An electroencephalogram signal recognition method and system based on joint modeling of space-time features
By constructing local temporal coding, spatial correlation coding, and global temporal coding modules, we have achieved joint modeling of the multi-scale temporal features and spatial correlation of EEG signals, which solves the problem of insufficient spatiotemporal information collaborative representation in existing technologies and improves the classification accuracy and stability of EEG signals related to schizophrenia.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HOHAI UNIV
- Filing Date
- 2026-04-21
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies in the analysis of EEG signals related to schizophrenia struggle to effectively combine the multi-scale temporal characteristics of EEG signals with the spatial correlation between EEG channels, resulting in insufficient spatiotemporal information co-representation capabilities and affecting the accuracy and stability of classification results.
A spatiotemporal feature-based joint modeling approach is adopted. Multi-scale temporal features are extracted through a local temporal coding module, and spatial relationships between EEG channels are modeled by a Transformer-based spatial correlation coding module. Feature fusion and aggregation are performed through a global temporal coding module, and finally, category identification is performed using an identification output module.
It improves the accuracy and stability of EEG signal category identification, can better capture the spatiotemporal information of EEG signals, and enhances the auxiliary discrimination ability between schizophrenia-related categories and healthy control categories.
Smart Images
Figure CN122398338A_ABST