├── CONTRIBUTING.md ├── LICENSE ├── README.md └── data_reader.py /CONTRIBUTING.md: -------------------------------------------------------------------------------- 1 | # How to Contribute 2 | 3 | We'd love to accept your patches and contributions to this project. There are 4 | just a few small guidelines you need to follow. 5 | 6 | ## Contributor License Agreement 7 | 8 | Contributions to this project must be accompanied by a Contributor License 9 | Agreement. You (or your employer) retain the copyright to your contribution; 10 | this simply gives us permission to use and redistribute your contributions as 11 | part of the project. Head over to to see 12 | your current agreements on file or to sign a new one. 13 | 14 | You generally only need to submit a CLA once, so if you've already submitted one 15 | (even if it was for a different project), you probably don't need to do it 16 | again. 17 | 18 | ## Code reviews 19 | 20 | All submissions, including submissions by project members, require review. We 21 | use GitHub pull requests for this purpose. Consult 22 | [GitHub Help](https://help.github.com/articles/about-pull-requests/) for more 23 | information on using pull requests. 24 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | 2 | Apache License 3 | Version 2.0, January 2004 4 | http://www.apache.org/licenses/ 5 | 6 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 7 | 8 | 1. Definitions. 9 | 10 | "License" shall mean the terms and conditions for use, reproduction, 11 | and distribution as defined by Sections 1 through 9 of this document. 12 | 13 | "Licensor" shall mean the copyright owner or entity authorized by 14 | the copyright owner that is granting the License. 15 | 16 | "Legal Entity" shall mean the union of the acting entity and all 17 | other entities that control, are controlled by, or are under common 18 | control with that entity. For the purposes of this definition, 19 | "control" means (i) the power, direct or indirect, to cause the 20 | direction or management of such entity, whether by contract or 21 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 22 | outstanding shares, or (iii) beneficial ownership of such entity. 23 | 24 | "You" (or "Your") shall mean an individual or Legal Entity 25 | exercising permissions granted by this License. 26 | 27 | "Source" form shall mean the preferred form for making modifications, 28 | including but not limited to software source code, documentation 29 | source, and configuration files. 30 | 31 | "Object" form shall mean any form resulting from mechanical 32 | transformation or translation of a Source form, including but 33 | not limited to compiled object code, generated documentation, 34 | and conversions to other media types. 35 | 36 | "Work" shall mean the work of authorship, whether in Source or 37 | Object form, made available under the License, as indicated by a 38 | copyright notice that is included in or attached to the work 39 | (an example is provided in the Appendix below). 40 | 41 | "Derivative Works" shall mean any work, whether in Source or Object 42 | form, that is based on (or derived from) the Work and for which the 43 | editorial revisions, annotations, elaborations, or other modifications 44 | represent, as a whole, an original work of authorship. For the purposes 45 | of this License, Derivative Works shall not include works that remain 46 | separable from, or merely link (or bind by name) to the interfaces of, 47 | the Work and Derivative Works thereof. 48 | 49 | "Contribution" shall mean any work of authorship, including 50 | the original version of the Work and any modifications or additions 51 | to that Work or Derivative Works thereof, that is intentionally 52 | submitted to Licensor for inclusion in the Work by the copyright owner 53 | or by an individual or Legal Entity authorized to submit on behalf of 54 | the copyright owner. For the purposes of this definition, "submitted" 55 | means any form of electronic, verbal, or written communication sent 56 | to the Licensor or its representatives, including but not limited to 57 | communication on electronic mailing lists, source code control systems, 58 | and issue tracking systems that are managed by, or on behalf of, the 59 | Licensor for the purpose of discussing and improving the Work, but 60 | excluding communication that is conspicuously marked or otherwise 61 | designated in writing by the copyright owner as "Not a Contribution." 62 | 63 | "Contributor" shall mean Licensor and any individual or Legal Entity 64 | on behalf of whom a Contribution has been received by Licensor and 65 | subsequently incorporated within the Work. 66 | 67 | 2. Grant of Copyright License. Subject to the terms and conditions of 68 | this License, each Contributor hereby grants to You a perpetual, 69 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 70 | copyright license to reproduce, prepare Derivative Works of, 71 | publicly display, publicly perform, sublicense, and distribute the 72 | Work and such Derivative Works in Source or Object form. 73 | 74 | 3. Grant of Patent License. Subject to the terms and conditions of 75 | this License, each Contributor hereby grants to You a perpetual, 76 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 77 | (except as stated in this section) patent license to make, have made, 78 | use, offer to sell, sell, import, and otherwise transfer the Work, 79 | where such license applies only to those patent claims licensable 80 | by such Contributor that are necessarily infringed by their 81 | Contribution(s) alone or by combination of their Contribution(s) 82 | with the Work to which such Contribution(s) was submitted. If You 83 | institute patent litigation against any entity (including a 84 | cross-claim or counterclaim in a lawsuit) alleging that the Work 85 | or a Contribution incorporated within the Work constitutes direct 86 | or contributory patent infringement, then any patent licenses 87 | granted to You under this License for that Work shall terminate 88 | as of the date such litigation is filed. 89 | 90 | 4. Redistribution. You may reproduce and distribute copies of the 91 | Work or Derivative Works thereof in any medium, with or without 92 | modifications, and in Source or Object form, provided that You 93 | meet the following conditions: 94 | 95 | (a) You must give any other recipients of the Work or 96 | Derivative Works a copy of this License; and 97 | 98 | (b) You must cause any modified files to carry prominent notices 99 | stating that You changed the files; and 100 | 101 | (c) You must retain, in the Source form of any Derivative Works 102 | that You distribute, all copyright, patent, trademark, and 103 | attribution notices from the Source form of the Work, 104 | excluding those notices that do not pertain to any part of 105 | the Derivative Works; and 106 | 107 | (d) If the Work includes a "NOTICE" text file as part of its 108 | distribution, then any Derivative Works that You distribute must 109 | include a readable copy of the attribution notices contained 110 | within such NOTICE file, excluding those notices that do not 111 | pertain to any part of the Derivative Works, in at least one 112 | of the following places: within a NOTICE text file distributed 113 | as part of the Derivative Works; within the Source form or 114 | documentation, if provided along with the Derivative Works; or, 115 | within a display generated by the Derivative Works, if and 116 | wherever such third-party notices normally appear. The contents 117 | of the NOTICE file are for informational purposes only and 118 | do not modify the License. You may add Your own attribution 119 | notices within Derivative Works that You distribute, alongside 120 | or as an addendum to the NOTICE text from the Work, provided 121 | that such additional attribution notices cannot be construed 122 | as modifying the License. 123 | 124 | You may add Your own copyright statement to Your modifications and 125 | may provide additional or different license terms and conditions 126 | for use, reproduction, or distribution of Your modifications, or 127 | for any such Derivative Works as a whole, provided Your use, 128 | reproduction, and distribution of the Work otherwise complies with 129 | the conditions stated in this License. 130 | 131 | 5. Submission of Contributions. Unless You explicitly state otherwise, 132 | any Contribution intentionally submitted for inclusion in the Work 133 | by You to the Licensor shall be under the terms and conditions of 134 | this License, without any additional terms or conditions. 135 | Notwithstanding the above, nothing herein shall supersede or modify 136 | the terms of any separate license agreement you may have executed 137 | with Licensor regarding such Contributions. 138 | 139 | 6. Trademarks. This License does not grant permission to use the trade 140 | names, trademarks, service marks, or product names of the Licensor, 141 | except as required for reasonable and customary use in describing the 142 | origin of the Work and reproducing the content of the NOTICE file. 143 | 144 | 7. Disclaimer of Warranty. Unless required by applicable law or 145 | agreed to in writing, Licensor provides the Work (and each 146 | Contributor provides its Contributions) on an "AS IS" BASIS, 147 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 148 | implied, including, without limitation, any warranties or conditions 149 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 150 | PARTICULAR PURPOSE. You are solely responsible for determining the 151 | appropriateness of using or redistributing the Work and assume any 152 | risks associated with Your exercise of permissions under this License. 153 | 154 | 8. Limitation of Liability. In no event and under no legal theory, 155 | whether in tort (including negligence), contract, or otherwise, 156 | unless required by applicable law (such as deliberate and grossly 157 | negligent acts) or agreed to in writing, shall any Contributor be 158 | liable to You for damages, including any direct, indirect, special, 159 | incidental, or consequential damages of any character arising as a 160 | result of this License or out of the use or inability to use the 161 | Work (including but not limited to damages for loss of goodwill, 162 | work stoppage, computer failure or malfunction, or any and all 163 | other commercial damages or losses), even if such Contributor 164 | has been advised of the possibility of such damages. 165 | 166 | 9. Accepting Warranty or Additional Liability. While redistributing 167 | the Work or Derivative Works thereof, You may choose to offer, 168 | and charge a fee for, acceptance of support, warranty, indemnity, 169 | or other liability obligations and/or rights consistent with this 170 | License. However, in accepting such obligations, You may act only 171 | on Your own behalf and on Your sole responsibility, not on behalf 172 | of any other Contributor, and only if You agree to indemnify, 173 | defend, and hold each Contributor harmless for any liability 174 | incurred by, or claims asserted against, such Contributor by reason 175 | of your accepting any such warranty or additional liability. 176 | 177 | END OF TERMS AND CONDITIONS 178 | 179 | APPENDIX: How to apply the Apache License to your work. 180 | 181 | To apply the Apache License to your work, attach the following 182 | boilerplate notice, with the fields enclosed by brackets "[]" 183 | replaced with your own identifying information. (Don't include 184 | the brackets!) The text should be enclosed in the appropriate 185 | comment syntax for the file format. We also recommend that a 186 | file or class name and description of purpose be included on the 187 | same "printed page" as the copyright notice for easier 188 | identification within third-party archives. 189 | 190 | Copyright [yyyy] [name of copyright owner] 191 | 192 | Licensed under the Apache License, Version 2.0 (the "License"); 193 | you may not use this file except in compliance with the License. 194 | You may obtain a copy of the License at 195 | 196 | http://www.apache.org/licenses/LICENSE-2.0 197 | 198 | Unless required by applicable law or agreed to in writing, software 199 | distributed under the License is distributed on an "AS IS" BASIS, 200 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 201 | See the License for the specific language governing permissions and 202 | limitations under the License. 203 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | ## Datasets used to train Generative Query Networks (GQNs) in the ‘Neural Scene Representation and Rendering’ paper. 2 | 3 | 4 | The following version of the datasets are available: 5 | 6 | * __rooms_ring_camera__. Scenes of a variable number of random objects 7 | captured in a square room of size 7x7 units. Wall textures, floor textures 8 | as well as the shapes of the objects are randomly chosen within a fixed pool 9 | of discrete options. There are 5 possible wall textures (red, green, cerise, 10 | orange, yellow), 3 possible floor textures (yellow, white, blue) and 7 11 | possible object shapes (box, sphere, cylinder, capsule, cone, icosahedron 12 | and triangle). Each scene contains 1, 2 or 3 objects. In this simplified 13 | version of the dataset, the camera only moves on a fixed ring and always 14 | faces the center of the room. This is the ‘easiest’ version of the dataset, 15 | use version for fast training. 16 | 17 | * __rooms_free_camera_no_object_rotations__. As in __rooom_ring_camera__, 18 | except the camera moves freely. However the objects themselves do not rotate 19 | around their axes, which makes the modeling task somewhat easier. This 20 | version is ‘medium’ difficulty. 21 | 22 | * __rooms_free_camera_with_object_rotations__. As in 23 | __rooms_free_camera_no_object_rotations__, the camera moves freely, however 24 | objects can rotate around their vertical axes across scenes. This is the 25 | ‘hardest’ version of the dataset. 26 | 27 | * __jaco__. a reproduction of the robotic Jaco arm is placed in the middle of 28 | the room along with one spherical target object. The arm has nine joints. As 29 | above, the appearance of the room is modified for each episode by randomly 30 | choosing a different texture for the walls and floor from a fixed pool of 31 | options. In addition, we modify both colour and position of the target 32 | randomly. Finally, the joint angles of the arm are also initialised at 33 | random within a range of physically sensible positions. 34 | 35 | * __shepard_metzler_5_parts__. Each object is composed of 7 randomly coloured 36 | cubes that are positioned by a self-avoiding random walk in 3D grid. As 37 | above, the camera is parametrised by its position, yaw and pitch, however it 38 | is constrained to only move around the object at a fixed distance from its 39 | centre. This is the ‘easy’ version of the dataset, where each object is 40 | composed of only 5 parts. 41 | 42 | * __shepard_metzler_7_parts__. This is the ‘hard’ version of the above 43 | dataset, where each object is composed of 7 parts. 44 | 45 | * __mazes__. Random mazes that were created using an OpenGL-based [DeepMind 46 | Lab](https://github.com/deepmind/lab) game engine (Beattie et al., 2016). 47 | Each maze is constructed out of an underlying 7 by 7 grid, with walls 48 | falling on the boundaries of the grid locations. However, the agent can be 49 | positioned at any continuous position in the maze. The mazes contain 1 or 2 50 | rooms, with multiple connecting corridors. The walls and floor textures of 51 | each maze are determined by random uniform sampling from a predefined set of 52 | textures. 53 | 54 | ### Usage example 55 | 56 | To select what dataset to load, instantiate a reader passing the correct 57 | `version` argument. Note that the constructor will set up all the queues used by 58 | the reader. To get tensors call `read` on the data reader passing in the desired 59 | batch size. 60 | 61 | ```python 62 | import tensorflow as tf 63 | 64 | root_path = 'path/to/datasets/root/folder' 65 | data_reader = DataReader(dataset='jaco', context_size=5, root=root_path) 66 | data = data_reader.read(batch_size=12) 67 | 68 | with tf.train.SingularMonitoredSession() as sess: 69 | d = sess.run(data) 70 | ``` 71 | 72 | ### Download 73 | 74 | Raw data files referred to in this document are available to download 75 | [here](https://console.cloud.google.com/storage/gqn-dataset). To download the 76 | datasets you can use 77 | the [`gsutil cp`](https://cloud.google.com/storage/docs/gsutil/commands/cp) 78 | command; see also the `gsutil` [installation instructions](https://cloud.google.com/storage/docs/gsutil_install). 79 | 80 | 81 | ### Notes 82 | 83 | This is not an official Google product. 84 | -------------------------------------------------------------------------------- /data_reader.py: -------------------------------------------------------------------------------- 1 | # Copyright 2018 Google LLC 2 | # 3 | # Licensed under the Apache License, Version 2.0 (the "License"); 4 | # you may not use this file except in compliance with the License. 5 | # You may obtain a copy of the License at 6 | # 7 | # https://www.apache.org/licenses/LICENSE-2.0 8 | # 9 | # Unless required by applicable law or agreed to in writing, software 10 | # distributed under the License is distributed on an "AS IS" BASIS, 11 | # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12 | # See the License for the specific language governing permissions and 13 | # limitations under the License. 14 | 15 | """Minimal data reader for GQN TFRecord datasets.""" 16 | 17 | from __future__ import absolute_import 18 | from __future__ import division 19 | from __future__ import print_function 20 | 21 | import collections 22 | import os 23 | import tensorflow as tf 24 | nest = tf.contrib.framework.nest 25 | 26 | DatasetInfo = collections.namedtuple( 27 | 'DatasetInfo', 28 | ['basepath', 'train_size', 'test_size', 'frame_size', 'sequence_size'] 29 | ) 30 | Context = collections.namedtuple('Context', ['frames', 'cameras']) 31 | Query = collections.namedtuple('Query', ['context', 'query_camera']) 32 | TaskData = collections.namedtuple('TaskData', ['query', 'target']) 33 | 34 | 35 | _DATASETS = dict( 36 | jaco=DatasetInfo( 37 | basepath='jaco', 38 | train_size=3600, 39 | test_size=400, 40 | frame_size=64, 41 | sequence_size=11), 42 | 43 | mazes=DatasetInfo( 44 | basepath='mazes', 45 | train_size=1080, 46 | test_size=120, 47 | frame_size=84, 48 | sequence_size=300), 49 | 50 | rooms_free_camera_with_object_rotations=DatasetInfo( 51 | basepath='rooms_free_camera_with_object_rotations', 52 | train_size=2034, 53 | test_size=226, 54 | frame_size=128, 55 | sequence_size=10), 56 | 57 | rooms_ring_camera=DatasetInfo( 58 | basepath='rooms_ring_camera', 59 | train_size=2160, 60 | test_size=240, 61 | frame_size=64, 62 | sequence_size=10), 63 | 64 | rooms_free_camera_no_object_rotations=DatasetInfo( 65 | basepath='rooms_free_camera_no_object_rotations', 66 | train_size=2160, 67 | test_size=240, 68 | frame_size=64, 69 | sequence_size=10), 70 | 71 | shepard_metzler_5_parts=DatasetInfo( 72 | basepath='shepard_metzler_5_parts', 73 | train_size=900, 74 | test_size=100, 75 | frame_size=64, 76 | sequence_size=15), 77 | 78 | shepard_metzler_7_parts=DatasetInfo( 79 | basepath='shepard_metzler_7_parts', 80 | train_size=900, 81 | test_size=100, 82 | frame_size=64, 83 | sequence_size=15) 84 | ) 85 | _NUM_CHANNELS = 3 86 | _NUM_RAW_CAMERA_PARAMS = 5 87 | _MODES = ('train', 'test') 88 | 89 | 90 | def _get_dataset_files(dateset_info, mode, root): 91 | """Generates lists of files for a given dataset version.""" 92 | basepath = dateset_info.basepath 93 | base = os.path.join(root, basepath, mode) 94 | if mode == 'train': 95 | num_files = dateset_info.train_size 96 | else: 97 | num_files = dateset_info.test_size 98 | 99 | length = len(str(num_files)) 100 | template = '{:0%d}-of-{:0%d}.tfrecord' % (length, length) 101 | return [os.path.join(base, template.format(i + 1, num_files)) 102 | for i in range(num_files)] 103 | 104 | 105 | def _convert_frame_data(jpeg_data): 106 | decoded_frames = tf.image.decode_jpeg(jpeg_data) 107 | return tf.image.convert_image_dtype(decoded_frames, dtype=tf.float32) 108 | 109 | 110 | class DataReader(object): 111 | """Minimal queue based TFRecord reader. 112 | 113 | You can use this reader to load the datasets used to train Generative Query 114 | Networks (GQNs) in the 'Neural Scene Representation and Rendering' paper. 115 | See README.md for a description of the datasets and an example of how to use 116 | the reader. 117 | """ 118 | 119 | def __init__(self, 120 | dataset, 121 | context_size, 122 | root, 123 | mode='train', 124 | # Optionally reshape frames 125 | custom_frame_size=None, 126 | # Queue params 127 | num_threads=4, 128 | capacity=256, 129 | min_after_dequeue=128, 130 | seed=None): 131 | """Instantiates a DataReader object and sets up queues for data reading. 132 | 133 | Args: 134 | dataset: string, one of ['jaco', 'mazes', 'rooms_ring_camera', 135 | 'rooms_free_camera_no_object_rotations', 136 | 'rooms_free_camera_with_object_rotations', 'shepard_metzler_5_parts', 137 | 'shepard_metzler_7_parts']. 138 | context_size: integer, number of views to be used to assemble the context. 139 | root: string, path to the root folder of the data. 140 | mode: (optional) string, one of ['train', 'test']. 141 | custom_frame_size: (optional) integer, required size of the returned 142 | frames, defaults to None. 143 | num_threads: (optional) integer, number of threads used to feed the reader 144 | queues, defaults to 4. 145 | capacity: (optional) integer, capacity of the underlying 146 | RandomShuffleQueue, defualts to 256. 147 | min_after_dequeue: (optional) integer, min_after_dequeue of the underlying 148 | RandomShuffleQueue, defualts to 128. 149 | seed: (optional) integer, seed for the random number generators used in 150 | the reader. 151 | 152 | Raises: 153 | ValueError: if the required version does not exist; if the required mode 154 | is not supported; if the requested context_size is bigger than the 155 | maximum supported for the given dataset version. 156 | """ 157 | 158 | if dataset not in _DATASETS: 159 | raise ValueError('Unrecognized dataset {} requested. Available datasets ' 160 | 'are {}'.format(dataset, _DATASETS.keys())) 161 | 162 | if mode not in _MODES: 163 | raise ValueError('Unsupported mode {} requested. Supported modes ' 164 | 'are {}'.format(mode, _MODES)) 165 | 166 | self._dataset_info = _DATASETS[dataset] 167 | 168 | if context_size >= self._dataset_info.sequence_size: 169 | raise ValueError( 170 | 'Maximum support context size for dataset {} is {}, but ' 171 | 'was {}.'.format( 172 | dataset, self._dataset_info.sequence_size-1, context_size)) 173 | 174 | self._context_size = context_size 175 | # Number of views in the context + target view 176 | self._example_size = context_size + 1 177 | self._custom_frame_size = custom_frame_size 178 | 179 | with tf.device('/cpu'): 180 | file_names = _get_dataset_files(self._dataset_info, mode, root) 181 | filename_queue = tf.train.string_input_producer(file_names, seed=seed) 182 | reader = tf.TFRecordReader() 183 | 184 | read_ops = [self._make_read_op(reader, filename_queue) 185 | for _ in range(num_threads)] 186 | 187 | dtypes = nest.map_structure(lambda x: x.dtype, read_ops[0]) 188 | shapes = nest.map_structure(lambda x: x.shape[1:], read_ops[0]) 189 | 190 | self._queue = tf.RandomShuffleQueue( 191 | capacity=capacity, 192 | min_after_dequeue=min_after_dequeue, 193 | dtypes=dtypes, 194 | shapes=shapes, 195 | seed=seed) 196 | 197 | enqueue_ops = [self._queue.enqueue_many(op) for op in read_ops] 198 | tf.train.add_queue_runner(tf.train.QueueRunner(self._queue, enqueue_ops)) 199 | 200 | def read(self, batch_size): 201 | """Reads batch_size (query, target) pairs.""" 202 | frames, cameras = self._queue.dequeue_many(batch_size) 203 | context_frames = frames[:, :-1] 204 | context_cameras = cameras[:, :-1] 205 | target = frames[:, -1] 206 | query_camera = cameras[:, -1] 207 | context = Context(cameras=context_cameras, frames=context_frames) 208 | query = Query(context=context, query_camera=query_camera) 209 | return TaskData(query=query, target=target) 210 | 211 | def _make_read_op(self, reader, filename_queue): 212 | """Instantiates the ops used to read and parse the data into tensors.""" 213 | _, raw_data = reader.read_up_to(filename_queue, num_records=16) 214 | feature_map = { 215 | 'frames': tf.FixedLenFeature( 216 | shape=self._dataset_info.sequence_size, dtype=tf.string), 217 | 'cameras': tf.FixedLenFeature( 218 | shape=[self._dataset_info.sequence_size * _NUM_RAW_CAMERA_PARAMS], 219 | dtype=tf.float32) 220 | } 221 | example = tf.parse_example(raw_data, feature_map) 222 | indices = self._get_randomized_indices(seed) 223 | frames = self._preprocess_frames(example, indices) 224 | cameras = self._preprocess_cameras(example, indices) 225 | return frames, cameras 226 | 227 | def _get_randomized_indices(self, seed): 228 | """Generates randomized indices into a sequence of a specific length.""" 229 | indices = tf.range(0, self._dataset_info.sequence_size) 230 | indices = tf.random_shuffle(indices, seed=seed) 231 | indices = tf.slice(indices, begin=[0], size=[self._example_size]) 232 | return indices 233 | 234 | def _preprocess_frames(self, example, indices): 235 | """Instantiates the ops used to preprocess the frames data.""" 236 | frames = tf.concat(example['frames'], axis=0) 237 | frames = tf.gather(frames, indices, axis=1) 238 | frames = tf.map_fn( 239 | _convert_frame_data, tf.reshape(frames, [-1]), 240 | dtype=tf.float32, back_prop=False) 241 | dataset_image_dimensions = tuple( 242 | [self._dataset_info.frame_size] * 2 + [_NUM_CHANNELS]) 243 | frames = tf.reshape( 244 | frames, (-1, self._example_size) + dataset_image_dimensions) 245 | if (self._custom_frame_size and 246 | self._custom_frame_size != self._dataset_info.frame_size): 247 | frames = tf.reshape(frames, (-1,) + dataset_image_dimensions) 248 | new_frame_dimensions = (self._custom_frame_size,) * 2 + (_NUM_CHANNELS,) 249 | frames = tf.image.resize_bilinear( 250 | frames, new_frame_dimensions[:2], align_corners=True) 251 | frames = tf.reshape( 252 | frames, (-1, self._example_size) + new_frame_dimensions) 253 | return frames 254 | 255 | def _preprocess_cameras(self, example, indices): 256 | """Instantiates the ops used to preprocess the cameras data.""" 257 | raw_pose_params = example['cameras'] 258 | raw_pose_params = tf.reshape( 259 | raw_pose_params, 260 | [-1, self._dataset_info.sequence_size, _NUM_RAW_CAMERA_PARAMS]) 261 | raw_pose_params = tf.gather(raw_pose_params, indices, axis=1) 262 | pos = raw_pose_params[:, :, 0:3] 263 | yaw = raw_pose_params[:, :, 3:4] 264 | pitch = raw_pose_params[:, :, 4:5] 265 | cameras = tf.concat( 266 | [pos, tf.sin(yaw), tf.cos(yaw), tf.sin(pitch), tf.cos(pitch)], axis=2) 267 | return cameras 268 | --------------------------------------------------------------------------------