Reliable detection of eczema areas for fully automated assessment of eczema severity from digital camera images

Assessing the severity of eczema in clinical research requires face-to-face skin examination by trained staff. Such approaches are resource-intensive for participants and staff, challenging during pandemics, prone to inter- intra-observer variation. Computer vision algorithms have been proposed automate assessment using digital camera images. However, they often require human intervention detect lesions cannot automatically assess from real-world images an end-to-end pipeline.We developed a model data augmentation pixel-level segmentation on 1345 provided dermatologists. We evaluated quality obtained compared that clinicians, robustness varying imaging conditions encountered real-life images, such as lighting, focus, blur performance downstream prediction when detected lesions.The lesion detection increased approximately 25% 40%, respectively, our previous model. The remained unchanged. Use alternative specialist labelling showed par with is used.