An Expert Smart Scalp Inspection System Using Deep Learning

With the advent of the "beauty economic era", in which people are paying more attention to beauty and health, the health of the scalp is being increasingly valued. However, current scalp care services are limited by problems such as they are not automatic and objective, and the results are not significant, which make them unacceptable to the public. Because of these reasons, in this study, we focus on the obstacles that hairdressers face and propose an expert inspection system that is suitable for determining scalp problems by utilizing deep learning, cloud computing techniques, and an embedded system. Dandruff is the most common scalp problem. In this work, we propose a convolutional neural network (CNN)-based method to analyze the severity of dandruff and evaluate the health of the scalp. The convolutional block attention module (CBAM) is adopted to improve the feature extraction performance of the CNN model. The depth separable convolution (DSC) and spinal fully connected (FC) are applied in this work to reduce the number of model parameters. Aside from offering a more effective smart scalp inspection process, this method also enables hairdressers and customers to track their scalp problems easily. In the future, we expect to reduce the stress of hairdressers and enhance customers' trust on scalp care services by using the smart health inspection offered by this system. Last but not the least, it has been shown that the method proposed in this research can achieve an accuracy of 85.03%, which is higher than those achieved by recently proposed methods.