Add model card for JOintGS
#1
by
nielsr
HF Staff
- opened
README.md
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
pipeline_tag: image-to-3d
|
| 3 |
+
license: cc-by-nc-sa-4.0
|
| 4 |
+
---
|
| 5 |
+
|
| 6 |
+
# JOintGS: Joint Optimization of Cameras, Bodies and 3D Gaussians for In-the-Wild Monocular Reconstruction
|
| 7 |
+
|
| 8 |
+
This repository contains the pre-trained checkpoints for **JOintGS**, a unified framework that jointly optimizes camera extrinsics, human poses, and 3D Gaussian representations for robust, animatable 3D human avatar reconstruction from monocular video with coarse initialization.
|
| 9 |
+
|
| 10 |
+
- **Paper:** [JOintGS: Joint Optimization of Cameras, Bodies and 3D Gaussians for In-the-Wild Monocular Reconstruction](https://huggingface.co/papers/2602.04317)
|
| 11 |
+
- **Repository:** [https://github.com/MiliLab/JOintGS](https://github.com/MiliLab/JOintGS)
|
| 12 |
+
|
| 13 |
+
## Introduction
|
| 14 |
+
|
| 15 |
+
Reconstructing high-fidelity animatable 3D human avatars from monocular RGB videos remains challenging, particularly in unconstrained in-the-wild scenarios where camera parameters and human poses from off-the-shelf methods (e.g., COLMAP, HMR2.0) are often inaccurate.
|
| 16 |
+
|
| 17 |
+
JOintGS enables a synergistic refinement mechanism where explicit foreground-background disentanglement allows mutual reinforcement: static background Gaussians anchor camera estimation via multi-view consistency; refined cameras improve human body alignment through accurate temporal correspondence; and optimized human poses enhance scene reconstruction by removing dynamic artifacts from static constraints.
|
| 18 |
+
|
| 19 |
+
## License
|
| 20 |
+
|
| 21 |
+
- **Source Code**: The software in the associated GitHub repository is licensed under the [MIT License](https://github.com/MiliLab/JOintGS/blob/main/LICENSE).
|
| 22 |
+
- **Model Weights**: The pre-trained checkpoints in this repository are released under the [CC BY-NC-SA 4.0 License](https://creativecommons.org/licenses/by-nc-sa/4.0/).
|
| 23 |
+
|
| 24 |
+
## Citation
|
| 25 |
+
|
| 26 |
+
```bibtex
|
| 27 |
+
@article{jointgs2026,
|
| 28 |
+
title={JOintGS: Joint Optimization of Cameras, Bodies and 3D Gaussians for In-the-Wild Monocular Reconstruction},
|
| 29 |
+
author={Jiuhai Chen and others},
|
| 30 |
+
journal={arXiv preprint arXiv:2602.04317},
|
| 31 |
+
year={2026}
|
| 32 |
+
}
|
| 33 |
+
```
|