FCN retina image blood vessel segmentation through combination of depth separable convolution and channel weighing

Abstract

The present invention relates to an FCN retina image blood vessel segmentation through combination of depth separable convolution and channel weighing. The method comprises the steps of: 1) performingCLAHE and Gamma correction to enhance a contrast ratio of a green channel of an eye bottom image; 2) in order to adapt network training, performing partitioning of the enhanced image to expand data;and 3) replacing a standard convolution mode with depth separable convolution to increase the width of the network, and introducing a channel weighing module to explicitly perform modeling of a dependency relation of a feature channel so as to improve the separability of the features. The depth separable convolution and the channel weighing are combined to be applied into the FCN network, an expert manual identification result is taken as supervision to perform experiment in a DRIVE database. The result shows that: the method can accurately perform segmentation of the retina image blood vessels and has high robustness.

Description

technical field [0001] The invention relates to a FCN retinal image blood vessel segmentation method combined with depth separable convolution and channel weighting, which is better than the prior art in terms of sensitivity, specificity and accuracy, and has good segmentation performance, belonging to medical image processing, field of deep learning. Background technique [0002] Studies have shown that diabetic retinopathy, arteriosclerosis, leukemia and other diseases will affect the fundus blood vessels, resulting in changes in their length, width, angle and vascular proliferation. Clinically, the retinal images of the fundus are often used to screen, analyze and diagnose diseases. Therefore, in order to quantitatively analyze diseases, fundus blood vessel segmentation has become a key step in retina-related work, which has guiding significance for the diagnosis of human diseases and is a manifestation of science benefiting mankind. [0003] The problem of blood vessel...

AI Technical Summary

Problems solved by technology

The manual segmentation method relies on the operator's technical experience, which is subject to more subjective factors, low repeatability, and low efficiency; the unsupervised segmentation method does not require prior label information, but the segmentation effect on pathological images with tissue damage is relatively low. Poor; supervised methods are mainly based on extracted features to train classifiers to achieve the purpose of identifying blood vessels and non-vascular, but the requirements for feature training classifiers are very high, and a large number of pre-segmented retinal blood vessel images are required as training samples. To ensure the accuracy of the model, the requirements for medical images are relatively high

Images