Neural image curvature estimation method and device based on topological structure

Abstract

The invention belongs to the field of neural image processing, and discloses a neural image curvature estimation method based on a topological structure, and the method comprises the steps of obtaining corneal neural image information, and separating the corneal neural image information through a deep learning network to obtain corneal neural network segments; calculating a curvature index of thecorneal neural network segment, and performing curvature index aggregation on the corneal neural network segment by adopting a weighted average aggregation algorithm to obtain a curvature aggregationindex; establishing a corneal nerve topological structure based on corneal nerve image information; analyzing the anisotropy of the corneal neural network segment, and fusing the curvature polymerization indexes by adopting an induced generalized ordered weighted average method to obtain the curvature of the corneal neural image; according to the invention, the morphological structures and functional anisotropy of different corneal nerve segments are fully considered, individualized analysis is performed on different nerve branches, clinical diagnosis experience and corneal nerve curvature automatic calculation can be effectively combined, and therefore clinical diagnosis based on corneal nerve images is achieved.

Description

technical field [0001] The invention belongs to the field of neural image processing, and in particular relates to a method and device for estimating curvature of neural images based on topology. Background technique [0002] Diabetes is currently a systemic disease with a high incidence in the world, which has seriously affected people's life, health and safety. Early screening, diagnosis and intervention can reduce the incidence of diabetes and its complications to a certain extent. Therefore, simple and effective diagnostic methods for diabetes Early screening is of great significance. Studies have shown that diabetes can cause morphological changes in retinal blood vessels and corneal nerves, and diabetic corneal neuropathy (Diabetic Corneal Neuropathy, hereinafter referred to as DCN) occurs before retinal vascular disease. Therefore, early screening and diagnosis of diabetes based on changes in corneal nerve morphology has very important clinical significance. At pres...

AI Technical Summary

Problems solved by technology

[0003] However, the currently widely used IVCM equipment does not have the function of quantifying the curvature of the corneal nerve, and the evaluation of the curvature of the corneal nerve largely depends on the doctor's clinical experience

With the continuous improvement of equipment technology and the popularization of corneal nerve detection, corneal nerve image data has shown explosive growth, so the manual diagnosis method relying on experts can no longer meet the actual clinical needs; at the same time, due to the differences in clinical experience of doctors and the subjectivity of manual diagnosis , may cause different doctors to give different diagnostic results for the same image of corneal neuropathy

It can be seen that the evaluation of diabetic corneal nerve curvature based on manual diagnosis has the disadvantages of low diagnostic efficiency, strong subjectivity, poor reproducibility, and inconsistent standards, which may easily lead to misdiagnosis or missed diagnosis, so that patients cannot receive corresponding medical treatment in time. treat

[0004] At the same time, most of the current automatic evaluation methods for the curvature of corneal nerve images are aimed at a single nerve segment. Clinically, the diagnosis of diseases is mostly based on the entire image as the smallest unit. The curvature of a single nerve cannot be directly applied to the automatic diagnosis of diabetes, so it is common The method is to aggregate the parameters of the corneal nerve segments contained in the image indiscriminately to achieve the curvature grading of the entire image. However, doctors actually include the selection of typical nerve structures when evaluating the curvature of the corneal nerve. Simply use The average method ignores the heterogeneity of nerve structure and function, cannot accurately describe the curvature of the corneal nerve in the entire image, and cannot be directly used for clinical examination of related diseases

Images