A Deep Learning Classifier for the Analysis of Tiger Tail Banding in Trichothiodystrophy

The study developed a U-Net convolutional neural network to classify tiger tail banding in trichothiodystrophy (TTD) hair using polarized light microscopy images. The model was trained on 35 TTD and 35 normal hair images, with data augmentation techniques applied. The research aimed to improve the diagnosis of TTD, which is characterized by a distinct banding pattern in hair, and differentiate it from xeroderma pigmentosum (XP), which does not affect hair. Future studies planned to include a third class for XP/TTD complex patients to further analyze banding patterns. The use of deep learning was suggested to aid in diagnosing hair abnormalities.