Color image encryption using non-dominated sorting genetic algorithm with local chaotic search based 5D chaotic map

The secure key generation is the predominant requirement of an image encryption. Chaotic maps are often considered by the researchers for secure key generation. However, chaotic maps suffer from hyper-tuning issue because the requirement of initial parameters. Therefore, an integrated non-dominated sorting genetic algorithm and local chaotic search based image encryption technique is proposed to tune the hyper-parameters of 5D chaotic map (TFCM). To implement TFCM, initially, the input image is decomposed into sub-bands using a dual-tree complex wavelet transform (DTCWT). These sub-bands are then diffused using the secret key obtained from the optimized 5D chaotic map. Finally, the inverse DTCWT is applied to obtain the final encrypted image. However, TFCM is computationally extensive for images with a larger size. Therefore, a parallel implementation of TFCM is also considered. Experimental analyses show that TFCM outperforms the competitive techniques in terms of NPCR, entropy, PSNR, and UACI by 0.9572%, 1.1576%, 1.0373%, and 1.0854%, respectively.