A Critical Appraisal of Data Augmentation Methods for Imaging-Based Medical Diagnosis Applications
This addresses a critical issue for medical diagnosis applications, highlighting the need for tailored augmentations to avoid false positives or negatives, but it is incremental as it critiques existing methods without proposing a new solution.
The paper tackles the problem of sub-optimal data augmentation in medical imaging, finding that common intensity-based methods distort MRI scans and cause texture loss, negatively affecting classification performance.
Current data augmentation techniques and transformations are well suited for improving the size and quality of natural image datasets but are not yet optimized for medical imaging. We hypothesize that sub-optimal data augmentations can easily distort or occlude medical images, leading to false positives or negatives during patient diagnosis, prediction, or therapy/surgery evaluation. In our experimental results, we found that utilizing commonly used intensity-based data augmentation distorts the MRI scans and leads to texture information loss, thus negatively affecting the overall performance of classification. Additionally, we observed that commonly used data augmentation methods cannot be used with a plug-and-play approach in medical imaging, and requires manual tuning and adjustment.