Secure and Efficient Outsourcing of PCA-Based Face Recognition

Face recognition has become increasingly popular in recent years. However, in some special cases, many face recognition calculations cannot be performed effectively due to the lack of sufficient computing power of the terminal, which poses a challenge to the practical application of face recognition technology. Cloud computing provides a good platform for solving this problem due to its abundant computing resources. However, cloud computing poses new challenges, such as how to protect clients' data privacy without reducing efficiency. In this paper, we review some of the results of previous research and analyze an outsourcing protocol for eigen decomposition and singular value decomposition. On this basis, we propose a secure and efficient outsourcing protocol for face recognition through principal component analysis. In the proposed protocol, information privacy is well protected, and computational resources are saved by means of conversions of the original image information. In addition, local verification is supported to cope with the laziness of the cloud. We show the feasibility and advancement of our protocol from both theoretical and experimental perspectives.

I. Introduction

Face recognition is a biometric recognition technology based on facial feature information that has a friendly human-machine interface, accurate identification, and high efficiency [1]. Face recognition systems have an extremely wide range of real-world applications, such as identity authentication [2], access control management [3], and public security [4]. However, face recognition algorithms generally have extremely high computation complexity. As the recognition scale increases, the dimension of the resulting matrix can reach thousands or even millions. In the case where the computing resources of the terminal device are severely limited, it is basically impossible to perform such complicated operations locally.