Android Based Palmprint Recognition System

Abstract

This paper aims to build a palmprint recognition system based on Android. Line detection and non-overlapping blocks methods used on the design of palmprint recognition system. Normalized Euclidean distance was used to match the two feature vector. The similarity degrees was measured by the smallest value. The method was applied to the system can result an accuracy rate about 91,32% (FMR = 5,2%, FNMR = 3,68%, T = 0,265). System testing was done as much as 2500 times to the 50 participants with 3 reference images of each user. The factor that influenced on system accuracy was the process of image acquisition. Palmprint images was taken directly by using camera on the android smartphone device.