Eyeball positioning method and device for retinopathy image, and electronic equipment

Abstract

The invention discloses an eyeball positioning method and an eyeball positioning device for a retinopathy image and electronic equipment. The method comprises the steps of acquiring retinopathy image data; carrying out data annotation on the retinopathy image data to obtain annotated data; establishing an initial deep learning network; inputting the retinopathy image data into the initial deep learning network, and outputting the data to obtain corresponding prediction data; comparing the annotated data and the prediction data corresponding to the retinopathy image data by use of a loss function to obtain a comparison result; adjusting a parameter in the initial deep learning network according to the comparison result until the comparison result reaches a preset threshold value to obtain a final deep learning network model; and processing to-be-detected retinopathy image data by use of the deep learning network model to obtain a corresponding eyeball center coordinate and an eyeball diameter. According to the eyeball positioning method and the eyeball positioning device for the retinopathy image and the electronic equipment provided by the invention, eyeball positioning of a retinopathy image can be completed relatively accurately.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an eyeball positioning method, device and electronic equipment for retinopathy images. Background technique [0002] Retinopathy, also known as Rieger's central retinitis, juvenile hemorrhagic maculopathy. The disease is an isolated exudative chorioretinal lesion in and around the macula, accompanied by subretinal neovascularization and hemorrhage. [0003] In order to observe retinopathy in detail, it is usually necessary to collect images of retinopathy for analysis. [0004] However, the inventor of the present invention found in the process of realizing the present invention that the prior art has at least the following problems: [0005] At present, the quality of retinopathy images is not high. When using traditional methods to locate the eyeball position and eye area, it is often affected by factors such as low resolution of retinopathy images, low brightness, a...

AI Technical Summary

Problems solved by technology

[0005] At present, the quality of retinopathy images is not high. When using traditional methods to locate the eyeball position and eyeball area, it is often affected by factors such as low resolution of retinopathy images, low brightness, and camera spot interference, resulting in inaccurate eyeball positioning.

Images