Multibiometric Secure System Based on Deep Learning

TL;DR

This paper introduces a novel multibiometric secure system leveraging deep learning and error-correction coding to create cancelable, secure biometric templates with high matching accuracy.

Contribution

It presents a new feature-level fusion framework with two architectures for robust multibiometric representation and a method to generate secure, cancelable templates using error-correcting codes.

Findings

Achieves state-of-the-art matching performance on a multimodal database.

Provides cancelability and security for biometric templates.

Demonstrates robustness of the proposed fusion architectures.

Abstract

In this paper, we propose a secure multibiometric system that uses deep neural networks and error-correction coding. We present a feature-level fusion framework to generate a secure multibiometric template from each user's multiple biometrics. Two fusion architectures, fully connected architecture and bilinear architecture, are implemented to develop a robust multibiometric shared representation. The shared representation is used to generate a cancelable biometric template that involves the selection of a different set of reliable and discriminative features for each user. This cancelable template is a binary vector and is passed through an appropriate error-correcting decoder to find a closest codeword and this codeword is hashed to generate the final secure template. The efficacy of the proposed approach is shown using a multimodal database where we achieve state-of-the-art matching performance, along with cancelability and security.