NOVA3R: Non-pixel-aligned Visual Transformer for Amodal 3D Reconstruction
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914371569975296 |
|---|---|
| author | Chen, Weirong Zheng, Chuanxia Zhang, Ganlin Vedaldi, Andrea Cremers, Daniel |
| author_facet | Chen, Weirong Zheng, Chuanxia Zhang, Ganlin Vedaldi, Andrea Cremers, Daniel |
| contents | We present NOVA3R, an effective approach for non-pixel-aligned 3D reconstruction from a set of unposed images in a feed-forward manner. Unlike pixel-aligned methods that tie geometry to per-ray predictions, our formulation learns a global, view-agnostic scene representation that decouples reconstruction from pixel alignment. This addresses two key limitations in pixel-aligned 3D: (1) it recovers both visible and invisible points with a complete scene representation, and (2) it produces physically plausible geometry with fewer duplicated structures in overlapping regions. To achieve this, we introduce a scene-token mechanism that aggregates information across unposed images and a diffusion-based 3D decoder that reconstructs complete, non-pixel-aligned point clouds. Extensive experiments on both scene-level and object-level datasets demonstrate that NOVA3R outperforms state-of-the-art methods in terms of reconstruction accuracy and completeness. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_04179 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | NOVA3R: Non-pixel-aligned Visual Transformer for Amodal 3D Reconstruction Chen, Weirong Zheng, Chuanxia Zhang, Ganlin Vedaldi, Andrea Cremers, Daniel Computer Vision and Pattern Recognition We present NOVA3R, an effective approach for non-pixel-aligned 3D reconstruction from a set of unposed images in a feed-forward manner. Unlike pixel-aligned methods that tie geometry to per-ray predictions, our formulation learns a global, view-agnostic scene representation that decouples reconstruction from pixel alignment. This addresses two key limitations in pixel-aligned 3D: (1) it recovers both visible and invisible points with a complete scene representation, and (2) it produces physically plausible geometry with fewer duplicated structures in overlapping regions. To achieve this, we introduce a scene-token mechanism that aggregates information across unposed images and a diffusion-based 3D decoder that reconstructs complete, non-pixel-aligned point clouds. Extensive experiments on both scene-level and object-level datasets demonstrate that NOVA3R outperforms state-of-the-art methods in terms of reconstruction accuracy and completeness. |
| title | NOVA3R: Non-pixel-aligned Visual Transformer for Amodal 3D Reconstruction |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.04179 |