- 期刊
Speech Encryption Scheme Based on BFV Homomorphic Encryption
摘要
To solve the problems of traditional speech encryption schemes in the cloud storage, such as data leakage, low efficiency of speech homomorphic encryption, and significant expansion of ciphertext data, a speech encryption scheme based on the Brakerski / Fan-Vercauteren (BFV) homomorphic encryption was proposed. Firstly, in order to reduce the computational complexity of homomorphic encryption, the original speech data was segmented by fixed-length segmentation technology to generate a series of two-dimensional arrays. Secondly, BFV homomorphic encryption algorithm was used to batch coding and encrypt the segmented speech arrays to improve the efficiency of speech homomorphic encryption. Finally, improved the ciphertext calculation algorithm and combined with the modulus switching and linearization technology to perform ciphertext operation on the encrypted speech data, which can reduce the length of the resulting ciphertext and ensure the correctness of the decryption. Experimental results show that the proposed scheme can resist the selective plaintext attack with low computational complexity and ciphertext expansion. The proposed scheme has higher security and encryption-decryption efficiency compared with the existing probabilistic statistics and additional homomorphic cryptosystem. It can be used to store speech data in cloud computing securely.