Principled Ultrasound Data Augmentation for Classification of Standard Planes

TL;DR

This paper introduces a systematic augmentation policy search for ultrasound image classification, significantly improving model performance and representation quality by optimizing data augmentation strategies.

Contribution

It proposes a principled augmentation policy search method that includes domain-specific transformations and non-linear strategies, enhancing ultrasound image classification accuracy.

Findings

Achieved an average F1-score improvement of 7.0% over naive augmentation.

Learned augmentations lead to better clustered and defined ultrasound image representations.

Principled augmentation strategies outperform ad-hoc methods in medical image analysis.

Abstract

Deep learning models with large learning capacities often overfit to medical imaging datasets. This is because training sets are often relatively small due to the significant time and financial costs incurred in medical data acquisition and labelling. Data augmentation is therefore often used to expand the availability of training data and to increase generalization. However, augmentation strategies are often chosen on an ad-hoc basis without justification. In this paper, we present an augmentation policy search method with the goal of improving model classification performance. We include in the augmentation policy search additional transformations that are often used in medical image analysis and evaluate their performance. In addition, we extend the augmentation policy search to include non-linear mixed-example data augmentation strategies. Using these learned policies, we show that principled data augmentation for medical image model training can lead to significant improvements in ultrasound standard plane detection, with an an average F1-score improvement of 7.0% overall over naive data augmentation strategies in ultrasound fetal standard plane classification. We find that the learned representations of ultrasound images are better clustered and defined with optimized data augmentation.