Performance Evaluation of Distance Measurement in Biometric Finger Knuckle Print Recognition

Biometric identification is a security method that utilize parts of human body. Finger Knuckle Print (FKP) is one of many parts in human that relatively unique. This research try to find out the best distance measurement as recognition method with implication of pre-processing CLAHE. Several variants of distance measurement that evaluated in this research are Euclidean, Manhattan, Minkowski, Canberra, Chi-square, Chebyshev and Bray Curtis. Besides that, feature extraction Principal Component Analysis (PCA) was applied as texture feature after preprocessing phase. The experiment results has shown several recognition method that used pre-processing CLAHE able to improve accuracy in recognizing FKP. However, in the other hands several distance has showed decreased accuracy. The highest accuracy that used pre-processing CLAHE are gathered when applied with Chi-Square distance until 95.15%.