Fig. 1. Traditional palm vein recognition system

Fig. 2. Proposed palm vein recognition system with pseudo image storage

Fig. 3. ROI extraction of palm vein image. (a) Original image; (b) location of ROI; (c) extracted ROI

Fig. 4. Original images and blocked images in three databases. (a) Original image in PolyU database; (b) 2×2 block image in PolyU database; (c) 4×4 block image in PolyU database; (d) 8×8 block image in PolyU database; (e) original image in CASIA database; (f) 2×2 block image in CASIA database; (g) 4×4 block image in CASIA database; (h) 8×8 block image in CASIA database; (i) ROI of original image in self-built database; (j) 2×2 block image in self-built database; (k) 4×4 block image in self-built databas

Fig. 5. Conversion between encrypted and decrypted images. (a) 2×2 blocked ROI; (b) pseudo image; (c) decrypted image

Fig. 6. Acquisition environment of PolyU database

Fig. 7. Acquisition environment of self-built database

Fig. 8. ROIs of palm vein in different databases. (a) PolyU database; (b) CASIA database; (c) self-built database

Fig. 9. ROC curve of PolyU database

Fig. 10. ROC curve of CASIA database

Fig. 11. ROC curve of self-built database

Table 1. Equal error rate of proposed algorithm with sample size of 100

Table 2. Recognition time of proposed algorithm with sample size of 100

Table 3. Equal error rate of proposed algorithm with sample size of 50

Table 4. Recognition time of proposed algorithm with sample size of 50

Table 5. Equal error rate of proposed algorithm and other algorithms

Table 6. Recognition time of proposed algorithm and other algorithms