AR-1-to-3: Single Image to Consistent 3D Object Generation via Next-View Prediction
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916889081413632 |
|---|---|
| author | Zhang, Xuying Zhou, Yupeng Wang, Kai Wang, Yikai Li, Zhen Jiao, Shaohui Zhou, Daquan Hou, Qibin Cheng, Ming-Ming |
| author_facet | Zhang, Xuying Zhou, Yupeng Wang, Kai Wang, Yikai Li, Zhen Jiao, Shaohui Zhou, Daquan Hou, Qibin Cheng, Ming-Ming |
| contents | Novel view synthesis (NVS) is a cornerstone for image-to-3d creation. However, existing works still struggle to maintain consistency between the generated views and the input views, especially when there is a significant camera pose difference, leading to poor-quality 3D geometries and textures. We attribute this issue to their treatment of all target views with equal priority according to our empirical observation that the target views closer to the input views exhibit higher fidelity. With this inspiration, we propose AR-1-to-3, a novel next-view prediction paradigm based on diffusion models that first generates views close to the input views, which are then utilized as contextual information to progressively synthesize farther views. To encode the generated view subsequences as local and global conditions for the next-view prediction, we accordingly develop a stacked local feature encoding strategy (Stacked-LE) and an LSTM-based global feature encoding strategy (LSTM-GE). Extensive experiments demonstrate that our method significantly improves the consistency between the generated views and the input views, producing high-fidelity 3D assets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_12929 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AR-1-to-3: Single Image to Consistent 3D Object Generation via Next-View Prediction Zhang, Xuying Zhou, Yupeng Wang, Kai Wang, Yikai Li, Zhen Jiao, Shaohui Zhou, Daquan Hou, Qibin Cheng, Ming-Ming Computer Vision and Pattern Recognition Novel view synthesis (NVS) is a cornerstone for image-to-3d creation. However, existing works still struggle to maintain consistency between the generated views and the input views, especially when there is a significant camera pose difference, leading to poor-quality 3D geometries and textures. We attribute this issue to their treatment of all target views with equal priority according to our empirical observation that the target views closer to the input views exhibit higher fidelity. With this inspiration, we propose AR-1-to-3, a novel next-view prediction paradigm based on diffusion models that first generates views close to the input views, which are then utilized as contextual information to progressively synthesize farther views. To encode the generated view subsequences as local and global conditions for the next-view prediction, we accordingly develop a stacked local feature encoding strategy (Stacked-LE) and an LSTM-based global feature encoding strategy (LSTM-GE). Extensive experiments demonstrate that our method significantly improves the consistency between the generated views and the input views, producing high-fidelity 3D assets. |
| title | AR-1-to-3: Single Image to Consistent 3D Object Generation via Next-View Prediction |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.12929 |