Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.19865 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917511241400320 |
|---|---|
| author | Chen, Yan-Ting Chen, Hao-Wei Hsiao, Tsu-Ching Lee, Chun-Yi |
| author_facet | Chen, Yan-Ting Chen, Hao-Wei Hsiao, Tsu-Ching Lee, Chun-Yi |
| contents | Accurate camera viewpoint estimation under sparse-view conditions remains challenging, particularly in two-view scenarios. Recent approaches leverage diffusion models such as Zero123 to synthesize novel views conditioned on relative viewpoint, showing promising results when repurposed for viewpoint estimation via optimization with MSE loss. However, existing methods often suffer from nonconvex loss landscape with numerous local minima, making them sensitive to initialization and reliant on naive multistart strategies. We analyze these optimization challenges and visualize failure cases, showing that geometric ambiguities, such as symmetry and self-similarity, can mislead gradient-based updates toward incorrect viewpoints. To address these limitations, we propose a score-based method that reshapes the optimization landscape to guide updates toward the ground-truth viewpoint, followed by a refinement stage using a viewpoint-conditioned diffusion model. Experiments show that our method improves convergence, reduces reliance on brute-force sampling, and achieves competitive accuracy with higher sample-efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_19865 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Landscape-Awareness for Geometric View Diffusion Model Chen, Yan-Ting Chen, Hao-Wei Hsiao, Tsu-Ching Lee, Chun-Yi Computer Vision and Pattern Recognition Accurate camera viewpoint estimation under sparse-view conditions remains challenging, particularly in two-view scenarios. Recent approaches leverage diffusion models such as Zero123 to synthesize novel views conditioned on relative viewpoint, showing promising results when repurposed for viewpoint estimation via optimization with MSE loss. However, existing methods often suffer from nonconvex loss landscape with numerous local minima, making them sensitive to initialization and reliant on naive multistart strategies. We analyze these optimization challenges and visualize failure cases, showing that geometric ambiguities, such as symmetry and self-similarity, can mislead gradient-based updates toward incorrect viewpoints. To address these limitations, we propose a score-based method that reshapes the optimization landscape to guide updates toward the ground-truth viewpoint, followed by a refinement stage using a viewpoint-conditioned diffusion model. Experiments show that our method improves convergence, reduces reliance on brute-force sampling, and achieves competitive accuracy with higher sample-efficiency. |
| title | Landscape-Awareness for Geometric View Diffusion Model |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.19865 |