Retinal fundus image preprocessing method

Abstract

The invention discloses a retinal fundus image preprocessing method. The retinal fundus image preprocessing method is characterized by including the steps of 1), reading an original image, namely reading the original image of the retinal fundus through a green channel; 2), removing blood vessel center light reflex; 3), removing salt-and-pepper noise; 4), smoothing noise; 5), conducting background extraction; 6), obtaining a shadow correction image; 7), obtaining a homogenization image; 8), solving a complementary image; 9), obtaining a blood vessel enhanced image; 10), outputting the enhanced image. The retinal fundus image preprocessing method has the advantages that the method enables the processed retinal image to be better in image quality and richer in information quantity, reserved enhancement of useful information such as blood vessels, optic disks and central foveae in the retinal image is achieved, and feature extraction, segmentation and recognition of the retinal image are improved.

Description

technical field [0001] The invention relates to feature extraction and image segmentation technology, in particular to a retinal fundus image preprocessing method. Background technique [0002] The arteries of the fundus retina are the only blood vessels that can be seen in the whole body of the human body. The blood vessels of the fundus can be observed intuitively, which reflects the blood vessels of the whole body. However, due to the influence of retinal vascular structure, imaging equipment, environment and noise, the quality of retinal images is generally poor. In order to facilitate the observation of retinal fundus, retinal images need to be preprocessed before use. [0003] Most of the current retinal vessel feature extraction and segmentation algorithms are based on enhanced retinal images, which requires that the processed retinal images should retain the original vessel texture information as much as possible, otherwise the final vessel segmentation accuracy will...

AI Technical Summary

Problems solved by technology

The above methods are to enhance blood vessel information to make it more conducive to the extraction of vascular network in the next step. These methods are widely used in image enhancement because of the simplicity of the algorithm, but their defects in practical applications are obvious: through the optimization algorithm to achieve two The parameters of the three-dimensional matched filter are selected to enhance the retinal image, but it can only enhance the blood vessels and small gray areas in the image; the adaptive histogram equalization method is used to normalize the brightness of the image, which greatly improves the image quality. The contrast between the blood vessel and the background, but it cannot suppress the background noise well, and may weaken the region of interest into the background: Contourlet wavelet transform realizes the global enhancement of the retinal image by adjusting the coefficient, but the determination of the coefficient is more complicated, and the enhancement result changes the width of the blood vessels in the image

[0004] In addition, some iconic features of the retina, such as the optic disc, blood vessels, and fovea, are prerequisites for later segmentation and recognition. However, retinal images are affected by illumination changes, low contrast, and noise, making blood vessel detection and segmentation more difficult. It is required to eliminate these effects as far as possible from the retinal image, and to achieve reserved enhancement of useful information such as blood vessels, optic discs, and fovea in the retinal image. The traditional single image enhancement method has been unable to meet such requirements.

Images