A Comparative Study of Cross-Device Finger Vein Recognition Using Classical and Deep Learning Approaches

Table 5

Number of mated and non-mated pairs per device pair.

	UTFV		ZkTeco		IDIAP		NTNU		PFV_L		PFV_C
	Mated	Non-mated	Mated	Non-mated	Mated	Non-mated	Mated	Non-mated	Mated	Non-mated	Mated	Non-mated
UTFV	528	10,000	1,056	10,000	888	10,000	1,032	10,000	528	10,000	600	10,000
ZkTeco	1,056	10,000	540	10,000	912	10,000	1,056	10,000	540	10,000	612	10,000
IDIAP	888	10,000	912	10,000	456	10,000	888	10,000	432	10,000	480	10,000
NTNU	1,032	10,000	1,056	10,000	888	10,000	528	10,000	534	10,000	606	10,000
PFV_L	528	10,000	540	10,000	432	10,000	534	10,000	84	6,888	366	10,000
PFV_C	600	10,000	612	10,000	480	10,000	606	10,000	366	10,000	84	6,888

The columns of the table represent the probe devices, while the rows are the reference.