Diabetic foot ulcer (DFU) classification method based on convolutional neural network

Abstract

A diabetic foot ulcer (DFU) classification method proposed in the invention based on a convolutional neural network includes the steps of creating a DFU data set, marking a region of interest (ROI), adding data, pre-processing training patches, conventional machine learning, and constructing a DFU network. The process includes: first collecting foot images of a DFU patient and a healthy person asdata sets and plotting ROIs, then adding data using a combination of various image processing techniques for network training, next, pre-processing the obtained patches, normalizing each pixel, and finally constructing a DFU network including input data, parallel convolutions, a fully connected layer, and an output classifier. The method of the invention utilizes an advanced convolutional neural network to process the input data more effectively and efficiently; and has high sensitivity and can effectively distinguish the characteristic differences between the healthy skin and the DFU, therebygreatly shortening the processing time.

Description

technical field [0001] The invention relates to the field of image classification, in particular to a method for classifying diabetic foot ulcers based on a convolutional neural network. Background technique [0002] Diabetic foot ulcers (DFU) are a major complication of diabetes that can put patients at risk if not managed properly when they occur. However, because DFU requires doctors to carefully diagnose the affected area, it also requires long-term treatment and expensive treatment and nursing costs, causing heavy economic burdens on patients and their families, especially in developing countries. The treatment of diseases accounts for 5% of the country's annual income. Therefore, if deep learning can be used to realize the automatic classification of DFU, the efficiency of diagnosis and treatment of DFU patients will be greatly improved. By developing automatic taggers, foot images can be automatically segmented and classified without the help of clinicians, and auto...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the inconvenience of the traditional DFU diagnosis method, the purpose of the present invention is to provide a method for classifying diabetic foot ulcers based on convolutional neural network. Firstly, the foot images of patients with diabetic foot ulcers and healthy people are collected as a data set and described. region, then augment the data by using a combination of various image processing techniques for network training, then preprocess the obtained patches, normalize each pixel, and finally build a DFU network, which includes the input data , parallel convolution, fully connected layers and output classifier

Images