Biometric verification of humans by means of hand geometry

TL;DR

This paper presents a hand geometry biometric identification system that uses a database of scanned hand images, feature extraction, and a neural network classifier, achieving up to 93.64% identification accuracy.

Contribution

It introduces a novel hand geometry biometric system with a new database, feature extraction method, and classifier, demonstrating high identification performance.

Findings

Maximum identification rate of 93.64%

Minimum detection cost function of 2.92%

Effective use of multilayer perceptron classifier

Abstract

This paper describes a hand geometry biometric identification system. We have acquired a database of 22 people, 10 acquisitions per person, using a conventional document scanner. We propose a feature extraction and classifier. The experimental results reveal a maximum identification rate equal to 93.64%, and a minimum value of the detection cost function equal to 2.92% using a multi layer perceptron classifier.