一种基于脑电信号的轻量级信号分类方法及装置
By improving the deep residual shrinkage network and multi-directional collaborative attention mechanism, the problems of low efficiency and poor adaptability of traditional EEG detection in primary healthcare scenarios have been solved, achieving efficient and accurate Parkinson's disease screening.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SICHUAN UNIV
- Filing Date
- 2026-04-03
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional EEG-based Parkinson's disease detection methods suffer from low model efficiency, susceptibility to environmental interference, and poor adaptability across devices and populations in primary healthcare settings.
An improved deep residual shrinking network is used, combined with blueprint separable convolutional layers and a multi-directional collaborative attention mechanism, to preprocess and extract features from EEG signals. Weighted cross-entropy loss and orthogonal regularization loss are used for training to suppress noise and improve the model's generalization ability.
This technology enables efficient and accurate Parkinson's disease screening in low-computing-power environments, reducing the number of model parameters and computational requirements, and improving the stability and adaptability of the detection, making it suitable for primary healthcare settings.
Smart Images
Figure CN121971108B_ABST