Semi-supervised Learning with Robust Loss in Brain Segmentation

TL;DR

This paper presents a semi-supervised deep learning approach for brain MRI segmentation that reduces labeling costs and incorporates robust loss to mitigate label noise, achieving competitive performance with fully supervised models.

Contribution

The work introduces a semi-supervised framework with robust loss for brain MRI segmentation, improving label efficiency and noise robustness.

Findings

Achieves competitive segmentation accuracy with less labeled data

Robust loss reduces impact of noisy labels

Framework lowers labeling costs for MRI datasets

Abstract

In this work, we used a semi-supervised learning method to train deep learning model that can segment the brain MRI images. The semi-supervised model uses less labeled data, and the performance is competitive with the supervised model with full labeled data. This framework could reduce the cost of labeling MRI images. We also introduced robust loss to reduce the noise effects of inaccurate labels generated in semi-supervised learning.