Palm Print Recognition for Identity Verification Using Deep Learning Techniques

Abstract

Biometric authentication methods are widely used today to improve security in digital systems. Among the available biometric traits, palm prints are considered reliable because the contain several unique features such as major palm lines, wrinkles, patterns, and texture details. These characteristics remain mostly unchanged over time and differ from person to person.

This paper describes a palm print recognition system developed using deep learning and proposed system includes image pre-processing, feature learning and classification. Pre-processing steps such as resizing, normalization, and noise removal are applied to enhance the quality of palm images. Unlike traditional palm print recognition methods that depend manually on extracted features, the CNN model automatically learns important features directly from the images.

The proposed system achieves more than 99% recognition accuracy and performs even in the variations in lighting conditions. The system can be effectively used in applications such as access control, digital banking and government-based identity verification systems.