Artificial intelligence (AI) based system for the diagnosis and classification of scalp health: AI-ScalpGrader

Abstract

Abstract Many people suffer from scalp disorders but common treatment devices have faults such as inaccuracy of results and inconvenience of use. This study proposes a deep learning-based intelligent scalp diagnosis and classification system, named artificial intelligence (AI)-ScalpGrader. The proposed system consists of a portable scalp imaging device (ASM-202), a mobile device app, a cloud-based AI training server, and a cloud-based management platform. The instrumentation diagnoses and classifies ten scalp symptoms (normal, drying, oily, sensitivity, atopy, seborrheic, trouble, dry dandruff, oily dandruff, and hair loss) based on seven dermatologist-based indices (microkeratin, sebaceous, erythema between hair follicles, follicular erythema/pustules, dandruff, and hair loss) by AI-based characterization of the symptoms and indices. EfficientNet, a convolutional neural network (CNN) learning model, is MBConvolution composed with depthwise convolution, squeeze excitation, and width scaling and was adopted to diagnose and classify scalp conditions through retraining of images in the system. The results and verification on the reliability of AI-based data show that the system is able to diagnose and classify these symptoms and severity of the indices with accuracy values from 87.3 to 91.3%. Therefore, the AI-ScalpGrader is a novel approach to diagnose and classify scalp status.