Local Binary Patterns for Dorsal Hand Vein Recognition

Abstrak

Biometric system is a more promising identification method than using cards and passwords. One of the physiological biometrics that is currently being developed is the dorsal hand vein pattern. The pattern of vein contains information that is able to authenticate individual identities for the purpose of personal authentication automatically. Vein patterns can be recognized by using a feature extraction approach. Therefore, in this study the Local Binary Patterns (LBP) method was used. LBP is used as a feature extraction method to optimize the texture feature of vein to obtain optimal accuracy and processing speed. The implementation of Local Binary Patterns (LBP) method for the dorsal hand vein recognition system was tested on 200 datasets. The test results show recognition accuracy up to 94% with an average accuracy of 90.67%.