Corneal disease diagnosis system, method and terminal based on anterior segment photography

By developing a corneal disease diagnostic model based on the fusion of learnable wavelets and Morse topological features, the accuracy and robustness of corneal disease diagnosis under low-cost anterior segment photography conditions are addressed, and effective identification of abnormalities in the overall morphology and texture frequency of the cornea is achieved.

CN122135939AActive Publication Date: 2026-06-02SHANGHAI FIRST PEOPLES HOSPITAL
9 Cites 0 Cited by

Patent Information

Application Number
CN202610247830.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-02
Publication Date
2026-06-02
Estimated Expiration
2046-03-02

AI Technical Summary

Technical Problem

Existing technologies struggle to simultaneously model the overall morphology and structure of the cornea and characterize abnormal texture frequencies under low-cost anterior segment photography conditions, thus limiting the accuracy and robustness of corneal disease diagnosis.

Method used

A corneal disease diagnostic model based on the fusion of learnable wavelets and Morse topology features is adopted. The model performs frequency domain decomposition and modulation through a frequency-aware modulation module, global topology response modeling through a topology-aware module, and cross-modal alignment and consistency sensing through a confidence-aware topology and frequency fusion module, thereby achieving accurate diagnosis of corneal diseases.

Benefits of technology

It improves the accuracy and robustness of corneal disease diagnosis, is suitable for large-scale screening and auxiliary diagnosis, and can effectively identify abnormalities in the overall morphology and texture frequency of the cornea.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a cornea disease diagnosis system, method and terminal based on anterior segment photography, and the system and method are characterized in that the acquired anterior segment photography is input into a cornea disease diagnosis model based on learning wavelet and Morse topological feature fusion to obtain a diagnosis result. In the model, a frequency sensing modulation module performs learnable frequency domain decomposition and sub-band level modulation on an input image to obtain frequency domain characteristics and reliability scores; the topology sensing module carries out global modeling on topological response long-range dependency relationships of different scales and positions of the image to obtain topology embedding features and response scores. And the confidence perception topology and frequency fusion module performs cross-modal alignment and modulation through a bidirectional cross attention mechanism, and perceives consistency to obtain a diagnosis result. According to the method, Morse theoretical modeling is introduced into a topological structure branch, stable cornea morphological characterization is constructed in a multi-scale mode, and efficient modeling long-range dependence is achieved in combination with a Mamba model; the frequency domain characterization branches highlight discriminative frequency components, interference is suppressed, and the texture feature stability is improved. Besides, a fusion mechanism is also designed, adaptive correction and complementation of heterogeneous features are realized, a reliable voting strategy is introduced, and the robustness and credibility of a diagnosis result are improved.
Need to check novelty before this filing date? Find Prior Art