Secure Evaluation of Discrete Sine Transform in Homomorphic Encrypted Domain

Huicong Zeng, Zhiwei Cai, Peijia Zheng, Hongmei Liu, Weiqi Luo

Abstract: Discrete sine transform (DST) is widely used in compression encoding, image processing, and speech enhancement. In cloud computing, outsourcing plaintext data to the cloud has the risk of private data leakage. In this paper, we study how to realize data computation outsourcing with privacy protection. We propose a scheme to implement DST in the encrypted domain that uses the scaling method to represent decimals. To improve the accuracy, we also propose a new scheme to implement high precision encrypted domain DST. We approximate complex numbers with high precision using integer coefficient polynomials whose independent variable is a unit root. With this representation, we can perform DST in the encryption domain with high precision. We conducted experiments to verify the effectiveness and efficiency. The two schemes have the advantages of high speed and high precision, respectively.