Chaos Based Novel Video Encryption Scheme to Secure Video/Image Information from Any AI-Generated Attacks

Abstract

With the rise of data-intensive activities, methods such as video encryption have gained prominence, paving the way for quantum encryption and AI-powered strategies. It ensures content protection, supports compliance, and enables controlled sharing across sectors like security, satellite & medical imaging, and E-commerce. This study introduces a novel efficient and guarded scheme for encrypting videos as well as images using a 4D hidden chaotic system and application in AI-developed real-time video. A chaotic system is used for developing a unique scrambling process and for the diffusion process. Such an encryption method is rare in recent findings. Various performance analyses are employed to assess the efficiency and resilience of the algorithm. The large key range enhances the algorithm's security against brute force attacks and the careful adjustment of keys empowers the chaos to create different sequences of pseudo-random numbers, which significantly influences the proposed algorithm. The correlation coefficient is close to zero and NPCR is at least 99.62%, UACI is at least 33.50%. After encryption, the histogram is uniformly distributed, key space is extensive with 10^135, and the three-dimensional intensity plot of the encrypted image also exhibits uniformity. Moreover, the encryption algorithm can overcome several noise attacks such as salt and pepper noise, Gaussian noise additionally provides strong resistance against clipping attacks, underscoring the algorithm's effectiveness and resilience.