Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2603.12759 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915859029557248 |
|---|---|
| author | Jiang, Lutao Cao, Zidong Chen, Weikai Zheng, Xu Lyu, Yuanhuiyi Li, Zhenyang HU, Zeyu Yin, Yingda Luo, Keyang Zhang, Runze Yan, Kai Qian, Shengju Fan, Haidi Peng, Yifan Wang, Xin Xiong, Hui Chen, Ying-Cong |
| author_facet | Jiang, Lutao Cao, Zidong Chen, Weikai Zheng, Xu Lyu, Yuanhuiyi Li, Zhenyang HU, Zeyu Yin, Yingda Luo, Keyang Zhang, Runze Yan, Kai Qian, Shengju Fan, Haidi Peng, Yifan Wang, Xin Xiong, Hui Chen, Ying-Cong |
| contents | Promptable instance segmentation is widely adopted in embodied and AR systems, yet the performance of foundation models trained on perspective imagery often degrades on 360° panoramas. In this paper, we introduce Segment Any 4K Panorama (SAP), a foundation model for 4K high-resolution panoramic instance-level segmentation. We reformulate panoramic segmentation as fixed-trajectory perspective video segmentation, decomposing a panorama into overlapping perspective patches sampled along a continuous spherical traversal. This memory-aligned reformulation preserves native 4K resolution while restoring the smooth viewpoint transitions required for stable cross-view propagation. To enable large-scale supervision, we synthesize 183,440 4K-resolution panoramic images with instance segmentation labels using the InfiniGen engine. Trained under this trajectory-aligned paradigm, SAP generalizes effectively to real-world 360° images, achieving +17.2 zero-shot mIoU gain over vanilla SAM2 of different sizes on real-world 4K panorama benchmark. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_12759 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SAP: Segment Any 4K Panorama Jiang, Lutao Cao, Zidong Chen, Weikai Zheng, Xu Lyu, Yuanhuiyi Li, Zhenyang HU, Zeyu Yin, Yingda Luo, Keyang Zhang, Runze Yan, Kai Qian, Shengju Fan, Haidi Peng, Yifan Wang, Xin Xiong, Hui Chen, Ying-Cong Computer Vision and Pattern Recognition Promptable instance segmentation is widely adopted in embodied and AR systems, yet the performance of foundation models trained on perspective imagery often degrades on 360° panoramas. In this paper, we introduce Segment Any 4K Panorama (SAP), a foundation model for 4K high-resolution panoramic instance-level segmentation. We reformulate panoramic segmentation as fixed-trajectory perspective video segmentation, decomposing a panorama into overlapping perspective patches sampled along a continuous spherical traversal. This memory-aligned reformulation preserves native 4K resolution while restoring the smooth viewpoint transitions required for stable cross-view propagation. To enable large-scale supervision, we synthesize 183,440 4K-resolution panoramic images with instance segmentation labels using the InfiniGen engine. Trained under this trajectory-aligned paradigm, SAP generalizes effectively to real-world 360° images, achieving +17.2 zero-shot mIoU gain over vanilla SAM2 of different sizes on real-world 4K panorama benchmark. |
| title | SAP: Segment Any 4K Panorama |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.12759 |