Research on typical feature recognition method of tongue diagnosis images in traditional Chinese medicine

Combined with the demands of modernization of tongue diagnosis in traditional Chinese medicine, this paper studied how to identify the moss color and thickness. An algorithm based on RGB and HSV combined color space analysis of moss color is proposed. It used the RGB color space to identify white and gray fur, and the HSV color space to identify yellow fur. This paper combines the characteristics of different color spaces to improve the recognition rate and accuracy. Experiments show that the algorithm has higher recognition rate than the single color space algorithm in the time-acceptable range. From the perspective of traditional Chinese medicine diagnosis, this paper designed a small particle ratio method to identify the thickness of tongue coatings. By binarizing the image, the proportion of fine particles was used to characterize the thickness of tongue fur. Convert three-dimensional features into two-dimensional features, and then extract one-dimensional values. Experiments show that compared with the gray level co-occurrence matrix method, the tongue coatings ratio method has a higher recognition rate in the time-acceptable range.