Augmentation for small object detection

TL;DR

This paper introduces a data augmentation technique that oversamples and copies small objects in images to improve small object detection performance, achieving significant gains on MS COCO.

Contribution

The authors propose a novel copy-pasting augmentation method that enhances small object detection by oversampling images with small objects, improving state-of-the-art results.

Findings

9.7% relative improvement in small object instance segmentation

7.1% improvement in small object detection

Effective augmentation strategy for small object detection

Abstract

In recent years, object detection has experienced impressive progress. Despite these improvements, there is still a significant gap in the performance between the detection of small and large objects. We analyze the current state-of-the-art model, Mask-RCNN, on a challenging dataset, MS COCO. We show that the overlap between small ground-truth objects and the predicted anchors is much lower than the expected IoU threshold. We conjecture this is due to two factors; (1) only a few images are containing small objects, and (2) small objects do not appear enough even within each image containing them. We thus propose to oversample those images with small objects and augment each of those images by copy-pasting small objects many times. It allows us to trade off the quality of the detector on large objects with that on small objects. We evaluate different pasting augmentation strategies, and ultimately, we achieve 9.7\% relative improvement on the instance segmentation and 7.1\% on the object detection of small objects, compared to the current state of the art method on MS COCO.