Finger Vein Recognition Using Perimeter and PCA

Biometrics is a way to identify or recognize the identity of a person based on physiological or behavioral characteristics. This method is applied in high-security systems such as banking, important buildings, and others. Finger vein pattern of veins is one of the relatively new biometric modality. Finger vein is hidden beneath skin make it difficult to forge. In this research, finger vein recognition using PCA is improved using perimeter measurements in image preprocessing stage. Result of the experiment showed that perimeter measurement makes reduction on PC number to achieve high accuracy. Highest accuracy in system with perimeter measurement was reached when we used K=1, cityblock distance and number of PC=180.