Weakly Supervised Volumetric Segmentation via Self-taught Shape Denoising Model

TL;DR

This paper introduces a weakly supervised 3D segmentation method that leverages shape priors and a novel learning strategy to improve medical image segmentation with minimal annotation effort.

Contribution

It proposes a self-taught shape denoising model that effectively captures 3D shape priors and integrates them into segmentation, reducing annotation costs.

Findings

Outperforms state-of-the-art methods on three organ segmentation benchmarks.

Achieves high accuracy with only 10% labeled slices.

Effective shape prior learning enhances segmentation quality.

Abstract

Weakly supervised segmentation is an important problem in medical image analysis due to the high cost of pixelwise annotation. Prior methods, while often focusing on weak labels of 2D images, exploit few structural cues of volumetric medical images. To address this, we propose a novel weakly-supervised segmentation strategy capable of better capturing 3D shape prior in both model prediction and learning. Our main idea is to extract a self-taught shape representation by leveraging weak labels, and then integrate this representation into segmentation prediction for shape refinement. To this end, we design a deep network consisting of a segmentation module and a shape denoising module, which are trained by an iterative learning strategy. Moreover, we introduce a weak annotation scheme with a hybrid label design for volumetric images, which improves model learning without increasing the overall annotation cost. The empirical experiments show that our approach outperforms existing SOTA strategies on three organ segmentation benchmarks with distinctive shape properties. Notably, we can achieve strong performance with even 10\% labeled slices, which is significantly superior to other methods.