AD judgment method and system based on gene difference and image research

By screening out differentially expressed genes and constructing a three-dimensional gray matter image dataset, combined with a lightweight 3D-ECA-EfficientNet model, the problem of existing methods being unable to effectively acquire multi-level pathological features was solved, achieving high-precision diagnosis of Alzheimer's disease.

CN122266718APending Publication Date: 2026-06-23YUNNAN UNIV
0 Cites 0 Cited by

Patent Information

Application Number
CN202610360819.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-24
Publication Date
2026-06-23

AI Technical Summary

Technical Problem

Existing Alzheimer's disease diagnosis methods based on convolutional neural networks cannot effectively acquire multi-level and multi-dimensional pathological-related global features, resulting in a loss of diagnostic accuracy and an inability to adaptively focus on lesion areas.

Method used

Based on gene differences, the most significantly different genes were screened out, and images of brain tissue (gray matter structure) with high expression were selected. The images were then processed using a lightweight 3D-ECA-EfficientNet model, combined with the ECA attention mechanism, to construct a three-dimensional gray matter image dataset for diagnosis.

Benefits of technology

It achieves a more accurate understanding of the early pathological features of Alzheimer's disease and higher classification accuracy, with a CN/AD classification accuracy of 97.51% and a CN/MCI classification accuracy of 93.36%.

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

Abstract

The invention discloses an AD judgment method and system based on gene difference and image research, and the method comprises the steps: screening out differential expression genes between an AD patient and CN by using a bioinformatics method, carrying out KEGG and GO analysis by selecting the gene with the most significant difference, and meanwhile, carrying out the MRI image double-sample T test of AD and CN, screening out highest expression brain tissue images of the differentially expressed genes from an HPA database, and determining that the grey matter part is directly influenced by diseases; the screened three-dimensional grey matter image is processed, and the processed image serves as input data to be input into a 3D-ECA-OfficientNet model; according to the model, the accuracy rates of 97.51% and 93.36% are obtained in CN / AD classification tasks and CN / MCI classification tasks respectively. According to the method, the early pathological features of the Alzheimer's disease can be known more accurately, and higher classification accuracy can be obtained.
Need to check novelty before this filing date? Find Prior Art