Fundus image focus segmentation and quantitative analysis method based on data enhancement

Abstract

The invention discloses a fundus image focus segmentation and quantitative analysis method based on data enhancement, and the method comprises the steps: P1, obtaining medical image data, and carryingout the preprocessing; P2, extracting the preprocessed image data to obtain a fundus focus image; P3, establishing a calculation model based on the focus image; P4, predicting a focus area on the medical image data based on the calculation model; P5, counting the boundary of the lesion area of each parallel layer, calculating the lesion area, and combining to obtain the lesion volume; P6, substituting the focus volume into the fundus image, inputting the fundus image into the training model, and training to obtain a focus segmentation model; P7, performing manual segmentation on the segmentation model, and meanwhile, performing automatic segmentation through a model contour to obtain two true value tags; and P8, performing quantitative analysis through true value label comparison results,boundary deviation is avoided, side effects are reduced, cross validation is carried out finally, safety and accuracy are guaranteed, operation stability is improved, and application and popularization are facilitated.

Description

technical field [0001] The present invention relates to the technical field of lesion segmentation and analysis, in particular to a data enhancement-based fundus image lesion segmentation and quantitative analysis method. Background technique [0002] In the medical system, the use of computers and imaging equipment for image acquisition and segmentation can effectively control the results of lesion judgment and improve treatment effects. [0003] However, in some image lesion segmentation and analysis methods, most of them are directly viewed and compared through ultrasound images, which is not conducive to controlling the area and distribution of lesions, and the accuracy is low, and it is not possible to verify the matching of images and results. Image quantitative analysis accuracy, to be proposed a new method. Contents of the invention [0004] The purpose of the present invention is to solve the shortcomings in the prior art, and propose a method for segmentation an...

AI Technical Summary

Problems solved by technology

[0003] However, in some image lesion segmentation and analysis methods, most of them are directly viewed and compared through ultrasound images, which is not conducive to controlling the area a