Tongue Image Identification System on Congestion of Fungiform Papillae (CFP)

摘要

Tongue diagnosis is a unique and important diagnostic method in Traditional Chinese Medicine (TCM). It is used to observe abnormal changes in the tongue for identifying diseases. However, due to its qualitative, subjective and experience-based nature, traditional tongue diagnosis has very limited applications in clinical medicine. To date, no work has sought to automatically recognize a distinctive diagnostic and textural feature of the tongue, Congestion of Fungiform Papillae (CFP) on the middle of tongue surface. In this paper, we present a novel computerized tongue inspection system for identifying the presence or absence of CFP images. We first define and partition a region of interest (ROI) for texture acquisition. After preprocessing for reflective points, we apply the Gabor filter banks and Bayesian Network to identify the texture blocks. Finally, a fusion strategy is employed to model the relationship between the image classes and the presence of CFP. The results are promising.