Quantitative Imaging Principles Improves Medical Image Learning

TL;DR

This paper introduces a novel self-supervised learning approach that incorporates quantitative imaging principles to enhance medical image quality and biological accuracy, leading to improved downstream performance and validation on clinical software.

Contribution

It proposes integrating physics-based quantitative imaging principles into SSL for medical images, which is a new approach to improve image quality and biological relevance.

Findings

Better initial states for supervised training with limited data

Generated images validate on clinical quantitative analysis software

Enhanced downstream task performance

Abstract

Fundamental differences between natural and medical images have recently favored the use of self-supervised learning (SSL) over ImageNet transfer learning for medical image applications. Differences between image types are primarily due to the imaging modality and medical images utilize a wide range of physics based techniques while natural images are captured using only visible light. While many have demonstrated that SSL on medical images has resulted in better downstream task performance, our work suggests that more performance can be gained. The scientific principles which are used to acquire medical images are not often considered when constructing learning problems. For this reason, we propose incorporating quantitative imaging principles during generative SSL to improve image quality and quantitative biological accuracy. We show that this training schema results in better starting states for downstream supervised training on limited data. Our model also generates images that validate on clinical quantitative analysis software.