Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2502.05176 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909585446535168 |
|---|---|
| author | Wu, Chung-Ho Chen, Yang-Jung Chen, Ying-Huan Lee, Jie-Ying Ke, Bo-Hsu Mu, Chun-Wei Tuan Huang, Yi-Chuan Lin, Chin-Yang Chen, Min-Hung Lin, Yen-Yu Liu, Yu-Lun |
| author_facet | Wu, Chung-Ho Chen, Yang-Jung Chen, Ying-Huan Lee, Jie-Ying Ke, Bo-Hsu Mu, Chun-Wei Tuan Huang, Yi-Chuan Lin, Chin-Yang Chen, Min-Hung Lin, Yen-Yu Liu, Yu-Lun |
| contents | Three-dimensional scene inpainting is crucial for applications from virtual reality to architectural visualization, yet existing methods struggle with view consistency and geometric accuracy in 360° unbounded scenes. We present AuraFusion360, a novel reference-based method that enables high-quality object removal and hole filling in 3D scenes represented by Gaussian Splatting. Our approach introduces (1) depth-aware unseen mask generation for accurate occlusion identification, (2) Adaptive Guided Depth Diffusion, a zero-shot method for accurate initial point placement without requiring additional training, and (3) SDEdit-based detail enhancement for multi-view coherence. We also introduce 360-USID, the first comprehensive dataset for 360° unbounded scene inpainting with ground truth. Extensive experiments demonstrate that AuraFusion360 significantly outperforms existing methods, achieving superior perceptual quality while maintaining geometric accuracy across dramatic viewpoint changes. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_05176 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AuraFusion360: Augmented Unseen Region Alignment for Reference-based 360° Unbounded Scene Inpainting Wu, Chung-Ho Chen, Yang-Jung Chen, Ying-Huan Lee, Jie-Ying Ke, Bo-Hsu Mu, Chun-Wei Tuan Huang, Yi-Chuan Lin, Chin-Yang Chen, Min-Hung Lin, Yen-Yu Liu, Yu-Lun Computer Vision and Pattern Recognition Three-dimensional scene inpainting is crucial for applications from virtual reality to architectural visualization, yet existing methods struggle with view consistency and geometric accuracy in 360° unbounded scenes. We present AuraFusion360, a novel reference-based method that enables high-quality object removal and hole filling in 3D scenes represented by Gaussian Splatting. Our approach introduces (1) depth-aware unseen mask generation for accurate occlusion identification, (2) Adaptive Guided Depth Diffusion, a zero-shot method for accurate initial point placement without requiring additional training, and (3) SDEdit-based detail enhancement for multi-view coherence. We also introduce 360-USID, the first comprehensive dataset for 360° unbounded scene inpainting with ground truth. Extensive experiments demonstrate that AuraFusion360 significantly outperforms existing methods, achieving superior perceptual quality while maintaining geometric accuracy across dramatic viewpoint changes. |
| title | AuraFusion360: Augmented Unseen Region Alignment for Reference-based 360° Unbounded Scene Inpainting |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2502.05176 |