PROJECT TITLE :

Efficient Privacy-preserving Outsourced Discrete Wavelet Transform in the Encrypted Domain

ABSTRACT:

Signal Processing in an encrypted domain is a potential tool that could protect sensitive signals from untrusted cloud servers and unauthorized users in a delegated computing setting. This would be accomplished without compromising the accuracy of large-scale signal analyzing and processing. The majority of the currently available methods encrypt each signal in a large bundle using Paillier's public key additively homomorphic encryption. This results in significant computational costs being incurred at local devices, which frequently have limited resources, while only guaranteeing the privacy of signal inputs. This paper proposes an efficient privacy-preserving outsourced discrete wavelet transform scheme (PPDWT), consisting of PPDWT-1 and PPDWT-2, that does not leverage public key (fully) homomorphic encryption in order to address these limitations. PPDWT stands for privacy-preserving outsourced discrete wavelet transform. Specifically, PPDWT-1 is proposed to achieve signal input privacy against the collusion between the honest-but-curious cloud and unauthorized users, and PPDWT-2 is proposed to protect both signal input privacy and coefficient privacy against collusion attacks. Both of these protections are provided by the PPDWT-1 and PPDWT-2 proposals. In order to encrypt batch signals, both constructions make use of the offline execution of any one-way trapdoor permutation only once. This makes it possible for signal processing to take place within the encrypted domain. Only users who have been granted permission can successfully decipher the results of the discrete wavelet transform using our method. Our method only incurs O(1) computational complexity, which is independent to the size of the signal inputs |l|, in contrast to the O(|l|) computational complexity on the user's end in existing state-of-the-art public key homomorphic encryption-based techniques. In the encrypted domain, we also discuss the expanding factor, the upper bound, and various extensions to the privacy-preserving discrete cosine/fourier transform. In conclusion, the universal composability (UC) model has provided a formal demonstration that our proposed PPDWT is secure. Case studies are then utilized in our analysis of the proposed methodology in order to demonstrate the methodology's viability and efficacy.