Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.07986 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913018421444608 |
|---|---|
| author | Chen, Tingxi Cheng, Zhengxue Zhong, Houqiang Wang, Su Xie, Rong Song, Li |
| author_facet | Chen, Tingxi Cheng, Zhengxue Zhong, Houqiang Wang, Su Xie, Rong Song, Li |
| contents | Egocentric video is crucial for next-generation 4D scene reconstruction, with applications in AR/VR and embodied AI. However, reconstructing dynamic first-person scenes is challenging due to complex ego-motion, occlusions, and hand-object interactions. Existing decomposition methods are ill-suited, assuming fixed viewpoints or merging dynamics into a single foreground. To address these limitations, we introduce DP-DeGauss, a dynamic probabilistic Gaussian decomposition framework for egocentric 4D reconstruction. Our method initializes a unified 3D Gaussian set from COLMAP priors, augments each with a learnable category probability, and dynamically routes them into specialized deformation branches for background, hands, or object modeling. We employ category-specific masks for better disentanglement and introduce brightness and motion-flow control to improve static rendering and dynamic reconstruction. Extensive experiments show that DP-DeGauss outperforms baselines by +1.70dB in PSNR on average with SSIM and LPIPS gains. More importantly, our framework achieves the first and state-of-the-art disentanglement of background, hand, and object components, enabling explicit, fine-grained separation, paving the way for more intuitive ego scene understanding and editing. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_07986 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DP-DeGauss: Dynamic Probabilistic Gaussian Decomposition for Egocentric 4D Scene Reconstruction Chen, Tingxi Cheng, Zhengxue Zhong, Houqiang Wang, Su Xie, Rong Song, Li Computer Vision and Pattern Recognition Egocentric video is crucial for next-generation 4D scene reconstruction, with applications in AR/VR and embodied AI. However, reconstructing dynamic first-person scenes is challenging due to complex ego-motion, occlusions, and hand-object interactions. Existing decomposition methods are ill-suited, assuming fixed viewpoints or merging dynamics into a single foreground. To address these limitations, we introduce DP-DeGauss, a dynamic probabilistic Gaussian decomposition framework for egocentric 4D reconstruction. Our method initializes a unified 3D Gaussian set from COLMAP priors, augments each with a learnable category probability, and dynamically routes them into specialized deformation branches for background, hands, or object modeling. We employ category-specific masks for better disentanglement and introduce brightness and motion-flow control to improve static rendering and dynamic reconstruction. Extensive experiments show that DP-DeGauss outperforms baselines by +1.70dB in PSNR on average with SSIM and LPIPS gains. More importantly, our framework achieves the first and state-of-the-art disentanglement of background, hand, and object components, enabling explicit, fine-grained separation, paving the way for more intuitive ego scene understanding and editing. |
| title | DP-DeGauss: Dynamic Probabilistic Gaussian Decomposition for Egocentric 4D Scene Reconstruction |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2604.07986 |