Deep Secure Encoding: An Application to Face Recognition

TL;DR

This paper introduces Deep Secure Encoding, a deep learning-based framework for face recognition that enhances security and cancelability of biometric templates by mapping faces to high-entropy codes and hashing them, achieving state-of-the-art results.

Contribution

It proposes a novel deep neural network approach for secure face recognition that combines high-entropy coding with hashing for improved security and performance.

Findings

Achieves state-of-the-art accuracy on CMU-PIE and Yale B datasets.

Provides cancelability and high security without unrealistic assumptions.

Works effectively in both identification and verification modes.

Abstract

In this paper we present Deep Secure Encoding: a framework for secure classification using deep neural networks, and apply it to the task of biometric template protection for faces. Using deep convolutional neural networks (CNNs), we learn a robust mapping of face classes to high entropy secure codes. These secure codes are then hashed using standard hash functions like SHA-256 to generate secure face templates. The efficacy of the approach is shown on two face databases, namely, CMU-PIE and Extended Yale B, where we achieve state of the art matching performance, along with cancelability and high security with no unrealistic assumptions. Furthermore, the scheme can work in both identification and verification modes.