ABSTRACTWith the rapid advancements in artificial intelligence (AI), ensuring the privacy and security of patient medical images has emerged as a pressing concern in the field of image privacy protection. Traditional medical image encryption methods, however, have often been criticized for their lack of flexibility and insufficient security measures. To address these challenges, proposed a new image encryption algorithm that uses the R3 (rotate, rescale, randomize) Model to apply the concepts of chaotic and hyperchaotic systems to new image encryption approaches. Based on the hyperchaotic behavior of the circle map system and the unpredictable behavior of the Chua system, the proposed algorithm performs scaling, rotation, and randomization on the target image. Basically, the input medical image is scaled and rotated using a chaotic system to minimize the association between neighboring pixels. To create an encrypted image, the permuted image is next subjected to a diffusion operation with the assistance of S‐box. By using these chaotic systems unpredictable behavior and sensitivity to beginning conditions, an encryption technique with a huge key space of 25208 is created, improving security overall and fortifying its resistance to brute‐force attacks. The speed at which the algorithm processes data and the minimal number of resources it consumes demonstrate its effectiveness, making it appropriate for real‐time applications. To evaluate the security and computational efficiency of the suggested encryption system, a series of grayscale medical photographs were subjected to extensive testing. In addition to having a large key space of about 25208, it also showed resistance to differential cryptanalysis, with NPCR and UACI values above 99.60% and 30.20%, respectively. The encrypted images' entropy was close to eight, suggesting improved system security. The technique is ideal for real‐time applications, with encryption speeds of 0.80 s for 512 × 512 images and 0.6 s for 256 × 256 images. The scheme's capacity to provide strong security with little computing overhead and complexity is demonstrated by experimental results.
Read full abstract