | * 9 | * gcc sod.c sobel_operator_img.c -lm -Ofast -march=native -Wall -std=c99 -o sod_img_proc 10 | * 11 | * Under Microsoft Visual Studio (>= 2015), just drop `sod.c` and its accompanying 12 | * header files on your source tree and you're done. If you have any trouble 13 | * integrating SOD in your project, please submit a support request at: 14 | * https://sod.pixlab.io/support.html 15 | */ 16 | /* 17 | * This simple program is a quick introduction on how to embed and start 18 | * experimenting with SOD without having to do a lot of tedious 19 | * reading and configuration. 20 | * 21 | * Make sure you have the latest release of SOD from: 22 | * https://pixlab.io/downloads 23 | * The SOD Embedded C/C++ documentation is available at: 24 | * https://sod.pixlab.io/api.html 25 | */ 26 | #include 27 | #include "sod.h" 28 | /* 29 | * Apply Sobel operator on an input image. 30 | */ 31 | int main(int argc, char *argv[]) 32 | { 33 | /* Input image (pass a path or use the test image shipped with the samples ZIP archive) */ 34 | const char *zInput = argc > 1 ? argv[1] : "./sobel.jpg"; 35 | /* Processed output image path */ 36 | const char *zOut = argc > 2 ? argv[2] : "./out_sobel.png"; 37 | /* Load the input image in the grayscale colorspace */ 38 | sod_img imgIn = sod_img_load_from_file(zInput, SOD_IMG_GRAYSCALE/* single channel colorspace (gray)*/); 39 | if (imgIn.data == 0) { 40 | /* Invalid path, unsupported format, memory failure, etc. */ 41 | puts("Cannot load input image..exiting"); 42 | return 0; 43 | } 44 | /* Apply Sobel operator. */ 45 | sod_img imgOut = sod_sobel_image(imgIn); 46 | /* Finally save our processed image to the specified path */ 47 | sod_img_save_as_png(imgOut, zOut); 48 | /* Cleanup */ 49 | sod_free_image(imgIn); 50 | sod_free_image(imgOut); 51 | return 0; 52 | } -------------------------------------------------------------------------------- /samples/test.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/test.png -------------------------------------------------------------------------------- /samples/text.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/symisc/sod/443d764e71ab5ecd25308fbfba0775c2f1b85d99/samples/text.jpg -------------------------------------------------------------------------------- /samples/train.txt: -------------------------------------------------------------------------------- 1 | # Single class pedestrian detector for SOD RealNets. 2 | # 3 | # This file can serve as template for your future RealNets models to be generated by the SOD training 4 | # interfaces which are documented at https://sod.pixlab.io/api.html#realnet_train. 5 | # 6 | # Copyright (C) PixLab| Symisc Systems - All right reserved. legal@symisc.net - https://sod.pixlab.io 7 | 8 | # The first thing to specify is where the training samples are located. 9 | # You must group your dataset on the same directory so can SOD load each entry 10 | # on a single run and pass the collected image set to the RealNet trainer. 11 | 12 | [paths] 13 | 14 | # Mandatory positive samples path (i.e. the pedestrian dataset that may contains hundred or thousand of images) 15 | pos = /var/pedestrian_dataset/positives 16 | 17 | # Background samples path (i.e. various negative samples holding anything [car, trees, bus, cat, etc] except a pedestrian!! very important) 18 | neg = /var/pedestrian_dataset/background 19 | 20 | # Optional test sample path 21 | #test = /var/pedestrian_dataset/test 22 | 23 | # True to recurse (scan) subdirectories on the root path of your dataset (positives, background and test paths) 24 | recurse = true 25 | 26 | # Everything below is an optional field and does not require that you mess with it unless 27 | # you know what you doing (i.e. Tune your model) 28 | 29 | [detector] 30 | 31 | # min_tree_depth = 6 # Minimum tree depth 32 | 33 | # max_tree_depth = 12 # Maximum tree depth 34 | 35 | # max_trees = 2048 # Maximum decision tress to generate for this model 36 | 37 | # tpr = 0.9975 # Minimum True Positive Rate (TPR) which must be a float value set between 0.1 .. 1 38 | 39 | # fpr = 0.5 # Maximum False Positive Rate (FPR) which must be a float value set between 0.1 .. 1 40 | 41 | # data_augment = false # Introduce small perturbation to the input positive samples 42 | 43 | # target_fpr = 1e-6 # Target false positive rate (FPR) to achieve. 44 | # When we hit this value or max_trees whichever occurs first, training is stopped. 45 | 46 | # normalize = false # Normalize the training positive samples 47 | 48 | # Information about your model 49 | name = pedestrian 50 | about = RealNets pedestrian detector (single class) - Copyright (C) 2017 - 2018 Symisc Systems 51 | 52 | 53 | -------------------------------------------------------------------------------- /sod.h: -------------------------------------------------------------------------------- 1 | #ifndef _SOD_H_ 2 | #define _SOD_H_ 3 | /* 4 | * SOD - An Embedded Computer Vision & Machine Learning Library. 5 | * Copyright (C) 2018 - 2023 PixLab| Symisc Systems. https://sod.pixlab.io 6 | * Version 1.1.9 7 | * 8 | * Symisc Systems employs a dual licensing model that offers customers 9 | * a choice of either our open source license (GPLv3) or a commercial 10 | * license. 11 | * 12 | * For information on licensing, redistribution of the SOD library, and for a DISCLAIMER OF ALL WARRANTIES 13 | * please visit: 14 | * https://pixlab.io/sod 15 | * or contact: 16 | * licensing@symisc.net 17 | * support@pixlab.io 18 | */ 19 | /* 20 | * This file is part of Symisc SOD - Open Source Release (GPLv3) 21 | * 22 | * SOD is free software : you can redistribute it and/or modify 23 | * it under the terms of the GNU General Public License as published by 24 | * the Free Software Foundation, either version 3 of the License, or 25 | * (at your option) any later version. 26 | * 27 | * SOD is distributed in the hope that it will be useful, 28 | * but WITHOUT ANY WARRANTY; without even the implied warranty of 29 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.See the 30 | * GNU General Public License for more details. 31 | * 32 | * You should have received a copy of the GNU General Public License 33 | * along with SOD. If not, see . 34 | */ 35 | /* Make sure we can call this stuff from C++ */ 36 | #ifdef __cplusplus 37 | extern "C" { 38 | #endif 39 | /* 40 | * Marker for exported interfaces. 41 | */ 42 | #if defined (_MSC_VER) || defined (__MINGW32__) || defined (__GNUC__) && defined (__declspec) 43 | #define SOD_APIIMPORT __declspec(dllimport) 44 | #define SOD_APIEXPORT __declspec(dllexport) 45 | #else 46 | #define SOD_APIIMPORT 47 | #define SOD_APIEXPORT 48 | #endif 49 | /* 50 | * The SOD_VERSION C preprocessor macro evaluates to a string literal 51 | * that is the SOD version in the format "X.Y.Z" where X is the major 52 | * version number and Y is the minor version number and Z is the release 53 | * number. 54 | */ 55 | #define SOD_VERSION "1.1.9" 56 | /* 57 | * The SOD_VERSION_NUMBER C preprocessor macro resolves to an integer 58 | * with the value (X*1000000 + Y*1000 + Z) where X, Y, and Z are the same 59 | * numbers used in [SOD_VERSION]. 60 | */ 61 | #define SOD_VERSION_NUMBER 1001009 62 | /* 63 | * Forward declarations. 64 | */ 65 | /* 66 | * RealNets handle documented at https://sod.pixlab.io/api.html#sod_realnet. */ 67 | typedef struct sod_realnet sod_realnet; 68 | /* 69 | * Convolutional/Recurrent Neural Networks (CNN/RNN) handle documented at https://sod.pixlab.io/api.html#sod_cnn. */ 70 | typedef struct sod_cnn sod_cnn; 71 | /* 72 | * bounding box structure documented at https://sod.pixlab.io/api.html#sod_box. */ 73 | typedef struct sod_box sod_box; 74 | /* 75 | * Image/Frame container documented at https://sod.pixlab.io/api.html#sod_img. */ 76 | typedef struct sod_img sod_img; 77 | /* 78 | * Point instance documented at https://sod.pixlab.io/api.html#sod_pts. */ 79 | typedef struct sod_pts sod_pts; 80 | /* 81 | * RealNets model handle documented at https://sod.pixlab.io/api.html#sod_realnet_model_handle. */ 82 | typedef unsigned int sod_realnet_model_handle; 83 | #ifdef SOD_ENABLE_NET_TRAIN 84 | /* 85 | * RealNets trainer handle documented at https://sod.pixlab.io/api.html#sod_realnet_trainer. */ 86 | typedef struct sod_realnet_trainer sod_realnet_trainer; 87 | #endif 88 | /* 89 | * A bounding box or bbox for short is represented by an instance of the `sod_box` structure. 90 | * A sod_box instance always store the coordinates of a rectangle obtained from a prior successful 91 | * call to one of the object detection routines of a sod_cnn or sod_realnet handle such as 92 | * `sod_cnn_predict()` or from the connected component labeling interface `sod_image_find_blobs()`. 93 | * 94 | * Besides the rectangle coordinates. The `zName` and `score` fields member of this structure hold 95 | * useful information about the object it surround. 96 | * 97 | * This structure and related interfaces are documented at https://sod.pixlab.io/api.html#sod_box. 98 | */ 99 | struct sod_box { 100 | int x; /* The x-coordinate of the upper-left corner of the rectangle */ 101 | int y; /* The y-coordinate of the upper-left corner of the rectangle */ 102 | int w; /* Rectangle width */ 103 | int h; /* Rectangle height */ 104 | float score; /* Confidence threshold. */ 105 | const char *zName; /* Detected object name. I.e. person, face, dog, car, plane, cat, bicycle, etc. */ 106 | void *pUserData; /* External pointer used by some modules such as the face landmarks, NSFW classifier, pose estimator, etc. */ 107 | }; 108 | /* 109 | * Internally, each in-memory representation of an input image or video frame 110 | * is kept in an instance of the `sod_img` structure. Basically, a `sod_img` is just a record 111 | * of the width, height and number of color channels in an image, and also the pixel values 112 | * for every pixel. Images pixels are arranged in CHW format. This means in a 3 channel 113 | * image with width 400 and height 300, the first 400 values are the 1st row of the 1st channel 114 | * of the image. The second 400 pixels are the 2nd row. after 120,000 values we get to pixels 115 | * in the 2nd channel, and so forth. 116 | * 117 | * This structure and related interfaces are documented at https://sod.pixlab.io/api.html#sod_img. 118 | */ 119 | struct sod_img { 120 | int h; /* Image/frame height */ 121 | int w; /* Image/frame width */ 122 | int c; /* Image depth/Total number of color channels e.g. 1 for grayscale images, 3 RGB, etc. */ 123 | union 124 | { 125 | float* data; /* Blob */ 126 | unsigned char* zdata; 127 | }; 128 | }; 129 | /* 130 | * An instance of the `sod_pts` structure describe a 2D point in space with integer coordinates 131 | * (usually zero-based). This structure is rarely manipulated by SOD and is used mostly by 132 | * the Hough line detection interface `sod_hough_lines_detect()` and line drawing routine `sod_image_draw_line()`. 133 | * 134 | * This structure and related interfaces are documented at https://sod.pixlab.io/api.html#sod_pts. 135 | */ 136 | struct sod_pts { 137 | int x; /* The x-coordinate, in logical units of the point offset. */ 138 | int y; /* The y-coordinate, in logical units of the point offset. */ 139 | }; 140 | /* 141 | * An integer configuration option that determines what property of a `sod_cnn` handle is to be configured. 142 | * Subsequent arguments vary depending on the configuration verb. 143 | * 144 | * The documentation (including expected arguments for each configuration verb) is available to consult 145 | * at https://sod.pixlab.io/c_api/sod_cnn_config.html. 146 | */ 147 | typedef enum { 148 | SOD_CNN_NETWORK_OUTPUT = 1, 149 | SOD_CNN_DETECTION_THRESHOLD, 150 | SOD_CNN_NMS, 151 | SOD_CNN_DETECTION_CLASSES, 152 | SOD_CNN_RAND_SEED, 153 | SOD_CNN_HIER_THRESHOLD, 154 | SOD_CNN_TEMPERATURE, 155 | SOD_CNN_LOG_CALLBACK, 156 | SOD_RNN_CALLBACK, 157 | SOD_RNN_TEXT_LENGTH, 158 | SOD_RNN_DATA_LENGTH, 159 | SOD_RNN_SEED 160 | }SOD_CNN_CONFIG; 161 | /* 162 | * RNN Consumer callback to be used in conjunction with the `SOD_RNN_CALLBACK` configuration verb. 163 | * 164 | * The documentation is available to consult at https://sod.pixlab.io/c_api/sod_cnn_config.html. 165 | */ 166 | typedef void (*ProcRnnCallback)(const char *, size_t, void *); 167 | /* 168 | * Log Consumer callback to be used in conjunction with the `SOD_CNN_LOG_CALLBACK` or 169 | * the `SOD_REALNET_TR_LOG_CALLBACK` configuration verb. 170 | * 171 | * The documentation is available to consult at https://sod.pixlab.io/c_api/sod_cnn_config.html. 172 | */ 173 | typedef void(*ProcLogCallback)(const char *, size_t, void *); 174 | /* 175 | * Macros to be used in conjunction with the `sod_img_load_from_file()` or `sod_img_load_from_mem()` interfaces. 176 | */ 177 | #define SOD_IMG_COLOR 0 /* Load full color channels. */ 178 | #define SOD_IMG_GRAYSCALE 1 /* Load an image in the grayscale colorpsace only (single channel). */ 179 | /* 180 | * Macros around a stack allocated `sod_img` instance. 181 | */ 182 | #define SOD_IMG_2_INPUT(IMG) (IMG.data) /* Pointer to raw binary contents (blobs) of an image or frame. */ 183 | #define SOD_IS_EMPTY_IMG(IMG) (!IMG.data) /* NIL pointer test (marker for an empty or broken image format). */ 184 | /* 185 | * Possible return value from each exported SOD interface defined below. 186 | */ 187 | #define SOD_OK 0 /* Everything went well */ 188 | #define SOD_UNSUPPORTED -1 /* Unsupported Pixel format */ 189 | #define SOD_OUTOFMEM -4 /* Out-of-Memory */ 190 | #define SOD_ABORT -5 /* User callback request an operation abort */ 191 | #define SOD_IOERR -6 /* IO error */ 192 | #define SOD_LIMIT -7 /* Limit reached */ 193 | /* 194 | * An integer configuration option that determines what property of a `sod_realnet_trainer` handle is to be configured. 195 | * Subsequent arguments vary depending on the configuration verb. 196 | * 197 | * The documentation (including expected arguments for each configuration verb) is available to consult 198 | * at https://sod.pixlab.io/c_api/sod_realnet_train_config.html. 199 | */ 200 | #ifdef SOD_ENABLE_NET_TRAIN 201 | typedef enum { 202 | SOD_REALNET_TR_LOG_CALLBACK = 1, 203 | SOD_REALNET_TR_OUTPUT_MODEL 204 | }SOD_REALNET_TRAINER_CONFIG; 205 | #endif /* SOD_ENABLE_NET_TRAIN */ 206 | /* 207 | * An integer configuration option that determines what property of a `sod_realnet` handle is to be configured. 208 | * Subsequent arguments vary depending on the configuration verb. 209 | * 210 | * The documentation (including expected arguments for each configuration verb) is available to consult 211 | * at https://sod.pixlab.io/c_api/sod_realnet_model_config.html. 212 | */ 213 | typedef enum { 214 | SOD_REALNET_MODEL_MINSIZE = 1, 215 | SOD_REALNET_MODEL_MAXSIZE, 216 | SOD_REALNET_MODEL_SCALEFACTOR, 217 | SOD_REALNET_MODEL_STRIDEFACTOR, 218 | SOD_RELANET_MODEL_DETECTION_THRESHOLD, 219 | SOD_REALNET_MODEL_NMS, 220 | SOD_REALNET_MODEL_DISCARD_NULL_BOXES, 221 | SOD_REALNET_MODEL_NAME, 222 | SOD_REALNET_MODEL_ABOUT_INFO 223 | }SOD_REALNET_MODEL_CONFIG; 224 | /* 225 | * SOD Embedded C/C++ API. 226 | * 227 | * The API documentation is available to consult at https://sod.pixlab.io/api.html. 228 | * The introduction course is available to consult at https://sod.pixlab.io/intro.html. 229 | */ 230 | #ifndef SOD_DISABLE_CNN 231 | /* 232 | * Convolutional/Recurrent Neural Networks (CNN/RNN) API. 233 | * 234 | * The interfaces are documented at https://sod.pixlab.io/api.html#cnn. 235 | */ 236 | SOD_APIEXPORT int sod_cnn_create(sod_cnn **ppOut, const char *zArch, const char *zModelPath, const char **pzErr); 237 | SOD_APIEXPORT int sod_cnn_config(sod_cnn *pNet, SOD_CNN_CONFIG conf, ...); 238 | SOD_APIEXPORT int sod_cnn_predict(sod_cnn *pNet, float *pInput, sod_box **paBox, int *pnBox); 239 | SOD_APIEXPORT void sod_cnn_destroy(sod_cnn *pNet); 240 | SOD_APIEXPORT float * sod_cnn_prepare_image(sod_cnn *pNet, sod_img in); 241 | SOD_APIEXPORT int sod_cnn_get_network_size(sod_cnn *pNet, int *pWidth, int *pHeight, int *pChannels); 242 | #endif /* SOD_DISABLE_CNN */ 243 | #ifndef SOD_DISABLE_REALNET 244 | /* 245 | * RealNets API. 246 | * 247 | * The interfaces are documented at https://sod.pixlab.io/api.html#realnet. 248 | */ 249 | SOD_APIEXPORT int sod_realnet_create(sod_realnet **ppOut); 250 | SOD_APIEXPORT int sod_realnet_load_model_from_mem(sod_realnet *pNet, const void * pModel, unsigned int nBytes, sod_realnet_model_handle *pOutHandle); 251 | #ifdef SOD_NO_MMAP 252 | SOD_APIEXPORT int sod_realnet_load_model_from_disk(sod_realnet *pNet, const char * zPath, sod_realnet_model_handle *pOutHandle); 253 | #endif 254 | SOD_APIEXPORT int sod_realnet_model_config(sod_realnet *pNet, sod_realnet_model_handle handle, SOD_REALNET_MODEL_CONFIG conf, ...); 255 | SOD_APIEXPORT int sod_realnet_detect(sod_realnet *pNet, const unsigned char *zGrayImg, int width, int height, sod_box **apBox, int *pnBox); 256 | SOD_APIEXPORT void sod_realnet_destroy(sod_realnet *pNet); 257 | #endif /* SOD_DISABLE_REALNET */ 258 | #ifdef SOD_ENABLE_NET_TRAIN 259 | /* 260 | * RealNets Training API. 261 | * 262 | * The interfaces are documented at https://sod.pixlab.io/api.html#realnet_train. 263 | */ 264 | SOD_APIEXPORT int sod_realnet_train_init(sod_realnet_trainer **ppOut); 265 | SOD_APIEXPORT int sod_realnet_train_config(sod_realnet_trainer *pTrainer, SOD_REALNET_TRAINER_CONFIG op, ...); 266 | SOD_APIEXPORT int sod_realnet_train_start(sod_realnet_trainer *pTrainer, const char *zConf); 267 | SOD_APIEXPORT void sod_realnet_train_release(sod_realnet_trainer *pTrainer); 268 | #endif /* SOD_ENABLE_NET_TRAIN */ 269 | /* 270 | * Image Processing API. 271 | * 272 | * The interfaces are documented at https://sod.pixlab.io/api.html#imgproc. 273 | */ 274 | SOD_APIEXPORT sod_img sod_make_empty_image(int w, int h, int c); 275 | SOD_APIEXPORT sod_img sod_make_image(int w, int h, int c); 276 | SOD_APIEXPORT int sod_grow_image(sod_img *pImg,int w, int h, int c); 277 | SOD_APIEXPORT sod_img sod_make_random_image(int w, int h, int c); 278 | SOD_APIEXPORT sod_img sod_copy_image(sod_img m); 279 | SOD_APIEXPORT void sod_free_image(sod_img m); 280 | 281 | #ifndef SOD_DISABLE_IMG_READER 282 | SOD_APIEXPORT sod_img sod_img_load_from_file(const char *zFile, int nChannels); 283 | SOD_APIEXPORT sod_img sod_img_load_from_mem(const unsigned char *zBuf, int buf_len, int nChannels); 284 | SOD_APIEXPORT int sod_img_set_load_from_directory(const char *zPath, sod_img ** apLoaded, int * pnLoaded, int max_entries); 285 | SOD_APIEXPORT void sod_img_set_release(sod_img *aLoaded, int nEntries); 286 | #ifndef SOD_DISABLE_IMG_WRITER 287 | SOD_APIEXPORT int sod_img_save_as_png(sod_img input, const char *zPath); 288 | SOD_APIEXPORT int sod_img_save_as_jpeg(sod_img input, const char *zPath, int Quality); 289 | SOD_APIEXPORT int sod_img_blob_save_as_png(const char * zPath, const unsigned char *zBlob, int width, int height, int nChannels); 290 | SOD_APIEXPORT int sod_img_blob_save_as_jpeg(const char * zPath, const unsigned char *zBlob, int width, int height, int nChannels, int Quality); 291 | SOD_APIEXPORT int sod_img_blob_save_as_bmp(const char * zPath, const unsigned char *zBlob, int width, int height, int nChannels); 292 | #endif /* SOD_DISABLE_IMG_WRITER */ 293 | #define sod_img_load_color(zPath) sod_img_load_from_file(zPath, SOD_IMG_COLOR) 294 | #define sod_img_load_grayscale(zPath) sod_img_load_from_file(zPath, SOD_IMG_GRAYSCALE) 295 | #endif /* SOD_DISABLE_IMG_READER */ 296 | 297 | SOD_APIEXPORT float sod_img_get_pixel(sod_img m, int x, int y, int c); 298 | SOD_APIEXPORT void sod_img_set_pixel(sod_img m, int x, int y, int c, float val); 299 | SOD_APIEXPORT void sod_img_add_pixel(sod_img m, int x, int y, int c, float val); 300 | SOD_APIEXPORT sod_img sod_img_get_layer(sod_img m, int l); 301 | 302 | SOD_APIEXPORT void sod_img_rgb_to_hsv(sod_img im); 303 | SOD_APIEXPORT void sod_img_hsv_to_rgb(sod_img im); 304 | SOD_APIEXPORT void sod_img_rgb_to_bgr(sod_img im); 305 | SOD_APIEXPORT void sod_img_bgr_to_rgb(sod_img im); 306 | SOD_APIEXPORT void sod_img_yuv_to_rgb(sod_img im); 307 | SOD_APIEXPORT void sod_img_rgb_to_yuv(sod_img im); 308 | 309 | /* Introduced in version 1.1.9 */ 310 | SOD_APIEXPORT void sod_constrain_image(sod_img im); 311 | SOD_APIEXPORT sod_img sod_img_mask_to_rgb(sod_img mask); 312 | SOD_APIEXPORT void sod_censor_image(sod_img im, int dx, int dy, int w, int h); 313 | SOD_APIEXPORT void sod_saturate_image(sod_img im, float sat); 314 | SOD_APIEXPORT void sod_saturate_exposure_image(sod_img im, float sat, float exposure); 315 | SOD_APIEXPORT void sod_hue_image(sod_img im, float hue); 316 | SOD_APIEXPORT void sod_exposure_image(sod_img im, float sat); 317 | SOD_APIEXPORT void sod_distort_image(sod_img im, float hue, float sat, float val); 318 | SOD_APIEXPORT void sod_random_distort_image(sod_img im, float hue, float saturation, float exposure); 319 | SOD_APIEXPORT sod_img sod_box_blur_image(sod_img im); 320 | SOD_APIEXPORT void sod_image_sepia_filter(sod_img rgb); 321 | SOD_APIEXPORT sod_img sod_gaussian_blur_image(sod_img im, int radius, double sigma); 322 | 323 | SOD_APIEXPORT sod_img sod_minutiae(sod_img bin, int *pTotal, int *pEp, int *pBp); 324 | SOD_APIEXPORT sod_img sod_gaussian_noise_reduce(sod_img grayscale); 325 | SOD_APIEXPORT sod_img sod_equalize_histogram(sod_img im); 326 | 327 | SOD_APIEXPORT sod_img sod_grayscale_image(sod_img im); 328 | SOD_APIEXPORT void sod_grayscale_image_3c(sod_img im); 329 | 330 | SOD_APIEXPORT sod_img sod_threshold_image(sod_img im, float thresh); 331 | SOD_APIEXPORT sod_img sod_otsu_binarize_image(sod_img im); 332 | SOD_APIEXPORT sod_img sod_binarize_image(sod_img im, int reverse); 333 | SOD_APIEXPORT sod_img sod_dilate_image(sod_img im, int times); 334 | SOD_APIEXPORT sod_img sod_erode_image(sod_img im, int times); 335 | 336 | SOD_APIEXPORT sod_img sod_sharpen_filtering_image(sod_img im); 337 | SOD_APIEXPORT sod_img sod_hilditch_thin_image(sod_img im); 338 | 339 | SOD_APIEXPORT sod_img sod_sobel_image(sod_img im); 340 | SOD_APIEXPORT sod_img sod_canny_edge_image(sod_img im, int reduce_noise); 341 | 342 | SOD_APIEXPORT sod_pts * sod_hough_lines_detect(sod_img im, int threshold, int *nPts); 343 | SOD_APIEXPORT void sod_hough_lines_release(sod_pts *pLines); 344 | 345 | SOD_APIEXPORT int sod_image_find_blobs(sod_img im, sod_box **paBox, int *pnBox, int(*xFilter)(int width, int height)); 346 | SOD_APIEXPORT void sod_image_blob_boxes_release(sod_box *pBox); 347 | 348 | SOD_APIEXPORT void sod_composite_image(sod_img source, sod_img dest, int dx, int dy); 349 | SOD_APIEXPORT void sod_flip_image(sod_img input); 350 | SOD_APIEXPORT sod_img sod_image_distance(sod_img a, sod_img b); 351 | 352 | SOD_APIEXPORT void sod_embed_image(sod_img source, sod_img dest, int dx, int dy); 353 | SOD_APIEXPORT sod_img sod_blend_image(sod_img fore, sod_img back, float alpha); 354 | SOD_APIEXPORT void sod_scale_image_channel(sod_img im, int c, float v); 355 | SOD_APIEXPORT void sod_translate_image_channel(sod_img im, int c, float v); 356 | 357 | SOD_APIEXPORT sod_img sod_resize_image(sod_img im, int w, int h); 358 | SOD_APIEXPORT sod_img sod_resize_max(sod_img im, int max); 359 | SOD_APIEXPORT sod_img sod_resize_min(sod_img im, int min); 360 | SOD_APIEXPORT sod_img sod_rotate_crop_image(sod_img im, float rad, float s, int w, int h, float dx, float dy, float aspect); 361 | SOD_APIEXPORT sod_img sod_rotate_image(sod_img im, float rad); 362 | 363 | SOD_APIEXPORT void sod_translate_image(sod_img m, float s); 364 | SOD_APIEXPORT void sod_scale_image(sod_img m, float s); 365 | SOD_APIEXPORT void sod_normalize_image(sod_img p); 366 | SOD_APIEXPORT void sod_transpose_image(sod_img im); 367 | 368 | SOD_APIEXPORT sod_img sod_crop_image(sod_img im, int dx, int dy, int w, int h); 369 | SOD_APIEXPORT sod_img sod_random_crop_image(sod_img im, int w, int h); 370 | SOD_APIEXPORT sod_img sod_random_augment_image(sod_img im, float angle, float aspect, int low, int high, int size); 371 | 372 | SOD_APIEXPORT void sod_image_draw_box(sod_img im, int x1, int y1, int x2, int y2, float r, float g, float b); 373 | SOD_APIEXPORT void sod_image_draw_box_grayscale(sod_img im, int x1, int y1, int x2, int y2, float g); 374 | SOD_APIEXPORT void sod_image_draw_circle(sod_img im, int x0, int y0, int radius, float r, float g, float b); 375 | SOD_APIEXPORT void sod_image_draw_circle_thickness(sod_img im, int x0, int y0, int radius, int width, float r, float g, float b); 376 | SOD_APIEXPORT void sod_image_draw_bbox(sod_img im, sod_box bbox, float r, float g, float b); 377 | SOD_APIEXPORT void sod_image_draw_bbox_width(sod_img im, sod_box bbox, int width, float r, float g, float b); 378 | SOD_APIEXPORT void sod_image_draw_line(sod_img im, sod_pts start, sod_pts end, float r, float g, float b); 379 | 380 | SOD_APIEXPORT unsigned char * sod_image_to_blob(sod_img im); 381 | SOD_APIEXPORT void sod_image_free_blob(unsigned char *zBlob); 382 | /* 383 | * OpenCV Integration API. The library must be compiled against OpenCV 384 | * with the compile-time directive SOD_ENABLE_OPENCV defined. 385 | * 386 | * The documentation is available to consult at https://sod.pixlab.io/api.html#cmpl_opencv. 387 | * The interfaces are documented at https://sod.pixlab.io/api.html#cvinter. 388 | */ 389 | #ifdef SOD_ENABLE_OPENCV 390 | /* 391 | * Change the include paths to the directory where OpenCV reside 392 | * if those headers are not found by your compiler. 393 | */ 394 | #include 395 | #include 396 | 397 | SOD_APIEXPORT sod_img sod_img_load_cv_ipl(IplImage* src); 398 | SOD_APIEXPORT sod_img sod_img_load_from_cv(const char *filename, int channels); 399 | SOD_APIEXPORT sod_img sod_img_load_from_cv_stream(CvCapture *cap); 400 | SOD_APIEXPORT int sod_img_fill_from_cv_stream(CvCapture *cap, sod_img *pImg); 401 | SOD_APIEXPORT void sod_img_save_to_cv_jpg(sod_img im, const char *zPath); 402 | #endif /* SOD_ENABLE_OPENCV */ 403 | /* 404 | * SOD Embedded Release Information & Copyright Notice. 405 | */ 406 | SOD_APIEXPORT const char * sod_lib_copyright(void); 407 | #define SOD_LIB_INFO "SOD Embedded - Release 1.1.9 under GPLv3/Commercial Licensing. Copyright (C) 2018 - 2023 PixLab| Symisc Systems, https://sod.pixlab.io" 408 | #ifdef __cplusplus 409 | } 410 | #endif 411 | #endif /* _SOD_H_ */ 412 | --------------------------------------------------------------------------------