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

Table 5

Relationship between the number of layers and recognition performance.

Layer	Convolution and pooling layers	Time (s)	Accuracy (%)	Error (%)	Sensitivity (%)	Specificity (%)	MCC	score
2	C1-S1	117.76	50.00	—	—	—	—	—
4	C1-S1-C2-S2	485.38	82.30	17.50	99.60	65.40	0.69	0.82
5	C1-S1-C2-S2-C3	1161.70	83.80	16.20	72.40	95.20	0.69	0.84
6	C1-S1-C2-S2-C3-S3	1765.49	89.30	10.70	92.00	86.60	0.79	0.89
7	C1-S1-C2-S2-C3-S3-C4	1779.64	85.40	14.60	97.00	73.80	0.73	0.85
8	C1-S1-C2-S2-C3-S3-C4-S4	1881.06	76.00	24.00	65.60	86.40	0.53	0.76