Flip Learning: Erase to Segment

TL;DR

This paper introduces Flip Learning, a weakly-supervised segmentation method for breast ultrasound images that uses box annotations and erases regions to improve segmentation accuracy with minimal supervision.

Contribution

The study proposes a novel Flip Learning framework that erases regions at the superpixel level using reinforcement learning, with a dual-reward system and a coarse-to-fine strategy.

Findings

Achieves competitive segmentation performance on a large dataset.

Effectively narrows the gap between weakly-supervised and fully-supervised methods.

Demonstrates robustness and potential for clinical application.

Abstract

Nodule segmentation from breast ultrasound images is challenging yet essential for the diagnosis. Weakly-supervised segmentation (WSS) can help reduce time-consuming and cumbersome manual annotation. Unlike existing weakly-supervised approaches, in this study, we propose a novel and general WSS framework called Flip Learning, which only needs the box annotation. Specifically, the target in the label box will be erased gradually to flip the classification tag, and the erased region will be considered as the segmentation result finally. Our contribution is three-fold. First, our proposed approach erases on superpixel level using a Multi-agent Reinforcement Learning framework to exploit the prior boundary knowledge and accelerate the learning process. Second, we design two rewards: classification score and intensity distribution reward, to avoid under- and over-segmentation, respectively. Third, we adopt a coarse-to-fine learning strategy to reduce the residual errors and improve the segmentation performance. Extensively validated on a large dataset, our proposed approach achieves competitive performance and shows great potential to narrow the gap between fully-supervised and weakly-supervised learning.