A brainprint recognition method for a space-time dual-branch feature encoder in a multi-paradigm scene

By using a dual-branch spatiotemporal feature extraction network, the problem of poor adaptability of brainprint recognition in multi-paradigm scenarios is solved, the recognition accuracy is improved and the deployment cost is reduced, making it suitable for diverse identity authentication scenarios.

CN122087544APending Publication Date: 2026-05-26XIDIAN UNIV +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
XIDIAN UNIV
Filing Date
2026-03-13
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing brainprint recognition technology has poor adaptability in multiple paradigm scenarios, resulting in insufficient recognition accuracy and difficulty in meeting the needs of practical applications.

Method used

A dual-branch spatiotemporal feature extraction network is adopted, which extracts the temporal features of EEG signals and the spatial positional relationship of electrodes through the temporal feature extraction branch and the spatial feature extraction branch respectively. Combined with the multi-head self-attention mechanism and feature enhancement module, feature fusion is performed to enhance the model's representation ability.

Benefits of technology

It enhances cross-paradigm recognition capabilities, improves the adaptability and accuracy of brainprint recognition under multiple paradigms, and has the potential for low-cost and high-efficiency deployment, making it suitable for application in edge computing devices.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122087544A_ABST
    Figure CN122087544A_ABST
Patent Text Reader

Abstract

This invention discloses a brainprint recognition method using a spatiotemporal dual-branch feature encoder for multi-paradigm scenarios, relating to the technical field of brainprint recognition. This method leverages the inherent characteristics of EEG signals, employing parallel, independent temporal feature extraction branches and spatial feature extraction branches to extract temporal features and spatial positional relationships between electrodes from the EEG signals, respectively, while simultaneously reducing mutual interference in spatiotemporal extraction. The temporal feature extraction branch uses a dual-kernel convolutional architecture to simultaneously capture global and local features; the spatial feature extraction branch uses a convolutional-attention hybrid architecture to extract local spatial features and global spatial positional relationships, modeling global positional relationships through a multi-head self-attention mechanism, and then using a designed feature enhancement module to scale the feature dimensions; finally, a spatiotemporal feature fusion module fuses temporal and spatial features, and a classification head performs classification to predict the identity category.
Need to check novelty before this filing date? Find Prior Art