Tongue quality and tongue coating separation method based on LGBM model

Abstract

The invention discloses a tongue quality and tongue coating separation method based on an LGBM model, and relates to the technical field of tongue body image processing, and the method comprises the steps: dividing tongue quality pixel points and tongue coating pixel points into N types of different colors according to the color features of a tongue body image; then performing sampling to respectively obtain characteristic values of each channel component of the sample in the RGB color space, the HSV color space and the LAB color space, and combining the characteristics to obtain a characteristic vector; utilizing a PCA dimension reduction method to perform dimension reduction on the feature vector to obtain M types of features as input features of the LGBM model; performing training optimization on the LGBM model through the training set to obtain a trained LGBM tongue body color classifier; and preprocessing the tongue body images to be classified, and inputting the preprocessed tongue body images into the LGBM tongue body color classifier to obtain corresponding tongue quality images and tongue coating images respectively. The method has the advantages that the classification precision is greatly improved, the classification time is short, and few resources are occupied.

Description

technical field [0001] The present invention relates to the technical field of tongue body image processing, and more specifically relates to a method for separating tongue body and coating based on an LGBM model. Background technique [0002] At present, the tongue body and tongue coating classification of the tongue body image is mainly carried out in the following two ways: [0003] 1. K-means classification method; [0004] 2. SVM classifier; [0005] The difficulty of tongue texture and tongue coating separation technology is that for different individuals and different photographing environments, tongue images vary widely. There is no clear standard for distinguishing tongue texture and tongue coating, and it is difficult to build a sample library of tongue texture or tongue coating. Supervised classification algorithm, however, for the K-means classification method, although it is an unsupervised classification algorithm, it can be used without preconditions and sam...

AI Technical Summary

Problems solved by technology

[0007] The object of the present invention is: in order to solve the problems of poor classification accuracy, poor classification effect and low classification efficiency of the existing methods for classifying the tongue and fur, the present invention provides a method for separating the tongue and fur based on the LGBM model

Images