Hybrid Image Encryption Approach Using XOR, AES, and Pixel Shuffling for Improved Security

Abstract

As digital communication continues to grow rapidly, there is an increasing need for secure and efficient techniques to send multimedia data, especially sensitive images. Symmetric cryptographic algorithms such as AES and XOR may be considered adequate for providing viable information security, but using them independently often sacrifices statistical strengths or increases computational load and cost. In this paper, we present a hybrid image encryption scheme that leverages three complementary approaches, including pixel shuffling, XOR diffusion, and AES confusion. The combination of shuffling, which is a very low overhead permutation approach, the XOR which is mathematically easy to work with, and AES which provides consistent strong non-linearity is able to produce an efficient encryption and decryption scheme with strong cryptographic qualities. A standard image, Lena RGB (256×256), was used for the experimental evaluation of the cryptographic scheme. The quantitative analysis characterized the implementation based on a variety of aspects, including entropy (8.0 theoretical), correlation coefficients, NPCR, UACI, and PSNR. The qualitative analysis included visual inspection and histogram analysis. Overall, results indicated that the hybrid scheme provided an improvement to stand-alone AES, XOR, and shuffling. Entropy approached 7.98, correlation coefficients were close to zero, NPCR and UACI estimates were 99.21% and 33.27%, respectively, demonstrating high resistance to a differential attack. However, PSNR estimates indicated perfect, lossless decryption. The hybrid implementation produced cipher images that exhibited flat histograms and high randomness. This hybrid strategy maintains strong cryptographic strength while allowing for a reasonable computational effort to locate secure image encryption mechanisms and will prove effective within implementation towards multimedia security and secure image transmission systems. Unlike traditional methods that independently utilize encryption techniques, this study presents a seamlessly integrated hybrid framework. It strategically merges pixel shuffling, XOR confusion and AES diffusion to enhance security while ensuring low computational overhead and lossless data reconstruction.