PolyTransform: Deep Polygon Transformer for Instance Segmentation
This work addresses the problem of accurate instance segmentation for applications like autonomous driving, but it is incremental as it builds on existing segmentation and polygon methods.
The paper tackles instance segmentation by combining segmentation networks with polygon-based methods to produce precise, geometry-preserving masks, achieving top performance on the Cityscapes dataset with significant improvements over the backbone network.
In this paper, we propose PolyTransform, a novel instance segmentation algorithm that produces precise, geometry-preserving masks by combining the strengths of prevailing segmentation approaches and modern polygon-based methods. In particular, we first exploit a segmentation network to generate instance masks. We then convert the masks into a set of polygons that are then fed to a deforming network that transforms the polygons such that they better fit the object boundaries. Our experiments on the challenging Cityscapes dataset show that our PolyTransform significantly improves the performance of the backbone instance segmentation network and ranks 1st on the Cityscapes test-set leaderboard. We also show impressive gains in the interactive annotation setting. We release the code at https://github.com/uber-research/PolyTransform.