IncepA-EEGNet: 融合Inception网络和注意力机制的P300信号检测方法

许萌,王丹,李致远,陈远方

IncepA-EEGNet: P300 signal detection method based on fusion of Inception network and attention mechanism

Meng XU,Dan WANG,Zhi-yuan LI,Yuan-fang CHEN

表 2 IncepA-EEGNet使用不同卷积核参数K的分类结果

Tab.2 Classification results using different convolution kernel parameters K on IncepA-EEGNet

受试者 K 卷积核大小 Acc R P F1 A 8 (8/4/2,1) 0.7323 0.6630 0.3432 0.4523 A 16 (16/8/4,1) 0.7384 0.6547 0.3485 0.4548 A 32 (32/16/8,1) 0.7425 0.6486 0.3520 0.4564 A 64 (64/32/16,1) 0.7314 0.6677 0.3430 0.4532 A 128 (128/64/32,1) 0.7553 0.6457 0.3670 0.4679 A 160 (160/80/40,1) 0.7514 0.6210 0.3582 0.4543 B 8 (8/4/2,1) 0.7817 0.7034 0.4122 0.5149 B 16 (16/8/4,1) 0.7855 0.7186 0.4168 0.5276 B 32 (32/16/8,1) 0.7848 0.7143 0.4154 0.5253 B 64 (64/32/16,1) 0.7899 0.6993 0.4215 0.5260 B 128 (128/64/32,1) 0.7914 0.7250 0.4261 0.5367 B 160 (160/80/40,1) 0.7889 0.6987 0.4199 0.5245