Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.09245 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908875960090624 |
|---|---|
| author | Sun, Jiacheng Lin, Jiaqi Hu, Wenlong Li, Haoyang Zhou, Xinghong Mao, Chenghai Peng, Yan Li, Xiaomao |
| author_facet | Sun, Jiacheng Lin, Jiaqi Hu, Wenlong Li, Haoyang Zhou, Xinghong Mao, Chenghai Peng, Yan Li, Xiaomao |
| contents | One of the bottlenecks for instance segmentation today lies in the conflicting requirements of high-resolution inputs and lightweight, real-time inference. To address this bottleneck, we present a Polygon Detection Transformer (Poly-DETR) to reformulate instance segmentation as sparse vertex regression via Polar Representation, thereby eliminating the reliance on dense pixel-wise mask prediction. Considering the box-to-polygon reference shift in Detection Transformers, we propose Polar Deformable Attention and Position-Aware Training Scheme to dynamically update supervision and focus attention on boundary cues. Compared with state-of-the-art polar-based methods, Poly-DETR achieves a 4.7 mAP improvement on MS COCO test-dev. Moreover, we construct a parallel mask-based counterpart to support a systematic comparison between polar and mask representations. Experimental results show that Poly-DETR is more lightweight in high-resolution scenarios, reducing memory consumption by almost half on Cityscapes dataset. Notably, on PanNuke (cell segmentation) and SpaceNet (building footprints) datasets, Poly-DETR surpasses its mask-based counterpart on all metrics, which validates its advantage on regular-shaped instances in domain-specific settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_09245 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Towards Instance Segmentation with Polygon Detection Transformers Sun, Jiacheng Lin, Jiaqi Hu, Wenlong Li, Haoyang Zhou, Xinghong Mao, Chenghai Peng, Yan Li, Xiaomao Computer Vision and Pattern Recognition One of the bottlenecks for instance segmentation today lies in the conflicting requirements of high-resolution inputs and lightweight, real-time inference. To address this bottleneck, we present a Polygon Detection Transformer (Poly-DETR) to reformulate instance segmentation as sparse vertex regression via Polar Representation, thereby eliminating the reliance on dense pixel-wise mask prediction. Considering the box-to-polygon reference shift in Detection Transformers, we propose Polar Deformable Attention and Position-Aware Training Scheme to dynamically update supervision and focus attention on boundary cues. Compared with state-of-the-art polar-based methods, Poly-DETR achieves a 4.7 mAP improvement on MS COCO test-dev. Moreover, we construct a parallel mask-based counterpart to support a systematic comparison between polar and mask representations. Experimental results show that Poly-DETR is more lightweight in high-resolution scenarios, reducing memory consumption by almost half on Cityscapes dataset. Notably, on PanNuke (cell segmentation) and SpaceNet (building footprints) datasets, Poly-DETR surpasses its mask-based counterpart on all metrics, which validates its advantage on regular-shaped instances in domain-specific settings. |
| title | Towards Instance Segmentation with Polygon Detection Transformers |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.09245 |