Line hand feature-based palm-print identification system using learning vector quantization

Palm print identification has become one of the most active research areas of image processing, computer vision, graphics, and visualization. The palm print identification problems which are due to affectedness, illumination, impediment, shadow, and blur are undertaken using various methods such as hierarchal feature-based, template matching, graphical record matching, and artificial neural network approach. We propose a line hand featurebased palm print identification system with GLCM feature extraction (including 5 differentiators - Angular Second Moment/Energy, Contrast, Homogeneity/Inverse Difference Moment, Entropy, Correlation) and Learning Vector Quantization (LVQ) artificial neural network as a classifier. The motivation of using this palm print identification is because of its unique solution to the stability-malleability dilemma, where it is the ability to preserve antecedently learnt knowledge (stability), and potency to accommodate new patterns indefinitely (malleability). Another motivation is also used for biometric authentication of a person's identity because of its profusion, where it not only has the information available on the fingerprint, but also it has a far more amount of details in terms of principal lines and wrinkles. Yet compared with the other biometric characteristics, palm print identification has several advantages : low-cost capturing device, low-resolution imaging, and low-officiousness. The experiments show a promising result by using the proposed method that obtained an identification rate of 98.75%.