Generating and Weighting Semantically Consistent Sample Pairs for Ultrasound Contrastive Learning
This work addresses the costly need for annotated medical datasets by improving generalization in ultrasound applications, though it is incremental as it builds on existing contrastive learning and meta-learning techniques.
The paper tackled the problem of domain gaps in medical ultrasound image analysis by proposing Meta-USCL, a self-supervised contrastive learning method that pre-trains deep neural networks on unlabeled ultrasound videos, achieving state-of-the-art results on tasks like pneumonia detection and breast cancer classification.
Well-annotated medical datasets enable deep neural networks (DNNs) to gain strong power in extracting lesion-related features. Building such large and well-designed medical datasets is costly due to the need for high-level expertise. Model pre-training based on ImageNet is a common practice to gain better generalization when the data amount is limited. However, it suffers from the domain gap between natural and medical images. In this work, we pre-train DNNs on ultrasound (US) domains instead of ImageNet to reduce the domain gap in medical US applications. To learn US image representations based on unlabeled US videos, we propose a novel meta-learning-based contrastive learning method, namely Meta Ultrasound Contrastive Learning (Meta-USCL). To tackle the key challenge of obtaining semantically consistent sample pairs for contrastive learning, we present a positive pair generation module along with an automatic sample weighting module based on meta-learning. Experimental results on multiple computer-aided diagnosis (CAD) problems, including pneumonia detection, breast cancer classification, and breast tumor segmentation, show that the proposed self-supervised method reaches state-of-the-art (SOTA). The codes are available at https://github.com/Schuture/Meta-USCL.