ComPC: Completing a 3D Point Cloud with 2D Diffusion Priors
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912201618489344 |
|---|---|
| author | Huang, Tianxin Yan, Zhiwen Zhao, Yuyang Lee, Gim Hee |
| author_facet | Huang, Tianxin Yan, Zhiwen Zhao, Yuyang Lee, Gim Hee |
| contents | 3D point clouds directly collected from objects through sensors are often incomplete due to self-occlusion. Conventional methods for completing these partial point clouds rely on manually organized training sets and are usually limited to object categories seen during training. In this work, we propose a test-time framework for completing partial point clouds across unseen categories without any requirement for training. Leveraging point rendering via Gaussian Splatting, we develop techniques of Partial Gaussian Initialization, Zero-shot Fractal Completion, and Point Cloud Extraction that utilize priors from pre-trained 2D diffusion models to infer missing regions and extract uniform completed point clouds. Experimental results on both synthetic and real-world scanned point clouds demonstrate that our approach outperforms existing methods in completing a variety of objects. Our project page is at \url{https://tianxinhuang.github.io/projects/ComPC/}. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_06814 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | ComPC: Completing a 3D Point Cloud with 2D Diffusion Priors Huang, Tianxin Yan, Zhiwen Zhao, Yuyang Lee, Gim Hee Computer Vision and Pattern Recognition 3D point clouds directly collected from objects through sensors are often incomplete due to self-occlusion. Conventional methods for completing these partial point clouds rely on manually organized training sets and are usually limited to object categories seen during training. In this work, we propose a test-time framework for completing partial point clouds across unseen categories without any requirement for training. Leveraging point rendering via Gaussian Splatting, we develop techniques of Partial Gaussian Initialization, Zero-shot Fractal Completion, and Point Cloud Extraction that utilize priors from pre-trained 2D diffusion models to infer missing regions and extract uniform completed point clouds. Experimental results on both synthetic and real-world scanned point clouds demonstrate that our approach outperforms existing methods in completing a variety of objects. Our project page is at \url{https://tianxinhuang.github.io/projects/ComPC/}. |
| title | ComPC: Completing a 3D Point Cloud with 2D Diffusion Priors |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2404.06814 |