Digital display division method of characteristic elements of back Sha images
Objective Through the study of the image data of the Sha,an objective method of digital display is proposed to accurately identify the characteristic elements of the Sha on the back,in order to avoid the deviation caused by different subjective cognition,help people quickly identify the relevant features of the Sha,provide a certain objective basis for the diagnosis of traditional Chinese medicine doctors,and reduce the subjective components.Methods First,the deep semantic segmentation algorithm model was used to segment the collected Sha images on the back.The purpose was to obtain only part of the area containing Sha and remove other irrelevant parts in the images,so as to minimize the impact on the subsequent experiments.Then,according to the segmented images,combined with the mapping area of the five viscera in the back of traditional Chinese medicine,the color features of the Sha were identified in three intervals by the method of key region detection.According to the traditional Chinese medicine subjective division of the shape of the Sha,the pixel statistics method was used to divide the shape of the Sha into flake and spot.Finally,the accuracy,precision,recall rate and F1 value were used to evaluate the objective identification results.Results The objective digital display method proposed in this paper achieved an accuracy rate of 80.56%in the color feature and 89.60%in the shape feature of the Sha images,which could accurately classify the feature elements of the Sha images and had certain feasibility.Conclusions The digital display method proposed in this paper can identify the color and shape features of Sha more accurately,and can avoid the different identification problems caused by subjective cognition to a large extent.It has a great application prospect in assisting doctors in diagnosis and teaching of traditional Chinese medicine image features recognition,and provides help for correctly understanding the feature information of Sha.
traditional Chinese medicine Shapromotion of information technologymachine learningfeature classificationimage processing
万方数据
维普
