Rethinking Ultrasound Augmentation: A Physics-Inspired Approach
This addresses the challenge of limited and artifact-prone ultrasound data for training deep learning models in medical imaging, though it is incremental as it builds on existing augmentation methods.
The authors tackled the problem of unrealistic ultrasound images from standard data augmentation by proposing physics-inspired transformations, achieving a 7% improvement in bone segmentation Dice score and a 5% increase in classification accuracy on a new spine ultrasound dataset.
Medical Ultrasound (US), despite its wide use, is characterized by artifacts and operator dependency. Those attributes hinder the gathering and utilization of US datasets for the training of Deep Neural Networks used for Computer-Assisted Intervention Systems. Data augmentation is commonly used to enhance model generalization and performance. However, common data augmentation techniques, such as affine transformations do not align with the physics of US and, when used carelessly can lead to unrealistic US images. To this end, we propose a set of physics-inspired transformations, including deformation, reverb and Signal-to-Noise Ratio, that we apply on US B-mode images for data augmentation. We evaluate our method on a new spine US dataset for the tasks of bone segmentation and classification.