Oral mucosal disease recognition method based on deep learning multi-feature fusion

Abstract

The invention relates to an oral mucosal disease recognition method based on deep learning multi-feature fusion. The method can be used for diagnosing and recognizing oral leukoplakia, oral lichen planus, oral cancer and recurrent oral ulceration, the diagnosis efficiency is improved, and the recognition accuracy is improved. According to the specific scheme, the method comprises the following steps: (1) collecting an oral mucosal disease image through a camera under white light; (2) preprocessing the collected images, (3) extracting texture features of four oral diseases by using a gray level co-occurrence matrix GLCM algorithm, and extracting bottom features of shapes and colors from HOG and HSV images by using a neural network model; (4) performing extraction of high-level features of the RGB image by using an Officientnet network model; (5) performing feature selection by using a Pearson coefficient in combination with a random forest algorithm, and selecting features having a larger relationship with a target value; (6) carrying out classification identification to train the training set; and (7) performing test verification on the trained model by using the divided test set image.

Description

technical field [0001] The invention relates to the field of image recognition classification and counting, in particular to an oral mucosal disease recognition method based on deep learning multi-feature fusion. Background technique [0002] With the development of modern society and the improvement of people's living standards, oral health has been paid more and more attention by the public. Oral disease is one of the most common diseases in my country. With the increase of the population and the aging of the population, the number of patients continues to rise. [0003] Oral mucosal diseases have the characteristics of large intra-class differences and high inter-class similarities, and are often difficult to distinguish in disease diagnosis. Even professional doctors sometimes cannot make accurate diagnoses. Traditional oral disease diagnosis methods rely on the accumulated experience of doctors to make preliminary judgments, and then conduct biopsies according to diffe...

AI Technical Summary

Problems solved by technology

Moreover, since oral leukoplakia and oral lichen planus are precancerous lesions, they are similar to oral cancer in color, characteristics, and diseased sites, which bring greater challenges to the identification work

[0005] Because oral cancer is similar to oral leukoplakia and lichen planus, and the characteristics are difficult to distinguish, it is difficult to obtain good results only by using deep learning methods. design research proposal

Images