Light-based processing reportedly raised skin-disease image classification accuracy from 68% to 98%

Tech Xplore reports that light-based image processing raised classification accuracy from 68% to 98% in a skin-disease image task. It says the approach preserves color information that grayscale conversion can discard.