Palmprint-Based Biometric Identification Using CNN, LDA, and KNN

Abstract

In recent decades, biometric systems have evolved significantly, making them an effective alternative to traditional means of identification and authentication. One of the recent trends in this field is the use of palmprint, due to its accuracy and speed in identity verification. Our aim of this work is to increase the accuracy of identifying individuals through the use of palm prints. We start by extracting the region of interest (ROI) from the original image after that, we extract features using the convolutional neural network (CNN) algorithm and then the dimensionality reduction phase using linear discriminant analysis (LDA) and then the features are classified using the KNN algorithm.