The Efficacy of Semantics-Preserving Transformations in Self-Supervised Learning for Medical Ultrasound

2025-04-10

AI Summary

This work addresses the lack of semantic alignment in data augmentation and preprocessing strategies for self-supervised learning on lung ultrasound (LUS) images. We propose the first semantics-preserving augmentation pipeline specifically designed for ultrasound imagery and systematically evaluate three augmentation paradigms—semantics-preserving, cropping-based, and ultrasound-specific preprocessing—across three downstream tasks: B-line detection, pleural effusion identification, and COVID-19 classification. Results demonstrate that semantics-preserving augmentation significantly improves performance on global-context-dependent tasks (e.g., COVID-19 classification), whereas cropping-based augmentation benefits local pattern recognition (e.g., B-lines and effusion detection). Ultrasound-specific preprocessing consistently enhances all downstream tasks. Our study establishes an interpretable, task-adaptive augmentation design paradigm for self-supervised learning in medical ultrasound, advancing both methodological rigor and clinical applicability.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Data augmentation is a central component of joint embedding self-supervised learning (SSL). Approaches that work for natural images may not always be effective in medical imaging tasks. This study systematically investigated the impact of data augmentation and preprocessing strategies in SSL for lung ultrasound. Three data augmentation pipelines were assessed: (1) a baseline pipeline commonly used across imaging domains, (2) a novel semantic-preserving pipeline designed for ultrasound, and (3) a distilled set of the most effective transformations from both pipelines. Pretrained models were evaluated on multiple classification tasks: B-line detection, pleural effusion detection, and COVID-19 classification. Experiments revealed that semantics-preserving data augmentation resulted in the greatest performance for COVID-19 classification - a diagnostic task requiring global image context. Cropping-based methods yielded the greatest performance on the B-line and pleural effusion object classification tasks, which require strong local pattern recognition. Lastly, semantics-preserving ultrasound image preprocessing resulted in increased downstream performance for multiple tasks. Guidance regarding data augmentation and preprocessing strategies was synthesized for practitioners working with SSL in ultrasound.

Problem

• Evaluating data augmentation impact on lung ultrasound SSL
• Comparing semantics-preserving vs. standard augmentation for medical imaging
• Optimizing preprocessing for COVID-19 and local pattern classification tasks

Innovation

• Semantics-preserving data augmentation for ultrasound
• Cropping-based methods for local pattern recognition
• Ultrasound image preprocessing boosts performance