Enhancing Weakly Supervised 3D Medical Image Segmentation through Probabilistic-aware Learning

TL;DR

This paper introduces a probabilistic-aware weakly supervised learning pipeline for 3D medical image segmentation, combining pseudo label generation, probabilistic self-attention, and confidence-based loss to improve accuracy with sparse annotations.

Contribution

The novel pipeline integrates probabilistic pseudo labels, a probabilistic transformer network, and a confidence-aware loss for improved weakly supervised 3D segmentation.

Findings

Achieves up to 18.1% improvement in Dice scores over existing methods.

Rivals fully supervised segmentation performance.

Effective on CT and MRI datasets.

Abstract

3D medical image segmentation is a challenging task with crucial implications for disease diagnosis and treatment planning. Recent advances in deep learning have significantly enhanced fully supervised medical image segmentation. However, this approach heavily relies on labor-intensive and time-consuming fully annotated ground-truth labels, particularly for 3D volumes. To overcome this limitation, we propose a novel probabilistic-aware weakly supervised learning pipeline, specifically designed for 3D medical imaging. Our pipeline integrates three innovative components: a Probability-based Pseudo Label Generation technique for synthesizing dense segmentation masks from sparse annotations, a Probabilistic Multi-head Self-Attention network for robust feature extraction within our Probabilistic Transformer Network, and a Probability-informed Segmentation Loss Function to enhance training with annotation confidence. Demonstrating significant advances, our approach not only rivals the performance of fully supervised methods but also surpasses existing weakly supervised methods in CT and MRI datasets, achieving up to 18.1% improvement in Dice scores for certain organs. The code is available at https://github.com/runminjiang/PW4MedSeg.