Segmentation Of Tongue Image Based On Saliency Detection

Tongue diagnosis of TCM(traditional Chinese medicine)is the main content of "look"diagnosis and a traditional diagnostic method which is the most characteristics of traditional Chinese medicine.Traditional Chinese medicine theory proposed that the human body is an organic and unified whole.The tongue image can reflect the physiological function and pathological changes of the human body.The traditional diagnosis of traditional Chinese medicine mainly relies on the individual observation of doctors,which may easily lead to the difficulty of quantification and the lack of objectivity and persuasiveness in the diagnosis of traditional Chinese medicine.Therefore,it restricts the modern development of traditional Chinese medicine diagnosis.In this paper,the tongue diagnosis information quantization was studied with the help of computer image processing algorithm,and the automatic segmentation of the tongue body is completed.The main work is as follows:1.Aiming at the problems of traditional automatic tongue segmentation and the deficiency of tongue image diagnosis in traditional Chinese medicine,this paper uses a method based on visual significance detection to conduct rough tongue segmentation.Firstly,the tongue image is preprocessed by adaptive histogram equalization with limited contrast.Then,on this basis,it is compared with many classical significance detection methods,and the significance detection method based on popular sorting is applied to the tongue image segmentation.Finally,the approximate region of the tongue image is obtained by threshold segmentation.The experimental results show that the algorithm can effectively achieve automatic rough segmentation of tongue region and the segmentation effect is satisfactory.2.Aiming at the problem of over-segmentation of tongue image in rough segmentation,this paper adopts the combination of visual significance detection technology and existing segmentation algorithm to accurately segment tongue image.Firstly,the algorithm iteratively runs GrabCut to segment the rough tongue segmentation results.After each iteration,the image morphology operation was used to obtain the input of the next iteration,and the final tongue body fine segmentation results were obtained through successive iterations.The results show that the proposed method has higher robustness and fewer iterations,and is significantly faster and more complete than the traditional method.