vLAR commited on
Commit
ab2d71b
ยท
verified ยท
1 Parent(s): 829a98c

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +24 -57
README.md CHANGED
@@ -143,41 +143,10 @@ homepage: https://vlar-group.github.io/PhysInOne.html
143
  <h1 align="center" style="font-size: 42px;">PhysInOne: Visual Physics Learning and Reasoning in One Suite</h1>
144
 
145
  <p align="center">
146
- Siyuan Zhou<sup>1โ€ </sup>
147
- Hejun Wang<sup>1โ€ </sup>
148
- Hu Cheng<sup>1โ€ </sup>
149
- Jinxi Li<sup>1โ€ </sup>
150
- Dongsheng Wang<sup>1โ€ </sup>
151
- Junwei Jiang<sup>1โ€ </sup> <br>
152
- Yixiao Jin<sup>1โ€ </sup>
153
- Jiayue Huang<sup>1โ€ </sup>
154
- Shiwei Mao<sup>1โ€ </sup>
155
- Shangjia Liu<sup>2</sup>
156
- Yafei Yang<sup>1</sup> <br>
157
- Hongkang Song<sup>1</sup>
158
- Shenxing Wei<sup>1</sup>
159
- Zihui Zhang<sup>1</sup>
160
- DataTeam<sup>1*</sup>
161
- Bing Wang<sup>2</sup>
162
- Zhihua Wang<sup>3</sup>
163
- Chuhang Zou<sup>4</sup>
164
- Bo Yang<sup>1โ€ก</sup>
165
- </p>
166
-
167
- <p align="center">
168
- <sup>1</sup> vLAR Group
169
- <sup>2</sup> The Hong Kong Polytechnic University
170
- <sup>3</sup> Syai Singapore
171
- <sup>4</sup> Meta
172
- </p>
173
-
174
- <p align="center">
175
- <sup>โ€ </sup> Equal Contribution
176
- <sup>โ€ก</sup> Corresponding Author
177
- </p>
178
-
179
- <p align="center">
180
- <sup>*</sup> {Peng Huang, Shijie Liu, Zhengli Hao, Hao Li, Yitian Li, Wenqi Zhou, Zhihan Zhao, Zongqi He, Hongtao Wen, Shouwang Huang, Peng Yun, Bowen Cheng, Pok Kazaf Fu, Wai Kit Lai, Jiahao Chen, Kaiyuan Wang, Zhixuan Sun, Ziqi Li, Haochen Hu, Di Zhang, Chun Ho Yuen}
181
  </p>
182
 
183
  <p align="center">
@@ -210,26 +179,26 @@ homepage: https://vlar-group.github.io/PhysInOne.html
210
  <a href="#release-timetable">๐Ÿš€ Release Timetable</a><br>
211
  <a href="#data-splits">๐Ÿ“Š Data Splits</a><br>
212
  <a href="#how-to-download">๐Ÿ“ฅ How to Download</a><br>
213
- &nbsp;&nbsp;&nbsp;<a href="#download-scripts">1. Download Scripts</a><br>
214
- &nbsp;&nbsp;&nbsp;<a href="#install-dependencies">2. Install Dependencies</a><br>
215
- &nbsp;&nbsp;&nbsp;<a href="#filter-cases">3. Filter Cases</a><br>
216
- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#filter-by-split">3.1 Split</a><br>
217
- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#filter-by-split-and-activity">3.2 Split + Activity</a><br>
218
- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#filter-by-phenomenon">3.3 Phenomenon</a><br>
219
- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#random-sampling">3.4 Random Sampling</a><br>
220
- &nbsp;&nbsp;&nbsp;<a href="#download-selected-cases">4. Download Cases</a><br>
221
  <a href="#dataset-structure">๐Ÿ“ฆ Dataset Structure</a><br>
222
  <a href="#annotation-details">๐Ÿ“ Annotation Details</a><br>
223
- &nbsp;&nbsp;&nbsp;โ€ข <a href="#depth">Depth</a><br>
224
- &nbsp;&nbsp;&nbsp;โ€ข <a href="#segmentation">Segmentation</a><br>
225
- &nbsp;&nbsp;&nbsp;โ€ข <a href="#captions">Captions</a><br>
226
- &nbsp;&nbsp;&nbsp;โ€ข <a href="#trajectories">Trajectories</a><br>
227
- &nbsp;&nbsp;&nbsp;โ€ข <a href="#cameras">Cameras</a><br>
228
- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;โ—ฆ <a href="#camera-json-structure">Camera JSON structure</a><br>
229
- &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;โ—ฆ <a href="#dynamic-camera-behavior">Dynamic camera behavior</a><br>
230
- &nbsp;&nbsp;&nbsp;โ€ข <a href="#point-clouds">Point Clouds</a><br>
231
  <a href="#data-fields">๐Ÿงฉ Data Fields</a><br>
232
- &nbsp;&nbsp;&nbsp;โ€ข <a href="#physical-phenomena">Physical Phenomena</a><br>
233
  <a href="#license">๐Ÿ“œ License</a><br>
234
  <a href="#citation">๐Ÿ“š Citation</a><br>
235
  <a href="#contact">๐Ÿ“ฎ Contact</a><br>
@@ -240,11 +209,9 @@ homepage: https://vlar-group.github.io/PhysInOne.html
240
  <h2 id="summary" style="margin-left: 0; font-size: 1.9em; line-height: 1.22; margin-top: 1.2em; margin-bottom: 0.5em;">๐Ÿ“Œ Summary</h2>
241
  <div style="margin-left: 24px;">
242
 
243
- **PhysInOne** is a large-scale synthetic dataset for visual physics learning and reasoning. It contains **153,810 dynamic 3D scenes** and **2 million annotated videos**, covering **71 basic physical phenomena** across four domains of everyday physics: **mechanics**, **optics**, **fluid dynamics**, and **magnetism**.
244
-
245
- Each scene may contain multi-object and multi-physics interactions in complex 3D environments. PhysInOne provides rich annotations including RGB videos, depth maps, object masks, 3D trajectories, camera poses, object meshes, material properties, and textual descriptions.
246
-
247
- The dataset is designed to support research on physics-aware video generation, future frame prediction, physical property estimation, motion transfer, physical reasoning, and world models.
248
 
249
  <div class="physinone-summary-grid" style="display: grid; grid-template-columns: repeat(2, minmax(0, 1fr)); gap: 10px; margin: 18px 0;">
250
 
 
143
  <h1 align="center" style="font-size: 42px;">PhysInOne: Visual Physics Learning and Reasoning in One Suite</h1>
144
 
145
  <p align="center">
146
+ vLAR Group |
147
+ The Hong Kong Polytechnic University |
148
+ Syai Singapore |
149
+ Meta
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
150
  </p>
151
 
152
  <p align="center">
 
179
  <a href="#release-timetable">๐Ÿš€ Release Timetable</a><br>
180
  <a href="#data-splits">๐Ÿ“Š Data Splits</a><br>
181
  <a href="#how-to-download">๐Ÿ“ฅ How to Download</a><br>
182
+ <a href="#download-scripts">1. Download Scripts</a><br>
183
+ <a href="#install-dependencies">2. Install Dependencies</a><br>
184
+ <a href="#filter-cases">3. Filter Cases</a><br>
185
+ <a href="#filter-by-split">3.1 Split</a><br>
186
+ <a href="#filter-by-split-and-activity">3.2 Split + Activity</a><br>
187
+ <a href="#filter-by-phenomenon">3.3 Phenomenon</a><br>
188
+ <a href="#random-sampling">3.4 Random Sampling</a><br>
189
+ <a href="#download-selected-cases">4. Download Cases</a><br>
190
  <a href="#dataset-structure">๐Ÿ“ฆ Dataset Structure</a><br>
191
  <a href="#annotation-details">๐Ÿ“ Annotation Details</a><br>
192
+ โ€ข <a href="#depth">Depth</a><br>
193
+ โ€ข <a href="#segmentation">Segmentation</a><br>
194
+ โ€ข <a href="#captions">Captions</a><br>
195
+ โ€ข <a href="#trajectories">Trajectories</a><br>
196
+ โ€ข <a href="#cameras">Cameras</a><br>
197
+ โ—ฆ <a href="#camera-json-structure">Camera JSON structure</a><br>
198
+ โ—ฆ <a href="#dynamic-camera-behavior">Dynamic camera behavior</a><br>
199
+ โ€ข <a href="#point-clouds">Point Clouds</a><br>
200
  <a href="#data-fields">๐Ÿงฉ Data Fields</a><br>
201
+ โ€ข <a href="#physical-phenomena">Physical Phenomena</a><br>
202
  <a href="#license">๐Ÿ“œ License</a><br>
203
  <a href="#citation">๐Ÿ“š Citation</a><br>
204
  <a href="#contact">๐Ÿ“ฎ Contact</a><br>
 
209
  <h2 id="summary" style="margin-left: 0; font-size: 1.9em; line-height: 1.22; margin-top: 1.2em; margin-bottom: 0.5em;">๐Ÿ“Œ Summary</h2>
210
  <div style="margin-left: 24px;">
211
 
212
+ <p style="text-align: justify;">
213
+ <strong>PhysInOne</strong> is a large-scale synthetic dataset for visual physics learning and reasoning, containing <strong>153,810 dynamic 3D scenes</strong> and <strong>2 million annotated videos</strong> across <strong>71 physical phenomena</strong> in mechanics, optics, fluid dynamics, and magnetism. Each scene features complex multi-object and multi-physics interactions with rich annotations, including RGB videos, depth maps, object masks, 3D trajectories, camera poses, object meshes, material properties, and textual descriptions. The dataset supports research on physics-aware video generation, future frame prediction, physical property estimation, motion transfer, physical reasoning, and world models.
214
+ </p>
 
 
215
 
216
  <div class="physinone-summary-grid" style="display: grid; grid-template-columns: repeat(2, minmax(0, 1fr)); gap: 10px; margin: 18px 0;">
217