Datasets:
Update dataset card / previews / scripts
Browse files
README.md
CHANGED
|
@@ -25,7 +25,11 @@ configs:
|
|
| 25 |
|
| 26 |
# DREAMS-AVATAR
|
| 27 |
|
| 28 |
-
**
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
|
| 30 |
32 calibrated, matted camera views of a full-body performance, with one SMPL-X body fitted
|
| 31 |
to all views at once by the
|
|
@@ -311,14 +315,14 @@ reprojection error rather than fitting a visual hull:
|
|
| 311 |
|
| 312 |
Sequence mode warm-starts frame `N` from frame `N-1` and only cold-starts the first frame of
|
| 313 |
a shard, so a ~1800 frame capture is practical. FLAME and joint offsets are off: this is
|
| 314 |
-
native SMPL-X, matching [DEGAS](https://initialneil.github.io/DEGAS
|
| 315 |
|
| 316 |
The `previews/` contact sheets and the 3-panel images above are the tracker's own QC output,
|
| 317 |
rendered on every 100th frame with the mesh rasterised through the true calibrated camera.
|
| 318 |
|
| 319 |
## Related
|
| 320 |
|
| 321 |
-
- [DEGAS](https://initialneil.github.io/DEGAS
|
| 322 |
- [Holistic-Tracker](https://github.com/initialneil/Holistic-Tracker): the monocular counterpart
|
| 323 |
|
| 324 |
## Licence
|
|
@@ -340,19 +344,16 @@ Please cite both this dataset and the DEGAS work the captures come from.
|
|
| 340 |
|
| 341 |
```bibtex
|
| 342 |
@misc{dreams-avatar-2026,
|
| 343 |
-
title = {DREAMS-AVATAR:
|
| 344 |
author = {Shao, Zhijing},
|
| 345 |
year = {2026},
|
| 346 |
url = {https://huggingface.co/datasets/initialneil/DREAMS-AVATAR}
|
| 347 |
}
|
| 348 |
|
| 349 |
-
@inproceedings{
|
| 350 |
title = {DEGAS: Detailed Expressions on Full-Body Gaussian Avatars},
|
| 351 |
-
author = {Shao, Zhijing and Wang, Duotun and Tian, Qing-Yao and Yang, Yao-Dong
|
| 352 |
-
and Meng, Hengyu and Cai, Zeyu and Dong, Bo and Zhang, Yu
|
| 353 |
-
and Zhang, Kang and Wang, Zeyu},
|
| 354 |
booktitle = {International Conference on 3D Vision (3DV)},
|
| 355 |
-
year = {2025}
|
| 356 |
-
url = {https://initialneil.github.io/DEGAS.html}
|
| 357 |
}
|
| 358 |
```
|
|
|
|
| 25 |
|
| 26 |
# DREAMS-AVATAR
|
| 27 |
|
| 28 |
+
**The DREAMS-Avatar dataset from the [DEGAS](https://initialneil.github.io/DEGAS) paper
|
| 29 |
+
(3DV 2025), re-registered to pure SMPL-X.**
|
| 30 |
+
|
| 31 |
+
These are the same multiview captures introduced as the *DREAMS-Avatar dataset* in DEGAS
|
| 32 |
+
(Fig. 1b); what is new here is the registration.
|
| 33 |
|
| 34 |
32 calibrated, matted camera views of a full-body performance, with one SMPL-X body fitted
|
| 35 |
to all views at once by the
|
|
|
|
| 315 |
|
| 316 |
Sequence mode warm-starts frame `N` from frame `N-1` and only cold-starts the first frame of
|
| 317 |
a shard, so a ~1800 frame capture is practical. FLAME and joint offsets are off: this is
|
| 318 |
+
native SMPL-X, matching [DEGAS](https://initialneil.github.io/DEGAS).
|
| 319 |
|
| 320 |
The `previews/` contact sheets and the 3-panel images above are the tracker's own QC output,
|
| 321 |
rendered on every 100th frame with the mesh rasterised through the true calibrated camera.
|
| 322 |
|
| 323 |
## Related
|
| 324 |
|
| 325 |
+
- [DEGAS](https://initialneil.github.io/DEGAS): Detailed Expressions on Full-Body Gaussian Avatars (3DV 2025), which introduced these captures as the DREAMS-Avatar dataset
|
| 326 |
- [Holistic-Tracker](https://github.com/initialneil/Holistic-Tracker): the monocular counterpart
|
| 327 |
|
| 328 |
## Licence
|
|
|
|
| 344 |
|
| 345 |
```bibtex
|
| 346 |
@misc{dreams-avatar-2026,
|
| 347 |
+
title = {DREAMS-AVATAR: the DREAMS-Avatar dataset re-registered to pure SMPL-X},
|
| 348 |
author = {Shao, Zhijing},
|
| 349 |
year = {2026},
|
| 350 |
url = {https://huggingface.co/datasets/initialneil/DREAMS-AVATAR}
|
| 351 |
}
|
| 352 |
|
| 353 |
+
@inproceedings{shao2025degas,
|
| 354 |
title = {DEGAS: Detailed Expressions on Full-Body Gaussian Avatars},
|
| 355 |
+
author = {Shao, Zhijing and Wang, Duotun and Tian, Qing-Yao and Yang, Yao-Dong and Meng, Hengyu and Cai, Zeyu and Dong, Bo and Zhang, Yu and Zhang, Kang and Wang, Zeyu},
|
|
|
|
|
|
|
| 356 |
booktitle = {International Conference on 3D Vision (3DV)},
|
| 357 |
+
year = {2025}
|
|
|
|
| 358 |
}
|
| 359 |
```
|