├── requirements.txt ├── in_img ├── bing304687-0.jpg ├── bing304687-1.jpg ├── bing304687-2.jpg ├── bing304687-3.jpg └── bing304687-4.jpg ├── out_img ├── web │ ├── bing304687-0.jpg │ ├── bing304687-1.jpg │ ├── bing304687-2.jpg │ ├── bing304687-3.jpg │ └── bing304687-4.jpg ├── onnx-cpu │ ├── bing304687-0.jpg │ ├── bing304687-1.jpg │ ├── bing304687-2.jpg │ ├── bing304687-3.jpg │ └── bing304687-4.jpg └── web_while │ ├── bing304687-0.jpg │ ├── bing304687-1.jpg │ ├── bing304687-2.jpg │ ├── bing304687-3.jpg │ └── bing304687-4.jpg ├── .idea ├── .gitignore ├── misc.xml ├── vcs.xml ├── inspectionProfiles │ ├── profiles_settings.xml │ └── Project_Default.xml ├── modules.xml └── Image_matting.iml ├── tokens.txt ├── main.py └── README.md /requirements.txt: -------------------------------------------------------------------------------- 1 | rembg==2.0.30 2 | requests==2.28.1 3 | -------------------------------------------------------------------------------- /in_img/bing304687-0.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/in_img/bing304687-0.jpg -------------------------------------------------------------------------------- /in_img/bing304687-1.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/in_img/bing304687-1.jpg -------------------------------------------------------------------------------- /in_img/bing304687-2.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/in_img/bing304687-2.jpg -------------------------------------------------------------------------------- /in_img/bing304687-3.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/in_img/bing304687-3.jpg -------------------------------------------------------------------------------- /in_img/bing304687-4.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/in_img/bing304687-4.jpg -------------------------------------------------------------------------------- /out_img/web/bing304687-0.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web/bing304687-0.jpg -------------------------------------------------------------------------------- /out_img/web/bing304687-1.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web/bing304687-1.jpg -------------------------------------------------------------------------------- /out_img/web/bing304687-2.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web/bing304687-2.jpg -------------------------------------------------------------------------------- /out_img/web/bing304687-3.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web/bing304687-3.jpg -------------------------------------------------------------------------------- /out_img/web/bing304687-4.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web/bing304687-4.jpg -------------------------------------------------------------------------------- /out_img/onnx-cpu/bing304687-0.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/onnx-cpu/bing304687-0.jpg -------------------------------------------------------------------------------- /out_img/onnx-cpu/bing304687-1.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/onnx-cpu/bing304687-1.jpg -------------------------------------------------------------------------------- /out_img/onnx-cpu/bing304687-2.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/onnx-cpu/bing304687-2.jpg -------------------------------------------------------------------------------- /out_img/onnx-cpu/bing304687-3.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/onnx-cpu/bing304687-3.jpg -------------------------------------------------------------------------------- /out_img/onnx-cpu/bing304687-4.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/onnx-cpu/bing304687-4.jpg -------------------------------------------------------------------------------- /out_img/web_while/bing304687-0.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web_while/bing304687-0.jpg -------------------------------------------------------------------------------- /out_img/web_while/bing304687-1.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web_while/bing304687-1.jpg -------------------------------------------------------------------------------- /out_img/web_while/bing304687-2.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web_while/bing304687-2.jpg -------------------------------------------------------------------------------- /out_img/web_while/bing304687-3.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web_while/bing304687-3.jpg -------------------------------------------------------------------------------- /out_img/web_while/bing304687-4.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Scipline/Image_matting/HEAD/out_img/web_while/bing304687-4.jpg -------------------------------------------------------------------------------- /.idea/.gitignore: -------------------------------------------------------------------------------- 1 | # 默认忽略的文件 2 | /shelf/ 3 | /workspace.xml 4 | # 基于编辑器的 HTTP 客户端请求 5 | /httpRequests/ 6 | # Datasource local storage ignored files 7 | /dataSources/ 8 | /dataSources.local.xml 9 | -------------------------------------------------------------------------------- /.idea/misc.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | -------------------------------------------------------------------------------- /.idea/vcs.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | -------------------------------------------------------------------------------- /.idea/inspectionProfiles/profiles_settings.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 6 | -------------------------------------------------------------------------------- /.idea/modules.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | -------------------------------------------------------------------------------- /.idea/Image_matting.iml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | -------------------------------------------------------------------------------- /.idea/inspectionProfiles/Project_Default.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 20 | -------------------------------------------------------------------------------- /tokens.txt: -------------------------------------------------------------------------------- 1 | neVfUwfrBYwGVUU8LjFtVayd 2 | HsWgUwJJJNvjVjmYMzDAfM3S 3 | zRJjjAdSmhaNF2JoiUCwtV1s 4 | mwQdb5oDNHZRbnGUTUPSJJeS 5 | c4XtEPVwUwGYJBZrXPNRkHSH 6 | LKvvZDGYwcakJkFR6HHkf2J7 7 | 8jSsx7iWVvwy4M49CMTYNwfX 8 | LiM4rfc1rr8FdWqikronMbdm 9 | W2Vz6osKSR4z2o7V7qVHFJzg 10 | 6oQcwGozeFyPS7NZPiRAKdw2 11 | 1psMbURiL3nCBrnsWMhHmkNJ 12 | 2URAxSUqCiFg2F8y8E29xbFb 13 | 7rwRz1ACanst1MLg1P1vNRHE 14 | L245rgUojVP9Nvo3nrsGTg3i 15 | kPFsVb8ajqaeegnLRwHQg5wg 16 | pxvvF152FPs2G47reAjdSQFv 17 | dMvFs2L8PWTjcMjbDo6jiQDb 18 | ggY6xY9yGE2frQAGst7M7FtS 19 | KhJUHr65KhauFvkRrUvXEUYm 20 | Ue1rcGF6pG1Fbnuq83a8jX6U 21 | YJBUt3pTq9A1G22553Kpcm95 22 | 3Jh2rJSFNHk4Fg31zYAaYHnY 23 | cn5C3F8xb628aKKX2DEQJs5q 24 | FCcqMzi9w95p499DNifnyqtc 25 | k6niQBfpxXsD1EfyQ8AagPRx 26 | eSt8w3vwT41B1KSMcAZ88dbp 27 | FPRKfddyw6QtX2GAzQHfGXzY 28 | EQrFePowinNS5Q7TeDnRFLse 29 | o6UwnQ9nfDgV8wRkjRtpiucW 30 | M8iCWbRVu71HTzJsox87tETP 31 | Bo6ndWzUR5voq4gEbWKQkK8h 32 | MwV1R4Ry1n5BsbLhFYcL6127 33 | iR6MQTFF71D4LdipGpPNZyYb 34 | Ngkqn7ZYXgot5rEMVM9h1Z9m 35 | EmuSuJ5mxS94oZzdQba8Nxuy -------------------------------------------------------------------------------- /main.py: -------------------------------------------------------------------------------- 1 | # !/usr/bin/env python3 2 | # -*- coding: UTF-8 -*- 3 | # 模型文件：https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2net.onnx 放在”C:\Users\Administrator\.u2net\u2net.onnx: 4 | # 官方文档：https://github.com/danielgatis/rembg 5 | """ 6 | apikey面板：https://www.remove.bg/dashboard#api-key 7 | 临时邮箱：https://www.linshiyouxiang.net 可用域名：@bestdefinitions.com 8 | 9 | 安装cv2，pip install opencv-contrib-python==4.5.5.64 10 | 适用于:python3.7 python3.8 python3.9 解释器 11 | """ 12 | from pathlib import Path 13 | 14 | import requests 15 | from rembg import remove, new_session 16 | 17 | 18 | def api_key(): 19 | f = open('tokens.txt', mode="r", encoding="utf-8") 20 | f.seek(0) 21 | token = f.readline().strip() 22 | f.seek(0) 23 | tokens = f.read() 24 | f.seek(0) 25 | f.close() 26 | f = open('tokens.txt', mode="w", encoding="utf-8") 27 | tokens = tokens.replace(token, "").strip() + "\n" + token 28 | f.write(tokens) 29 | f.close() 30 | return token 31 | 32 | 33 | def main(rebg_type): 34 | # color Ex:red,green / hex:81d4fa ? " 35 | bgcolor = "white" 36 | # bgcolor = "" # 留空为空白背景 37 | api = api_key() 38 | if not Path("out_img").exists(): 39 | Path("out_img").mkdir() 40 | p = Path('in_img') 41 | # 获取in_img目录下所有.png,.jpg文件 42 | for file in p.glob('*.[pj][np]g'): 43 | input_path = str(file) 44 | output_path = str(p.parent / ("out_img/" + file.name)) 45 | if Path(output_path).exists(): 46 | continue 47 | with open(input_path, 'rb') as i: 48 | with open(output_path, 'wb') as o: 49 | photo_data = i.read() 50 | if rebg_type == "web": 51 | output = rebg_web(photo_data, api, bgcolor) 52 | while not output: 53 | api = api_key() 54 | output = rebg_web(photo_data, api, bgcolor) 55 | if output == 400: 56 | output = rebg_cpu(photo_data) 57 | elif rebg_type == "onnx": 58 | output = rebg_cpu(photo_data) 59 | o.write(output) 60 | 61 | 62 | def rebg_web(photo_data, api, bgcolor=None): 63 | """ 64 | 借助https://www.remove.bg网站免费用户每个月50额度处理 65 | :param photo_data: 66 | :param api: 67 | :param bgcolor: 68 | :return: 69 | """ 70 | # 官方API教程：https://www.remove.bg/api#sample-code 71 | response = requests.post( 72 | 'https://api.remove.bg/v1.0/removebg', 73 | files={'image_file': photo_data}, 74 | data={'size': 'auto', 'bg_color': bgcolor}, 75 | headers={'X-Api-Key': api}) 76 | if response.status_code == 200: 77 | return response.content 78 | elif response.status_code == 400: 79 | print("抠图失败，Could not identify foreground in image") 80 | return 400 81 | elif response.status_code == 402: 82 | print("额度已用尽，Insufficient credits") 83 | elif response.status_code == 403: 84 | print("Api Key出错，Api Key fail") 85 | print(api) 86 | elif response.status_code == 429: 87 | print("超过速度限制，Rate limit exceeded") 88 | else: 89 | print("未知错误") 90 | print(response.status_code, response.text) 91 | 92 | 93 | def rebg_cpu(photo_data): 94 | # 选择模型 95 | session = new_session(model_name="u2net") 96 | return remove(photo_data, session=session) 97 | 98 | 99 | if __name__ == "__main__": 100 | main(rebg_type="web") 101 | # main(rebg_type="onnx") 102 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | ## 使用说明： 2 | ``` 3 | 任选其一 4 | main(rebg_type="web") # 更精准，低清晰度，需要网络 5 | main(rebg_type="onnx") # 更快速，更清晰，需要性能 6 | ``` 7 | 在线apikey，[remove.bg网站](https://www.remove.bg/dashboard#api-key)。[临时邮箱](https://www.linshiyouxiang.net) ,可用域名：@bestdefinitions.com 8 | 9 | 思路来源：https://github.com/huanghaozi/AutoMatting (在此基础上，批量获取，自动切换apikey解决频繁，错误，额度等问题，自动换背景图，规范导入导出文件名，不足之处就是没有大佬的GUI) 10 | 11 | - 必要依赖：`pip install -r requiremens.txt` 12 | 13 | - 可选依赖： 14 | 15 | 安装**cv2**，`pip install opencv-contrib-python==4.5.5.64`。适用于:python3.7 python3.8 python3.9 解释器 16 | 17 | 安装**Pillow**，`pip install Pillow==9.3.0`。适配rembg==2.0.30 18 | 19 | 20 | 以下内容为[rembg](https://github.com/danielgatis/rembg)库官方文档，仅用于存档。 21 | 22 | --- 23 | 24 | ## Rembg 25 | 26 | [![Downloads](https://pepy.tech/badge/rembg)](https://pepy.tech/project/rembg) 27 | [![Downloads](https://pepy.tech/badge/rembg/month)](https://pepy.tech/project/rembg/month) 28 | [![Downloads](https://pepy.tech/badge/rembg/week)](https://pepy.tech/project/rembg/week) 29 | [![License](https://img.shields.io/badge/License-MIT-blue.svg)](https://img.shields.io/badge/License-MIT-blue.svg) 30 | [![Hugging Face Spaces](https://img.shields.io/badge/🤗%20Hugging%20Face-Spaces-blue)](https://huggingface.co/spaces/KenjieDec/RemBG) 31 | [![Streamlit App](https://img.shields.io/badge/🎈%20Streamlit%20Community-Cloud-blue)](https://bgremoval.streamlit.app/) 32 | 33 | 34 | Rembg is a tool to remove images background. 35 | 36 |

37 | 38 | 39 | 40 | 41 | 42 | 43 |

44 | 45 |

46 | 47 | 48 | 49 | 50 | 51 | 52 |

53 | 54 |

55 | 56 | 57 | 58 | 59 | 60 | 61 |

62 | 63 | **If this project has helped you, please consider making a [donation](https://www.buymeacoffee.com/danielgatis).** 64 | 65 | ## Requirements 66 | 67 | ``` 68 | python: >3.7, <3.11 69 | ``` 70 | 71 | ## Installation 72 | 73 | CPU support: 74 | 75 | ```bash 76 | pip install rembg 77 | ``` 78 | 79 | GPU support: 80 | 81 | ```bash 82 | pip install rembg[gpu] 83 | ``` 84 | 85 | ## Usage as a cli 86 | 87 | After the installation step you can use rembg just typing `rembg` in your terminal window. 88 | 89 | The `rembg` command has 3 subcommands, one for each input type: 90 | - `i` for files 91 | - `p` for folders 92 | - `s` for http server 93 | 94 | You can get help about the main command using: 95 | 96 | ``` 97 | rembg --help 98 | ``` 99 | 100 | As well, about all the subcommands using: 101 | 102 | ``` 103 | rembg --help 104 | ``` 105 | 106 | ### rembg `i` 107 | 108 | Used when input and output are files. 109 | 110 | Remove the background from a remote image 111 | 112 | ``` 113 | curl -s http://input.png | rembg i > output.png 114 | ``` 115 | 116 | Remove the background from a local file 117 | 118 | ``` 119 | rembg i path/to/input.png path/to/output.png 120 | ``` 121 | 122 | Remove the background specifying a model 123 | 124 | ``` 125 | rembg i -m u2netp path/to/input.png path/to/output.png 126 | ``` 127 | 128 | Remove the background returning only the mask 129 | 130 | ``` 131 | rembg i -om path/to/input.png path/to/output.png 132 | ``` 133 | 134 | 135 | Remove the background applying an alpha matting 136 | 137 | ``` 138 | rembg i -a path/to/input.png path/to/output.png 139 | ``` 140 | 141 | ### rembg `p` 142 | 143 | Used when input and output are folders. 144 | 145 | Remove the background from all images in a folder 146 | 147 | ``` 148 | rembg p path/to/input path/to/output 149 | ``` 150 | 151 | Same as before, but watching for new/changed files to process 152 | 153 | ``` 154 | rembg p -w path/to/input path/to/output 155 | ``` 156 | 157 | ### rembg `s` 158 | 159 | Used to start http server. 160 | 161 | To see the complete endpoints documentation, go to: `http://localhost:5000/docs`. 162 | 163 | Remove the background from an image url 164 | 165 | ``` 166 | curl -s "http://localhost:5000/?url=http://input.png" -o output.png 167 | ``` 168 | 169 | Remove the background from an uploaded image 170 | 171 | ``` 172 | curl -s -F file=@/path/to/input.jpg "http://localhost:5000" -o output.png 173 | ``` 174 | 175 | ## Usage as a library 176 | 177 | Input and output as bytes 178 | 179 | ```python 180 | from rembg import remove 181 | 182 | input_path = 'input.png' 183 | output_path = 'output.png' 184 | 185 | with open(input_path, 'rb') as i: 186 | with open(output_path, 'wb') as o: 187 | input = i.read() 188 | output = remove(input) 189 | o.write(output) 190 | ``` 191 | 192 | Input and output as a PIL image 193 | 194 | ```python 195 | from rembg import remove 196 | from PIL import Image 197 | 198 | input_path = 'input.png' 199 | output_path = 'output.png' 200 | 201 | input = Image.open(input_path) 202 | output = remove(input) 203 | output.save(output_path) 204 | ``` 205 | 206 | Input and output as a numpy array 207 | 208 | ```python 209 | from rembg import remove 210 | import cv2 211 | 212 | input_path = 'input.png' 213 | output_path = 'output.png' 214 | 215 | input = cv2.imread(input_path) 216 | output = remove(input) 217 | cv2.imwrite(output_path, output) 218 | ``` 219 | 220 | How to iterate over files in a performatic way 221 | 222 | ```python 223 | from pathlib import Path 224 | from rembg import remove, new_session 225 | 226 | session = new_session() 227 | 228 | for file in Path('path/to/folder').glob('*.png'): 229 | input_path = str(file) 230 | output_path = str(file.parent / (file.stem + ".out.png")) 231 | 232 | with open(input_path, 'rb') as i: 233 | with open(output_path, 'wb') as o: 234 | input = i.read() 235 | output = remove(input, session=session) 236 | o.write(output) 237 | ``` 238 | 239 | ## Usage as a docker 240 | 241 | Just replace the `rembg` command for `docker run danielgatis/rembg`. 242 | 243 | Try this: 244 | 245 | ``` 246 | docker run danielgatis/rembg i path/to/input.png path/to/output.png 247 | ``` 248 | 249 | ## Models 250 | 251 | All models are downloaded and saved in the user home folder in the `.u2net` directory. 252 | 253 | The available models are: 254 | 255 | - u2net ([download](https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2net.onnx), [source](https://github.com/xuebinqin/U-2-Net)): A pre-trained model for general use cases. 256 | - u2netp ([download](https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2netp.onnx), [source](https://github.com/xuebinqin/U-2-Net)): A lightweight version of u2net model. 257 | - u2net_human_seg ([download](https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2net_human_seg.onnx), [source](https://github.com/xuebinqin/U-2-Net)): A pre-trained model for human segmentation. 258 | - u2net_cloth_seg ([download](https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2net_cloth_seg.onnx), [source](https://github.com/levindabhi/cloth-segmentation)): A pre-trained model for Cloths Parsing from human portrait. Here clothes are parsed into 3 category: Upper body, Lower body and Full body. 259 | - silueta ([download](https://github.com/danielgatis/rembg/releases/download/v0.0.0/silueta.onnx), [source](https://github.com/xuebinqin/U-2-Net/issues/295)): Same as u2net but the size is reduced to 43Mb. 260 | 261 | ### How to train your own model 262 | 263 | If You need more fine tunned models try this: 264 | https://github.com/danielgatis/rembg/issues/193#issuecomment-1055534289 265 | 266 | 267 | ## Some video tutorials 268 | 269 | - https://www.youtube.com/watch?v=3xqwpXjxyMQ 270 | - https://www.youtube.com/watch?v=dFKRGXdkGJU 271 | - https://www.youtube.com/watch?v=Ai-BS_T7yjE 272 | - https://www.youtube.com/watch?v=dFKRGXdkGJU 273 | - https://www.youtube.com/watch?v=D7W-C0urVcQ 274 | 275 | ## References 276 | 277 | - https://arxiv.org/pdf/2005.09007.pdf 278 | - https://github.com/NathanUA/U-2-Net 279 | - https://github.com/pymatting/pymatting 280 | 281 | ## Buy me a coffee 282 | 283 | Liked some of my work? Buy me a coffee (or more likely a beer) 284 | 285 | Buy Me A Coffee 286 | 287 | ## License 288 | 289 | Copyright (c) 2020-present [Daniel Gatis](https://github.com/danielgatis) 290 | 291 | Licensed under [MIT License](./LICENSE.txt) 292 | --------------------------------------------------------------------------------