Abstract
Facial diagnosis is an important and very intuitive diagnostic method in Traditional Chinese Medicine (TCM). However, due to its qualitative and experience-based subjective property, traditional facial diagnosis has a certain limitation in clinical medicine, even in modern information diagnosis. The computerized inspection method provides quantitative model to evaluate facial complexion. This paper proposes the methodology of facial color analysis for diagnosis in TCM. Firstly, five facial parts (forehead, nose, two cheeks and jaw) are segmented and extracted with both global and local color features. Then, the dimensionality reduction method is applied for low dimensional representation of facial parts. Furthermore, four typical classifiers are evaluated with their recognition performances under these given facial parts. Finally, these five parts would be weighted fusion with their best classification results for higher performance. Experiments have shown that the local feature is valuable in complexion recognition, and facial cheeks are most discriminative parts which is consistent with the TCM doctors experience. Moreover, the fusion of parts can make the recognition rate higher than any single part.