RefineSeg: Dual Coarse-to-Fine Learning for Medical Image Segmentation

TL;DR

RefineSeg introduces a dual coarse-to-fine learning framework that effectively refines noisy, coarse annotations into accurate medical image segmentations using transition matrices, reducing annotation costs.

Contribution

The paper presents a novel matrix-based modeling approach for refining coarse annotations in medical image segmentation, enabling training with only noisy, coarse labels.

Findings

Outperforms existing weakly supervised methods

Achieves segmentation accuracy close to fully supervised models

Validated on multiple public and private datasets

Abstract

High-quality pixel-level annotations of medical images are essential for supervised segmentation tasks, but obtaining such annotations is costly and requires medical expertise. To address this challenge, we propose a novel coarse-to-fine segmentation framework that relies entirely on coarse-level annotations, encompassing both target and complementary drawings, despite their inherent noise. The framework works by introducing transition matrices in order to model the inaccurate and incomplete regions in the coarse annotations. By jointly training on multiple sets of coarse annotations, it progressively refines the network's outputs and infers the true segmentation distribution, achieving a robust approximation of precise labels through matrix-based modeling. To validate the flexibility and effectiveness of the proposed method, we demonstrate the results on two public cardiac imaging datasets, ACDC and MSCMRseg, and further evaluate its performance on the UK Biobank dataset. Experimental results indicate that our approach surpasses the state-of-the-art weakly supervised methods and closely matches the fully supervised approach.