Multi-organ Self-supervised Contrastive Learning for Breast Lesion Segmentation
This work addresses the challenge of limited annotated data in medical imaging for breast lesion segmentation, though it is incremental in applying existing contrastive learning methods to multi-organ datasets.
The paper tackled the problem of breast tumor segmentation in ultrasound images by exploring multi-organ self-supervised contrastive learning for pre-training, resulting in improved performance over supervised baselines and achieving comparable results with only half the labeled data.
Self-supervised learning has proven to be an effective way to learn representations in domains where annotated labels are scarce, such as medical imaging. A widely adopted framework for this purpose is contrastive learning and it has been applied to different scenarios. This paper seeks to advance our understanding of the contrastive learning framework by exploring a novel perspective: employing multi-organ datasets for pre-training models tailored to specific organ-related target tasks. More specifically, our target task is breast tumour segmentation in ultrasound images. The pre-training datasets include ultrasound images from other organs, such as the lungs and heart, and large datasets of natural images. Our results show that conventional contrastive learning pre-training improves performance compared to supervised baseline approaches. Furthermore, our pre-trained models achieve comparable performance when fine-tuned with only half of the available labelled data. Our findings also show the advantages of pre-training on diverse organ data for improving performance in the downstream task.