Method for detecting atrophic arc of high myopia fundus image based on machine learning

Abstract

The invention discloses a method for detecting an atrophic arc of a high myopia fundus image based on machine learning, which comprises the following steps of: obtaining a fundus image, and randomly selecting a training set and a test set, removing noise and enhancing the fundus image, and extracting a main blood vessel contour, carrying out region-of-interest halving operation based on density distribution of the main blood vessels, and obtaining a positioning region of choroidal atrophic arcs of the optic disc and the periphery of the optic disc in the fundus image according to the brightness information and the shape information, taking the positioning region image as input, and extracting texture features of the positioning region image in combination with a complete local binary pattern operator and a binary Gabor pattern operator, and based on the extracted textural features, training a detection model on the training set by using a gradient elevator in machine learning, and finely adjusting parameters of the detection model to finally obtain a detection model of the atrophic arc of the high myopia fundus image. Automatic detection of choroidal atrophic arcs in fundus imagesof patients with high myopia is realized, the precision is high, and the speed is high.

Description

technical field [0001] The invention relates to a method for detecting atrophy arcs in fundus images of high myopia based on machine learning, and belongs to the field of medical image processing. Background technique [0002] The symptoms of high myopia are mainly excessive elongation of the anterior and posterior axis of the eyeball, and in the fundus, it is manifested as a choroidal atrophy arc around the optic disc, that is, a bright arc-shaped area around the optic disc that is slightly darker than the optic disc. In the fundus images of patients with high myopia, the choroid around the optic disc is pulled away from the temporal side of the nipple under the traction of the sclera, exposing the sclera behind it, forming a white arc-shaped spot. If the posterior pole of the eyeball continues to expand and extend, the detachment of the choroid gradually extends from the temporal side of the papilla to the periphery of the optic disc, and finally forms a ring-shaped spot. ...

AI Technical Summary

Problems solved by technology

[0003] At present, the early recognition of high myopia mainly relies on regular vision testing and the experience and judgment of doctors. However, manual testing is not only time-consuming and laborious, but also has a large number of myopia patients, and ophthalmic medical resources are far from enough to meet the needs of the majority of myopia patients.

Images