Extended Feature Pyramid Network for Small Object Detection

TL;DR

This paper introduces EFPN, an enhanced feature pyramid network with a high-resolution level and a feature texture transfer module, significantly improving small object detection accuracy while maintaining efficiency.

Contribution

The paper proposes EFPN with a novel FTT module and a balanced loss function, advancing small object detection by addressing feature coupling and area imbalance issues.

Findings

Achieves state-of-the-art results on Tsinghua-Tencent 100K

Outperforms existing methods on MS COCO small object categories

Maintains efficiency in computation and memory usage

Abstract

Small object detection remains an unsolved challenge because it is hard to extract information of small objects with only a few pixels. While scale-level corresponding detection in feature pyramid network alleviates this problem, we find feature coupling of various scales still impairs the performance of small objects. In this paper, we propose extended feature pyramid network (EFPN) with an extra high-resolution pyramid level specialized for small object detection. Specifically, we design a novel module, named feature texture transfer (FTT), which is used to super-resolve features and extract credible regional details simultaneously. Moreover, we design a foreground-background-balanced loss function to alleviate area imbalance of foreground and background. In our experiments, the proposed EFPN is efficient on both computation and memory, and yields state-of-the-art results on small traffic-sign dataset Tsinghua-Tencent 100K and small category of general object detection dataset MS COCO.