Finger Vein Verification on Different Datasets Based on Deep Learning with Triplet Loss

Table 6

The accuracy, precision, recall, ROC AUC, and best distance of corresponding training-validation sets in VGGNet16 model. The values at .

Training-validation	Accuracy	Precision	Recall	ROC AUC	Best distances
FV-USM-HKPU; (F-H)				0.9958		0.8821
FV-USM-SDUMLA-HMT; (F-S)				0.9801		0.6750
HKPU-FV-USM; (H-F)				0.9968		0.9343
HKPU-SDUMLA-HMT; (H-S)				0.9879		0.7478
SDUMLA-HMT-FV-USM; (S-F)				0.9953		0.8896
SDUMLA-HMT-HKPU; (S-H)				0.9954		0.8574