Real-time bit-level encryption of full high-definition video without pixel diffusion

Abstract

Despite the widespread adoption of Shannon’s confusion-diffusion architecture in image encryption, the implementation of diffusion to sequentially establish inter-pixel dependencies to attain plaintext sensitivity constrains algorithmic parallelism, while the need for multiple rounds of diffusion operations to meet sensitivity metrics incurs considerable computational overhead. Therefore, reliance on diffusion operations to achieve plaintext sensitivity is a primary factor limiting the computational efficiency and throughput of video encryption algorithms, rendering them unsuitable for real-time encryption of high-resolution video. To address this limitation, this paper proposes a real-time video encryption protocol based on heterogeneous parallel computing. The protocol incorporates the SHA-256 hashes of original frames as input, employs multiple CPU threads to concurrently generate encryption-related data, and utilizes numerous GPU threads to encrypt pixels simultaneously. By leveraging the high input sensitivity of the SHA hash, the protocol achieves the required plaintext sensitivity with a single round of confusion and XOR operations, thereby reducing computational overhead. By eliminating the reliance on diffusion, it enables the allocation of a dedicated GPU thread to encrypt each pixel within every channel, thus enhancing the algorithm’s parallelism. Experimental results show that the proposed approach not only exhibits desired statistical properties and robust security but also achieves delay-free bit-level encryption of 1920×1080 (Full HD) video at 30 FPS, with an average encryption time of 25.84 ms on a server equipped with an Intel Xeon Gold 6226R CPU and an NVIDIA GeForce RTX 3090 GPU. Additionally, the protocol is applied to a video monitoring system, enabling delay-free encryption of 640×480 video at 24 FPS on an NVIDIA Jetson Xavier NX with a Carmel ARM CPU and Volta GPU, demonstrating its practical feasibility and confirming that the desired metrics can be achieved without resorting to complex encryption procedures.