Path Aggregation Network for Instance Segmentation

TLDR

Information propagation in neural networks is crucial for performance. The paper proposes PANet to boost information flow in proposal‑based instance segmentation. PANet enhances the feature hierarchy with bottom‑up path augmentation, adaptive feature pooling across levels, and a complementary branch to improve mask prediction. PANet is simple to implement, incurs minimal overhead, and achieves top performance on COCO 2017 instance segmentation, second on object detection, and state‑of‑the‑art results on MVD and Cityscapes.

Abstract

The way that information propagates in neural networks is of great importance. In this paper, we propose Path Aggregation Network (PANet) aiming at boosting information flow in proposal-based instance segmentation framework. Specifically, we enhance the entire feature hierarchy with accurate localization signals in lower layers by bottom-up path augmentation, which shortens the information path between lower layers and topmost feature. We present adaptive feature pooling, which links feature grid and all feature levels to make useful information in each level propagate directly to following proposal subnetworks. A complementary branch capturing different views for each proposal is created to further improve mask prediction. These improvements are simple to implement, with subtle extra computational overhead. Yet they are useful and make our PANet reach the 1st place in the COCO 2017 Challenge Instance Segmentation task and the 2nd place in Object Detection task without large-batch training. PANet is also state-of-the-art on MVD and Cityscapes.