Identity recognition with palm vein feature using local binary pattern rotation Invariant
Annisa Yuditya Pratiwi, Tjokorda Agung Budi Wirayuda, Kurniawan Nur Ramadhani · 2016
Biometrics are being used as one of many alternatives in Recognition System. Palm vein is one of many features that could be used in biometrics. Several advantages of palm vein is that they were not easily damaged and rather hard to be duplicated because it is located deep inside skin layers. Palm vein is also cannot be seen with plain eyes or ordinary cameras. The research in this biometrics recognition system is done using Local Binary Pattern Rotation Invariant (LBPROT) as a feature extraction method on palm vein image. The matching process itself was done by using Cosine Distance to determine the nearest distance to which image. Test result shows that the optimal configuration of the method is by using 16 regions, 15 radius length, 8 neighborhood pixel, and 256×256 pixel image resolution. Another result shown by setting the threshold level in 0.085 is that the FAR score reached 0.11698 and the FRR score reached 0.1175, with 11.7% EER score and recognition accuracy rate in 96%.