A Novel Algorithm for Feature Level Fusion Using SVM Classifier for Multibiometrics-Based Person Identification

Table 2

Comparison of the proposed approach with Besbes et al. [3], Jagadeesan et al. [18], and Conti et al. [25] (FAR, FRR, and response time (mean) in seconds).

Author	Methods	Modalities	FAR (RBF; SVM)	FRR (RBF; SVM)	FAR; (Poly SVM)	FRR; (Poly SVM)	Training time (RBF; SVM)	Testing time; (RBF; SVM)	Training time (Poly SVM)	Testing time; (Poly SVM)
Besbes et al. [3]	Feature level fusion	Fingerprint	14%	22%	16%	22%	5.7	0.35	5.89	0.29
		Iris	12%	15%	14%	18%	4.82	0.28	5.2	0.32
		Fusion	4%	10%	4%	10%	4.3	0.20	4.98	0.24
Jagadeesan et al. [18]	Feature level fusion	Fingerprint	2%	17%	8%	18%	6.02	0.31	6.88	0.34
		Iris	1%	12%	6%	15%	4.34	0.26	5.20	0.28
		Fusion	0%	9%	0%	9%	4.98	0.20	5.01	0.21
Conti et al.; [25]	Feature level fusion	Fingerprint	2%	17%	8%	19%	5.47	0.28	5.90	0.30
		Iris	1%	10%	6%	16%	5.69	0.25	5.76	0.27
		Fusion	0%	8%	0%	9%	4.58	0.19	4.69	0.22
Proposed system	Feature level fusion	Fingerprint	4%	13%	8%	15%	5.22	0.2	4.78	0.21
		Iris	2%	12%	4%	16%	3.25	0.25	4.94	0.16
		Fusion	0%	6%	0%	7%	3.27	0.12	3.98	0.19