Method for extracting blood vessels in an eye fundus image based on multi-feature fusion and electronic equipment

Abstract

The embodiment of the invention provides a method for extracting blood vessels in an eye fundus image based on multi-feature fusion and electronic equipment, and the method comprises the steps: employing a plurality of different classifiers to extract a plurality of different features of the eye fundus blood vessels based on a to-be-processed eye fundus image, and carrying out the feature fusion of the plurality of different features; obtaining a segmented image of the fundus vessel by using the trained denseCRF model based on the comprehensive features after feature fusion; and performing morphological analysis processing on the segmented image, and extracting a binary image of the fundus vessel. According to the embodiment of the invention, the blood vessel image in the fundus image is extracted through multi-feature fusion of the fundus blood vessel, and the blood vessel edge image, including the fundus blood vessel edge image of the blurred image, can be extracted more effectivelyand more accurately.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of image processing, and more specifically, to a method and electronic equipment for extracting blood vessels in fundus images based on multi-feature fusion. Background technique [0002] For the analysis of fundus images, more and more researchers pay attention to image semantic segmentation methods, such as threshold segmentation, edge segmentation, gene coding segmentation, wavelet transform segmentation, cluster segmentation, etc. With the development of artificial intelligence recognition technology, neural network (CNN) has attracted extensive attention of researchers and has been applied in the field of image segmentation. [0003] At present, there are more image semantic segmentation methods based on the full convolutional network FCN+CRF, and the image segmentation algorithm represented by FCN+denseCRF. The training samples are trained by the FCN model to obtain the spatial sh...

AI Technical Summary

Problems solved by technology

Because the training features are relatively single, the classification results obtained by the trained model are not fine enough and are not sensitive to the details in the image, so the extraction accuracy is not good enough, and the edge accuracy is not high, especially for blood vessel edge images of some blurred images. Insufficient detail extraction

Images