AI06 ALUDWIG: An Automated AI-Based Assessment of Female Androgenic Alopecia

The study introduces ALUDWIG, an automated AI-based tool designed to assess the severity of female androgenic alopecia (FAGA) more objectively and reproducibly than traditional methods like the Ludwig scale, which suffer from high intraobserver variability. The study analyzed 62 retrospective and 34 prospective images of FAGA, using convolutional models for head cropping and hair loss segmentation. The models were trained on a larger dataset and optimized using Bayesian techniques. The prospective evaluation showed an accuracy of 53% against investigator scores, with a correlation coefficient of 50% for alopecia percentage and 34% for categorical grade. The tool demonstrated potential for standardizing FAGA evaluation from a single image, offering a promising alternative to current scoring methods.