Soft Computing   2016 年

2016.02.12

Private image data, especially including the biometric data with an authentication property, has received more and more attention along with the development of researches on big data. Consequently, to protect the private image data while enabling outsourced image computations becomes a major concern. In this present paper, we study the privacy-preserving face recognition by using a method that is different from the method of fuzzy classification recognition, which is scale-invariant feature transform (SIFT) as our key technical tool. We first propose a scheme in which the client encrypts his private image data locally and outsources the corresponding results to a company. The latter performs most of the computations, but remains ignorant of the original data. To prevent some potential adversary from forging the data, we introduce a third party who can decrypt any given valid ciphertext. Based on these ideas, we adopt the BCP double-decryption cryptosystem for our scheme. Some analyses show that our proposed scheme is secure, efficient and scalable.