A method for evaluating that quality of fundus image

Abstract

A method for evaluate that quality of fundus image is disclosed. Considering the influence of blur, overexposure and underexposure on the segmentation accuracy of fundus images, statistical feature vectors, texture feature vectors and shape feature vectors are extracted to form feature vectors. Then support vector regression is used to train the feature vectors of all distorted fundus images, anda prediction model is constructed. In the test phase, by calculating a feature vector of a distorted fundus image to be used as a test, and according to the prediction model constructed in the training stage, the segmentation accuracy value of the distorted fundus image used for testing is predicted, Since the obtained feature vector information can better reflect the distortion of the distorted fundus image on the accuracy of segmentation, Therefore, the correlation between the predicted segmentation accuracy value and the real segmentation accuracy value is effectively improved, that is, thefundus image quality can be accurately and automatically evaluated, and the correlation between the objective evaluation result and the subjective perception is effectively improved.

Description

technical field [0001] The invention relates to an image quality evaluation method, in particular to a fundus image quality evaluation method. Background technique [0002] The fundus image is captured by a special fundus camera. The fundus image includes the main physiological structures such as the optic disc, macula and blood vessels in the retina, and is an important type of image in medical imaging. Among them, the optic disc appears as a nearly circular bright color area in the normal fundus image, which has the strongest contrast with the background area, and is the initial area of ​​the optic nerve and blood vessels; the macula is rich in lutein, so it is in the normal fundus The image shows a dark area, and the dark area has no blood vessel structure, and there is an inward sunken area called the fovea in the center of the macula; the blood vessels start from the optic disc area and extend to the entire interior of the eyeball, presenting a tree-like distribution th...

AI Technical Summary

Problems solved by technology

[0003] Although the quality evaluation methods of fundus images are becoming more and more mature, most of them still rely on the subjective judgment of medical researchers. Although this subjective judgment method has certain reference value in evaluating the overall quality of fundus images , but it is difficult to evaluate the impact of fundus image quality on segmentation by subjective means

Fundus images can be used to extract blood vessel information, locate the optic disc, locate the macular area, etc. These applications provide huge information for the diagnosis and treatment of ophthalmic diseases, but the fundus images will inevitably be disturbed by noise during the acquisition and transmission process , leading to the loss of effective retinal information collected, which in turn affects the doctor's diagnosis

Images