Computer-Aided Diagnosis Research of a Lung Tumor Based on a Deep Convolutional Neural Network and Global Features

Table 4

Effect of feature maps on the experimental results.

Feature maps			Time (s)	Accuracy (%)	Error (%)	Sensitivity (%)	Specificity (%)	MCC	score
FM1	FM2	FM3
2	4	8	214.06	86.50	13.50	80.40	92.60	0.74	0.86
3	6	12	443.72	85.40	14.60	81.20	89.60	0.71	0.85
6	12	18	779.64	85.30	14.70	80.80	89.80	0.71	0.85
6	12	24	1161.70	89.30	10.70	92.00	86.60	0.79	0.89
6	12	36	1189.80	86.80	13.20	98.20	75.40	0.76	0.87
6	12	120	1392.19	87.00	13.00	98.60	75.40	0.76	0.87
6	12	150	1463.30	85.10	14.90	78.80	91.40	0.71	0.85
6	12	200	1582.14	86.80	13.20	83.80	89.80	0.74	0.87
6	16	32	2054.51	86.70	13.30	81.60	91.80	0.74	0.87
6	16	120	1812.94	87.90	12.10	93.80	82.00	0.76	0.88
6	16	200	2070.36	85.30	14.70	77.80	92.80	0.71	0.85
16	32	200	8616.49	71.70	28.30	100.00	43.40	0.53	0.69