Diabetic retinopathy grade classification method based on deep learning

Abstract

The invention provides a diabetic retinopathy grade classification method based on deep learning. The diabetic retinopathy grade classification method comprises the steps of: constructing a sample library; removing backgrounds and noise of ophthalmoscope photographs in the sample library; normalizing the images of different brightness and different intensity to the same range by adopting a local mean value subtracting method; adopting random stretching and rotating methods for different samples for data augmentation, and constructing a training set and a test set; training an initial deep learning network model by establishing an input portion architecture, a multi-branch feature transformation portion architecture and an output portion architecture separately; and inputting samples to betested into the trained initial deep learning network model for diabetic retinopathy grade classification. Compared with the traditional processing method, the diabetic retinopathy grade classification method gets rid of the dependence on prior knowledge, and has good generalization ability; and by adopting the designed multiple grades, a small-sized convolution kernel can be used for extracting very tiny lesion features, thereby making the classification results more reliable.

Description

technical field [0001] The present invention relates to the technical field of medical and health image processing, in particular, to a method for grading diabetic retinopathy based on deep learning. Background technique [0002] Diabetic retinopathy (Diabetic Retinopathy, DR) is a complication caused by diabetes and is one of the main blinding diseases. Traditional retinal image processing methods include 4 stages: preprocessing, anatomical structure analysis, lesion detection, and lesion diagnosis. Each stage of retinal image processing requires more than one image processing technology, and the complex structure of retinal images is easy to cross-affect with various lesions, coupled with the influence of complex background changes, the processing of retinal images faces various difficulties. Unavoidable external factors make the techniques for processing retinal images complex, poor in generalization, and highly dependent on prior knowledge. In this process, if any tech...

AI Technical Summary

Problems solved by technology

Each stage of retinal image processing requires more than one image processing technology, and the structure of retinal images is complex, and it is easy to cross-affect with various lesions, coupled with the influence of complex background changes, the processing of retinal images faces various difficulties

Unavoidable external factors make the techniques for processing retinal images complex, poor in generalization, and highly dependent on prior knowledge

In this process, if any technical link is wrong or the effect is not ideal, it will lead to the failure of subsequent work or the large error in the image classification and detection results.

Images