Traditional Chinese medicine face color identifying and retrieving method based on image analysis

Abstract

The invention provides a traditional Chinese medicine face color identifying and retrieving method based on image analysis, which belongs to the field of image analysis and identification. According to the invention, a retrieving and identifying platform facing traditional Chinese medicine face diagnosis is designed in view of strong subjective dependence and lack of objective quantitative basis in the traditional Chinese medicine face diagnosis. According to the invention, the method is characterized by comprising the following steps of: dividing human face through technologies like human face detection feature positioning when the user inputs a face image to be inquired; extracting the face colors of corresponding areas of internal organs; inputting characteristic vector of the face image to be inquired into a classifier so as to obtain a face color identification result in the identification module; and calculating the similarity of the characteristic vector of the face image to be inquired and the data in a face image characteristic database in the retrieving module; sequencing the characteristic vectors according to the similarity from large to small; returning the similar face image; and giving disease description of the similar face image. According to the invention, the precision rate and recall ratio are about 70% and 65%; and certain reference value is obtained.

Description

technical field [0001] The present invention takes the facial image of traditional Chinese medicine as the research object, according to the corresponding relationship between the complexion and the health status of viscera, uses computer vision theory and image analysis and recognition technology to extract the complexion information of the input facial image, and retrieves it from the database according to the similarity of the color feature vector Similar faces can be found, and qualitative diagnosis results of similar faces can be given, providing an objective identification and retrieval platform. Background technique [0002] The essence of traditional Chinese medicine lies in "treatment based on syndrome differentiation", and the four diagnostic methods of "look, smell, ask, and feel" are the important basis for completing "treatment based on syndrome differentiation". "Watch" is one of the most important contents of the four diagnostic methods of traditional Chinese ...

AI Technical Summary

Problems solved by technology

Traditional medical image retrieval methods usually use manual annotation based on text information, which is time-consuming and laborious, and the annotation results vary from person to person

When searching based on text information, due to the different understanding of image content by individuals, the retrieval results are often not ideal, which seriously affects the role of images in the diagnosis process.

Images