基于多脑区注意力机制胶囊融合网络的EEG-fNIRS情感识别

刘悦,张雪英,陈桂军,黄丽霞,孙颖

EEG-fNIRS emotion recognition based on multi-brain attention mechanism capsule fusion network

Yue LIU,Xueying ZHANG,Guijun CHEN,Lixia HUANG,Ying SUN

表 1 MBA-CF-cCapsNet模型的实验参数

Tab.1 Experimental parameter of MBA-CF-cCapsNet model

模型 层 参数 尺寸 多脑区注意力机制 Graph SAGE in_channels 5 Graph SAGE out_channels 5 Maxpooling Kernel C FC1 in_features 30 FC1 out_features 20 FC2 in_features 20 FC2 out_features 6 ReLu — — Sigmoid — — 卷积层 Conv1_1 Kernel 3×3×128 Conv1_2 Kernel 5×5×128 初级胶囊模块 Conv2_1 Kernel 3×3×128 Conv2_2 Kernel 5×5×128 Conv3 Kernel 1×1×256 胶囊融合模块 Maxpooling Kernel 8 ReLu — — Tanh — — 分类胶囊模块 动态路由机制 Wij 8×16