Diffusate detection method of retina fundus image

Abstract

The invention discloses a diffusate detection method of a retina fundus image. The method includes a fundus image shot by use of a color digital non-mydriatic fundus camera and comprises the following steps: 1, adjusting the size of an original image; selecting an RGB channel of the adjusted image; 3, positioning an optic disc of the adjusted image; 4, positioning a view field of the adjusted image; 5, performing dual-scale positioning on a diffusate suspicious area; 6, performing feature extraction on the diffusate suspicious area; and 7, performing classification detection on the diffusate suspicious area, and generating the fundus image marking a diffusate. According to the method provided by the invention, fundus images obtained under different acquisition conditions can be processed, the detection of the diffusate is robust and effective, the diffusate positions of the fundus images after processing are visual and obvious, and further diagnosis of an ophthalmological doctor is further facilitated.

Description

technical field [0001] The invention belongs to the field of image processing, and in particular relates to a method for detecting exudate of retinal fundus images. Background technique [0002] Diabetic retinopathy (Diabetic Retinopathy) is a serious widespread eye disease, it is the leading cause of blindness in adults. Early detection and treatment of retinopathy can minimize vision loss. Intraretinal exudate is one of the early visible signs of diabetic retinopathy, and it is also an important basis for lesion staging and diagnosis. Exudates can be divided into hard exudates (hardexudates) and soft exudates (softexudates). Due to the increased permeability of blood vessels, lipids leak out of the serum, and accumulate to form spot-like lesions of various sizes with clear edges called hard exudate; bright white, cotton-wool-like spot lesions are called soft exudates. exudate. [0003] At present, the detection of intraretinal exudate can be mainly divided into three c...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the task requirements of large-scale fundus image screening scenarios, the present invention proposes an exudate detection method for retinal fundus images in order to effectively deal with imaging quality problems such as uneven illumination and large contrast range of fundus images. The outflow area is more obvious, which is convenient for doctors to diagnose fundus lesions more accurately and further distinguish the severity of lesions

Images