Real-time chaotic video encryption based on multi-threaded parallel confusion and diffusion

Abstract

Due to the strong correlation among adjacent pixels, image encryption schemes typically perform multiple rounds of confusion and diffusion to protect images against various attacks. This is time-consuming and cannot meet the real-time requirements for video encryption. Existing works, therefore, realize video encryption by simplifying encryption process or selectively encrypting specific pixels, resulting in lower security compared to image encryption. This paper proposes a real-time chaotic video encryption strategy based on parallel computing. It splits video frame into sub-frames, creates a dedicated set of threads to concurrently perform confusion and diffusion operations on their respective sub-frames, and efficiently outputs encrypted frames. To assess its performance, two cryptosystems are implemented using different chaotic systems. Encryption speed evaluation demonstrates a significant acceleration in byte generation, confusion, and diffusion phases, enabling real-time encryption and decryption on different X86 platforms. The average encryption time is less than 42 ms, despite performing five rounds of confusion and diffusion operations on each frame. Statistical and security analysis prove that the deployed cryptosystems exhibit exceptional statistical properties and provide resistance to different attacks. Moreover, our method is adopted to implement a multi-user, real-time mobile video secure communication system using embedded systems, validating the feasibility of the proposed strategy for various application scenarios.