FAPN: Feature Augmented Pyramid Network for polyp segmentation

Yanzhou Su, Jian Cheng, Murong Yi, Haijun Liu

Abstract: Highlights•We propose a feature augmented pyramid network (FAPN) for polyp segmentation.•A cross-embedding module is introduced to merge the multi-level feature.•A predictive calibration module is proposed to highlight feature of regions of interest.•A hierarchical feature fusion module is used to integrate multiple stages’ features.•FAPN achieves the state-of-the-art performance in five polyp segmentation benchmarks.

External IDs:dblp:journals/bspc/SuCYL22