Retina optic disc segmentation method combining U-Net and region growing PCNN

Abstract

The invention discloses a U-Net and region growing PCNN combined retinal optic disk segmentation method. The method comprises the steps of performing graying processing on a retinal optic disk data set picture; performing CLAHE processing on the data set picture after graying processing to enhance the contrast between the optic disc and the background in the retinal optic disc image; partitioningthe retinal optic disc image; constructing and training a U-Net neural network model and roughly extracting pictures; constructing a regional growth PCNN neural network model; and carrying out retinaloptic disc segmentation by using the region growing PCNN. On one hand, the invention provides an improved U-Net retina optic disc image rough extraction method, and through the rough extraction, thebackground is significantly inhibited, the optic disc area is highlighted, and the picture contrast is increased, so that the picture quality of a data set is improved; on the other hand, the invention provides an optic disk image segmentation method based on the improved region growing PCNN, the PCNN segmentation performance is improved by changing a seed selection mode, a PCNN initial ignition threshold selection mode and a region growing end condition, and the segmentation of the complete optic disk is realized.

Description

technical field [0001] The invention relates to the technical field of optic disc recognition, in particular to a retinal optic disc segmentation method combining U-Net and region growing PCNN. Background technique [0002] Glaucoma is extremely harmful and is the leading cause of irreversible blindness. If it is not intervened and treated in its early stage, it will easily cause blindness. The size, shape and depth of the optic disc are closely related to the onset of glaucoma and the degree of symptoms. Therefore, the identification of the optic disc plays a crucial role in the diagnosis and treatment of glaucoma and has great research value. However, doctors need to review a large number of fundus images to make a diagnosis, which is a time-consuming and cumbersome process that is easily affected by subjective experience and fatigue. When the doctor is tired, the risk of missing certain details in the fundus image will increase, resulting in missed diagnosis and misdiag...

AI Technical Summary

Problems solved by technology

[0017] In one aspect of the present invention, the U-Net model is improved to improve the U-Net model and propose a rough extraction method based on the improved U-Net retinal optic disc image, through This kind of rough extraction can significantly suppress the background, highlight the optic disc area, weaken noise interference, increase the contrast of the image, and thus improve the image quality of the dataset

Images