Classification and Identification of Male Hair Loss based on Deep Learning

In recent years, the public has paid more and more attention to the problem of hair loss, which not only leads to a negative impact on personal image to a certain extent but also may cause patients to have psychological inferiority feelings and even mental illness. Detecting the stages of hair loss in men is essential for treating hair loss. The Hamilton-Norwood classification is the most commonly used method to describe male hair loss. However, clinicians often estimate the stage of male hair loss by visual inspection combined with the Hamilton-Norwood classification scale, which is subjective and time-consuming. In this paper, three deep learning methods, VGG16, ResNet-50, and RegNet-64, were used to detect the stage of male hair loss automatically. The experimental results show that RegNet-64 achieves 93.08% accuracy in classifying male baldness, indicating that deep learning is a good choice for assessing the severity of male hair loss.