Deep-Learning-Based Scalp Image Analysis Using Limited Data

The World Health Organization and Korea National Insurance assert that the number of alopecia patients is increasing every year, approximately 70 percent adults suffer from scalp problems. Although a genetic problem, it difficult to diagnose at an early stage. deep-learning-based approaches have been effective for medical image analyses, challenging generate deep learning models detection analysis because creating dataset challenging. In this paper, we present approach generating model specialized achieves high accuracy by applying data preprocessing, augmentation, ensemble analyses. We use containing 526 good, 13,156 mild, 3742 moderate, 825 severe images. was further augmented normalization, geometry-based augmentation (rotate, vertical flip, horizontal crop, affine transformation), PCA augmentation. compare performance single using ResNet, ResNeXt, DenseNet, XceptionNet, ensembles these models. best result achieved when ResNet were combined achieve 95.75 F1 score 87.05.