Application of Deep Learning Techniques for Biometric Systems

Abstract

This thesis focuses on developing and evaluating a biometric recognition system, specifically utilizing 2D palmprint recognition integrated with the GoogLeNet deep neural network architecture. The theoretical background encompasses the significance of biometrics, the various types of biometric systems, including multimodal systems, and an overview of deep learning, its types, applications, and benefits. The proposed biometric recognition system employs the GoogLeNet architecture for both classification and feature extraction. Using the PolyU Palmprint Database, experiments and results include parameter selection and vthe evaluation of both unimodal and multimodal biometric systems. A comparative study is conducted to assess the effectiveness of the proposed system. In conclusion, this thesis provides insights into the development, implementation, and evaluation of a novel biometric recognition system, highlighting its potential effectiveness in real-world applications.