Automating Hair Loss Labels for Universally Scoring Alopecia from Images

The document discusses a study that developed an artificial intelligence system, HairComb, to automatically score alopecia severity from images. The system was trained on 1,000 images and tested on 200 images, showing a high level of agreement with scores given by dermatologists. The study involved 404 participants and the algorithm showed a 92% accuracy rate. HairComb uses a convolutional neural network to determine hair loss patterns and hair density, providing a percentage hair loss at every pixel. The algorithm was validated as a standalone tool for any alopecia subtype and showed strong agreement with manual scores. The study concludes that HairComb could be a useful tool in assessing alopecia severity, potentially improving diagnosis and treatment planning.