Diabetic retinopathy sign detection method and device

Abstract

The present invention relates to a diabetic retinopathy sign detection method and device. The method comprises: receiving a fundus image to be detected; using a convolutional neural network (CNN) model to perform processing of the fundus image to be detected, and obtaining lesion area samples corresponding to the fundus image to be detected; according to the lesion area samples, constructing SIFT (Scale-Invariant Feature Transform) feature descriptors corresponding to the fundus image to be detected; and according to the SIFT feature descriptors, using an SVM (Support Vector Machine) classifier to determine the lesion type of the fundus image to be detected. The CNN model is employed to perform rough classification of the diabetic retinopathy to five out a lesion area, the SVM classifier is employed to perform fine classification of the lesion type of the lesion area to reduce interference generated by difference between the data size and the data when the neural network is directly configured to perform lesion classification and improve the classification precision.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method and device for detecting signs of diabetic retinopathy. Background technique [0002] In recent years, due to the acceleration of urbanization, the improvement of people's living standards, the change of diet structure and the increasingly tense pace of life, the incidence of diabetes in the population is increasing. As of 2016, the total number of diabetic patients in my country has reached 100000000. For patients with early-stage diabetes, an effective diagnosis can prevent the progression of the disease and greatly increase the likelihood that the patient will be cured. Diabetic retinopathy (hereinafter referred to as diabetic retinopathy) is the most common eye complication of diabetes and the main cause of blindness in adults aged 30-69. Therefore, the timely diagnosis of sugar reticulum plays a pivotal role in the prevention and diagnosis of diabetes. ...

AI Technical Summary

Problems solved by technology

[0006] Most of the more mature and effective methods for sugar net screening are based on neural networks. However, the current neural network-based methods use the entire image as a unit for model training and screening. Due to the large differences between images in different periods of sugar net Due to large and other reasons, the neural network-based method requires a large amount of data when it is implemented, and the existing data cannot meet the needs of model training, and the effect of the method is not perfect after implementation.

[0007] Most of the existing methods except the neural network can only detect certain lesion characteristics, and most of the sugar mesh fundus images are in the state of coexistence of multiple lesions, which makes these methods poor in detection effect and only have theoretical research significance. It has no use value, and the method of neural network can only realize the screening of images, and cannot detect and classify specific lesions and analyze sugar nets

Images