Fast decryption utilizing correlation calculation for BSS-based speech encryption system

Original BSS (blind source separation)-based speech encryption system utilizes BSS to perform decryption, but the complexity of BSS algorithms limits the decryption speed and its real-time application. Correlation calculation is an algebraic decryption method originally proposed for BSS-based image cryptosystem to overcome two limitations of BSS: permutation and scaling ambiguity and difficulty for separating strongly correlated images. Considering that there is similarity between encryption matrix for speech encryption and image encryption, this paper analyzes the correlation of speech signals with key signals, and then utilizes the correlation calculation to achieve speech decryption. The experiment results show that correlation calculation decryption nicely simplifies BSS-based speech encryption system, largely speeds up the speech decryption, and slightly improves the quality of decrypted speech signals.