表 2. Classification accuracy of different models (%).

不同模型的分类正确率（%）

受试	FBCSP	FBCSP-DNN	CNN	Shallow-ConvNet	DFFN	Deep-ConvNet	EEGNet	本文方法
注：粗体数字为结果最优者
受试1	86.41	87.70	77.39	79.51	83.46	81.28	84.38	86.85
受试2	65.28	68.21	60.14	56.25	69.30	59.52	52.12	71.76
受试3	90.28	85.59	82.92	88.89	90.29	88.12	87.82	93.54
受试4	66.67	66.71	72.28	78.90	71.07	63.88	63.18	75.68
受试5	62.50	67.00	75.83	57.29	65.41	70.24	66.57	80.75
受试6	45.49	67.40	68.98	53.82	69.45	64.82	57.58	67.77
受试7	89.58	91.00	76.03	91.67	88.18	87.12	87.82	91.97
受试8	83.32	83.80	76.85	81.25	85.76	82.14	78.36	86.21
受试9	79.51	87.60	84.66	79.17	90.54	83.66	82.79	92.68
标准差	15.22	10.62	11.35	14.44	10.26	11.01	13.71	9.53
平均值	74.46	78.33	75.00	74.31	79.71	75.64	73.39	83.03