PolarMask: Single Shot Instance Segmentation with Polar Representation

In this paper, we introduce an anchor-box free and single shot instance segmentation method, which is conceptually simple, fully convolutional and can be used as a mask prediction module for instance segmentation, by easily embedding it into most off-the-shelf detection methods. Our method, termed PolarMask, formulates the instance segmentation problem as instance center classification and dense distance regression in a polar coordinate. Moreover, we propose two effective approaches to deal with sampling high-quality center examples and optimization for dense distance regression, respectively, which can significantly improve the performance and simplify the training process. Without any bells and whistles, PolarMask achieves 32.9% in mask mAP with single-model and single-scale training/testing on challenging COCO dataset. For the first time, we demonstrate a much simpler and flexible instance segmentation framework achieving competitive accuracy. We hope that the proposed PolarMask framework can serve as a fundamental and strong baseline for single shot instance segmentation tasks. Code is available at: github.com/xieenze/PolarMask.

PDF Abstract CVPR 2020 PDF CVPR 2020 Abstract


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Instance Segmentation COCO test-dev PolarMask (ResNeXt-101-FPN) mask AP 32.9% # 100
AP50 55.4% # 35
AP75 33.8% # 32
APS 15.5% # 36
APM 35.1% # 33
APL 46.3% # 33
Instance Segmentation COCO test-dev PolarMask (ResNet-101-FPN) mask AP 30.4% # 101
AP50 51.9% # 38
AP75 31% # 33
APS 13.4% # 37
APM 32.4% # 34
APL 42.8% # 34