SIFT Based Vein Recognition Models: Analysis and Improvement

Table 1

Influence of selected CEs on the number of detected SIFT keypoints. Values represent the number variations by different CEs on left and right hands with male and female. (The number in the bracket represents the increased times against the number of original images.)

Image set		Linear/nonlinear functional enhancement (LNFE)
	Original	IN	GLS	CS	GC08	LS	UM	HBF	HF	INE
FL	8	19	74	18	10	13	75	18	8	8
FR	12	19	87	17	9	12	77	17	9	9
ML	13	13	46	13	14	18	93	16	13	13
MR	15	16	53	15	14	17	120	16	15	14
Average	12	17 (1.4)	65 (5.4)	16 (1.3)	12 ()	15 (1.3)	91 (7.6)	17 (1.4)	11 (0.9)	11 (0.9)
Image set	HE and its variants (HEs)					Hierarchical enhancement models (HEM)
	Original	HE	AHE	CLAHE	CLHE	IHE	DHE	RASF	HHE	HEHBF
FL	8	25	348	89	23	99	502	86	132	53
FR	12	31	375	25	32	118	438	80	157	53
ML	13	39	299	66	36	134	459	101	137	84
MR	15	40	327	40	38	164	517	82	170	89
Average	12	34 (2.8)	337 (28.1)	55 (4.6)	32 (2.7)	129 (10.8)	479 (39.9)	87 (7.3)	149 (12.4)	70 (5.8)