To overcome the small-sample issue, we propose a novel Dual Prototype Learning (DPL) method to optimize feature space distribution, making same-class features more compact and different-class features more separated. The DPL acts as a regularization technique, enhancing the model’s generalization ab...
Oct 31, 2024 SST-DPN This is the official repository to the paper "A Spatial-Spectral and Temporal Dual Prototype Network for Motor Imagery Brain-Computer Interface". Abstract To extract powerful spatial-spectral features, we design a lightweight attention mechanism that explicitly models the relatio...