Spaces:
Configuration error
Configuration error
File size: 11,058 Bytes
1ba539f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 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 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 |
# Supplementary Material
## Training and test data
We provide a [website](https://zju3dv.github.io/zju_mocap/) for visualization.
The multi-view videos are captured by 23 cameras. We train our model on the "0, 6, 12, 18" cameras and test it on the remaining cameras.
The following table shows the detailed frame numbers for training and test of each video. Since the video length of each subject is different, we choose the appropriate number of frames for training and test.
**Note that since rendering is very slow, we test our model every 30 frames. For example, although the frame range of video 313 is "0-59", we only test our model on the 0-th and 30-th frames.**
| Video | 313 | 315 | 377 | 386 | 387 | 390 | 392 | 393 | 394 |
| :-----: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| Number of frames | 1470 | 2185 | 617 | 646 | 654 | 1171 | 556 | 658 | 859 |
| Frame Range (Training) | 0-59 | 0-399 | 0-299 | 0-299 | 0-299 | 700-999 | 0-299 | 0-299 | 0-299 |
| Frame Range (Unseen human poses) | 60-1060 | 400-1400 | 300-617 | 300-646 | 300-654 | 0-700 | 300-556 | 300-658 | 300-859 |
## Evaluation metrics
**We save our rendering results on novel views of training frames and unseen human poses at [here](https://zjueducn-my.sharepoint.com/:u:/g/personal/pengsida_zju_edu_cn/Ea3VOUy204VAiVJ-V-OGd9YBxdhbtfpS-U6icD_rDq0mUQ?e=cAcylK).**
As described in the paper, we evaluate our model in terms of the PSNR and SSIM metrics.
A straightforward way for evaluation is calculating the metrics on the whole image. Since we already know the 3D bounding box of the target human, we can project the 3D box to obtain a `bound_mask` and make the colors of pixels outside the mask as zero, as shown in the following figure.
As a result, the PSNR and SSIM metrics appear very high performances, as shown in the following table.
<table style="text-align: center">
<tr>
<td></td>
<td colspan="2">Training frames</td>
<td colspan="2">Unseen human poses</td>
</tr>
<tr>
<td></td>
<td>PSNR</td>
<td>SSIM</td>
<td>PSNR</td>
<td>SSIM</td>
</tr>
<tr>
<td>313</td>
<td>35.21 </td>
<td>0.985 </td>
<td>29.02 </td>
<td>0.964 </td>
</tr>
<tr>
<td>315</td>
<td>33.07 </td>
<td>0.988 </td>
<td>25.70 </td>
<td>0.957 </td>
</tr>
<tr>
<td>392</td>
<td>35.76 </td>
<td>0.984 </td>
<td>31.53 </td>
<td>0.971 </td>
</tr>
<tr>
<td>393</td>
<td>33.24 </td>
<td>0.979 </td>
<td>28.40 </td>
<td>0.960 </td>
</tr>
<tr>
<td>394</td>
<td>34.31 </td>
<td>0.980 </td>
<td>29.61 </td>
<td>0.961 </td>
</tr>
<tr>
<td>377</td>
<td>33.86 </td>
<td>0.985 </td>
<td>30.60 </td>
<td>0.977 </td>
</tr>
<tr>
<td>386</td>
<td>36.07 </td>
<td>0.984 </td>
<td>33.05 </td>
<td>0.974 </td>
</tr>
<tr>
<td>390</td>
<td>34.48 </td>
<td>0.980 </td>
<td>30.25 </td>
<td>0.964 </td>
</tr>
<tr>
<td>387</td>
<td>31.39 </td>
<td>0.975 </td>
<td>27.68 </td>
<td>0.961 </td>
</tr>
<tr>
<td></td>
<td>34.15 </td>
<td>0.982 </td>
<td>29.54 </td>
<td>0.966 </td>
</tr>
</table>
To overcome this problem, a solution is only calculating the metrics on pixels inside the `bound_mask`. Since the SSIM metric requires the input to have the image format, we first compute the 2D box that bounds the `bound_mask` and then crop the corresponding image region.
```python
def ssim_metric(rgb_pred, rgb_gt, batch):
mask_at_box = batch['mask_at_box'][0].detach().cpu().numpy()
H, W = int(cfg.H * cfg.ratio), int(cfg.W * cfg.ratio)
mask_at_box = mask_at_box.reshape(H, W)
# convert the pixels into an image
img_pred = np.zeros((H, W, 3))
img_pred[mask_at_box] = rgb_pred
img_gt = np.zeros((H, W, 3))
img_gt[mask_at_box] = rgb_gt
# crop the object region
x, y, w, h = cv2.boundingRect(mask_at_box.astype(np.uint8))
img_pred = img_pred[y:y + h, x:x + w]
img_gt = img_gt[y:y + h, x:x + w]
# compute the ssim
ssim = compare_ssim(img_pred, img_gt, multichannel=True)
return ssim
```
The following table lists corresponding results.
<table style="text-align: center">
<tr>
<td></td>
<td colspan="2">Training frames</td>
<td colspan="2">Unseen human poses</td>
</tr>
<tr>
<td></td>
<td>PSNR</td>
<td>SSIM</td>
<td>PSNR</td>
<td>SSIM</td>
</tr>
<tr>
<td>313</td>
<td>30.56 </td>
<td>0.971 </td>
<td>23.95 </td>
<td>0.905 </td>
</tr>
<tr>
<td>315</td>
<td>27.24 </td>
<td>0.962 </td>
<td>19.56 </td>
<td>0.852 </td>
</tr>
<tr>
<td>392</td>
<td>29.44 </td>
<td>0.946 </td>
<td>25.76 </td>
<td>0.909 </td>
</tr>
<tr>
<td>394</td>
<td>28.44 </td>
<td>0.940 </td>
<td>23.80 </td>
<td>0.878 </td>
</tr>
<tr>
<td>393</td>
<td>27.58 </td>
<td>0.939 </td>
<td>23.25 </td>
<td>0.893 </td>
</tr>
<tr>
<td>377</td>
<td>27.64 </td>
<td>0.951 </td>
<td>23.91 </td>
<td>0.909 </td>
</tr>
<tr>
<td>386</td>
<td>28.60 </td>
<td>0.931 </td>
<td>25.68 </td>
<td>0.881 </td>
</tr>
<tr>
<td>387</td>
<td>25.79 </td>
<td>0.928 </td>
<td>21.60 </td>
<td>0.870 </td>
</tr>
<tr>
<td>390</td>
<td>27.59 </td>
<td>0.926 </td>
<td>23.90 </td>
<td>0.870 </td>
</tr>
<tr>
<td></td>
<td>28.10 </td>
<td>0.944 </td>
<td>23.49 </td>
<td>0.885 </td>
</tr>
</table>
## Results of other methods on ZJU-MoCap
We save rendering results of other methods on novel views of training frames and unseen human poses at [here](https://zjueducn-my.sharepoint.com/:u:/g/personal/pengsida_zju_edu_cn/EQaPRQww70NDqEXeSG-fOeAB5JXFSWiWDW223h5nmkHvwQ?e=mdofbl), including Neural Volumes, Multi-view Neural Human Rendering, and Deferred Neural Human Rendering. **Note that we only generate novel views of training frames for Neural Volumes.**
The following table lists quantitative results of Neural Volumes.
<table style="text-align: center">
<tr>
<td></td>
<td>PSNR</td>
<td>SSIM</td>
</tr>
<tr>
<td>313</td>
<td>20.09 </td>
<td>0.831 </td>
</tr>
<tr>
<td>315</td>
<td>18.57 </td>
<td>0.824 </td>
</tr>
<tr>
<td>392</td>
<td>22.88 </td>
<td>0.726 </td>
</tr>
<tr>
<td>394</td>
<td>22.08 </td>
<td>0.843 </td>
</tr>
<tr>
<td>393</td>
<td>21.29 </td>
<td>0.842 </td>
</tr>
<tr>
<td>377</td>
<td>21.15 </td>
<td>0.842 </td>
</tr>
<tr>
<td>386</td>
<td>23.21 </td>
<td>0.820 </td>
</tr>
<tr>
<td>387</td>
<td>20.74 </td>
<td>0.838 </td>
</tr>
<tr>
<td>390</td>
<td>22.49 </td>
<td>0.825 </td>
</tr>
<tr>
<td></td>
<td>21.39 </td>
<td>0.821 </td>
</tr>
</table>
The following table lists quantitative results of Multi-view Neural Human Rendering.
<table style="text-align: center">
<tr>
<td></td>
<td colspan="2">Training frames</td>
<td colspan="2">Unseen human poses</td>
</tr>
<tr>
<td></td>
<td>PSNR</td>
<td>SSIM</td>
<td>PSNR</td>
<td>SSIM</td>
</tr>
<tr>
<td>313</td>
<td>26.68 </td>
<td>0.935 </td>
<td>23.05 </td>
<td>0.893 </td>
</tr>
<tr>
<td>315</td>
<td>19.81 </td>
<td>0.874 </td>
<td>18.88 </td>
<td>0.844 </td>
</tr>
<tr>
<td>392</td>
<td>24.73 </td>
<td>0.902 </td>
<td>23.66 </td>
<td>0.893 </td>
</tr>
<tr>
<td>394</td>
<td>25.01 </td>
<td>0.906 </td>
<td>22.87 </td>
<td>0.874 </td>
</tr>
<tr>
<td>393</td>
<td>23.47 </td>
<td>0.894 </td>
<td>22.27 </td>
<td>0.885 </td>
</tr>
<tr>
<td>377</td>
<td>23.79 </td>
<td>0.918 </td>
<td>21.94 </td>
<td>0.885 </td>
</tr>
<tr>
<td>386</td>
<td>25.02 </td>
<td>0.879 </td>
<td>23.70 </td>
<td>0.853 </td>
</tr>
<tr>
<td>387</td>
<td>22.65 </td>
<td>0.858 </td>
<td>20.97 </td>
<td>0.866 </td>
</tr>
<tr>
<td>390</td>
<td>23.72 </td>
<td>0.873 </td>
<td>22.65 </td>
<td>0.858 </td>
</tr>
<tr>
<td></td>
<td>23.87 </td>
<td>0.893 </td>
<td>22.22 </td>
<td>0.872 </td>
</tr>
</table>
The following table lists quantitative results of Deferred Neural Human Rendering.
<table style="text-align: center">
<tr>
<td></td>
<td colspan="2">Training frames</td>
<td colspan="2">Unseen human poses</td>
</tr>
<tr>
<td></td>
<td>PSNR</td>
<td>SSIM</td>
<td>PSNR</td>
<td>SSIM</td>
</tr>
<tr>
<td>313</td>
<td>25.78 </td>
<td>0.929 </td>
<td>22.56 </td>
<td>0.889 </td>
</tr>
<tr>
<td>315</td>
<td>19.44 </td>
<td>0.869 </td>
<td>18.38 </td>
<td>0.841 </td>
</tr>
<tr>
<td>392</td>
<td>24.96 </td>
<td>0.905 </td>
<td>24.08 </td>
<td>0.900 </td>
</tr>
<tr>
<td>394</td>
<td>24.84 </td>
<td>0.903 </td>
<td>22.67 </td>
<td>0.871 </td>
</tr>
<tr>
<td>393</td>
<td>23.50 </td>
<td>0.896 </td>
<td>22.45 </td>
<td>0.888 </td>
</tr>
<tr>
<td>377</td>
<td>23.74 </td>
<td>0.917 </td>
<td>22.07 </td>
<td>0.886 </td>
</tr>
<tr>
<td>386</td>
<td>24.93 </td>
<td>0.877 </td>
<td>23.70 </td>
<td>0.851 </td>
</tr>
<tr>
<td>387</td>
<td>22.44 </td>
<td>0.888 </td>
<td>20.64 </td>
<td>0.862 </td>
</tr>
<tr>
<td>390</td>
<td>24.33 </td>
<td>0.881 </td>
<td>22.90 </td>
<td>0.864 </td>
</tr>
<tr>
<td></td>
<td>23.77 </td>
<td>0.896 </td>
<td>22.16 </td>
<td>0.872 </td>
</tr>
</table>
|