Pulmonary Nodule Detection Model Based on SVM and CT Image Feature-Level Fusion with Rough Sets

Table 4

Statistics of effectiveness before and after rough set reduction.

	Serial number	Accuracy (%)	Sensibility (%)	Specificity (%)	Processing time (s)
Before reduction	1	96.42	92.86	100	1.0610
	2	91.96	83.93	100	0.6170
	3	95.54	100	91.07	0.5490
	4	89.28	100	78.57	0.5630
	5	95.54	91.07	100	0.5470
	6	98.21	96.43	100	0.5460
	7	94.64	89.29	100	0.5460
	8	95.53	91.07	100	0.5460
	9	91.96	83.93	100	0.5460
	10	97.32	100	96.64	0.5300
	Mean	94.64	92.86	96.43	0.6051
After reduction (Rs1)	1	100	100	100	0.9370
	2	100	100	100	0.4360
	3	100	100	100	0.3870
	4	100	100	100	0.4210
	5	100	100	100	0.4210
	6	100	100	100	0.3900
	7	100	100	100	0.4060
	8	91.67	100	83.33	0.4060
	9	100	100	100	0.3740
	10	100	100	100	0.3930
	Mean	99.17	100	98.33	0.4571
Increase after reduction		4.53	7.14	1.9	0.148