ISDA: Position-Aware Instance Segmentation with Deformable Attention

23 Feb 2022  ·  Kaining Ying, Zhenhua Wang, Cong Bai, Pengfei Zhou ·

Most instance segmentation models are not end-to-end trainable due to either the incorporation of proposal estimation (RPN) as a pre-processing or non-maximum suppression (NMS) as a post-processing. Here we propose a novel end-to-end instance segmentation method termed ISDA. It reshapes the task into predicting a set of object masks, which are generated via traditional convolution operation with learned position-aware kernels and features of objects. Such kernels and features are learned by leveraging a deformable attention network with multi-scale representation. Thanks to the introduced set-prediction mechanism, the proposed method is NMS-free. Empirically, ISDA outperforms Mask R-CNN (the strong baseline) by 2.6 points on MS-COCO, and achieves leading performance compared with recent models. Code will be available soon.

PDF Abstract

Datasets


Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Instance Segmentation COCO test-dev ISDA (ResNet-50) APL 55.7 # 17
Instance Segmentation COCO test-dev ISDA (ours) mask AP 38.7 # 81
AP50 62 # 21
AP75 41.1 # 23
APS 17 # 33
APM 41.2 # 25

Methods