Hybrid Chaotic-DNA Cryptosystem With Galois Field Diffusion For Secure Color Image Encryption

Security issues in digital image transmission have become an important concern in multimedia communication. Because of their large image data volume, high redundancy and strong correlation among the neighboring pixels, the conventional encryption algorithms are not suitable for practical image encryption. In this paper, a chaotic color image encryption scheme based on Deoxyribonucleic Acid (DNA) coding calculation, multiplication arithmetic over the Galois field GF(17) and improved one-dimensional (1D) chaotic maps is presented. To address the limitations of classical 1D maps, including a narrow chaotic range and non-uniformity, three modified 1D chaotic maps, namely, Sine-Logistic Map (SLM), Chebyshev-Logistic Map (CLM) and Sine-Chebyshev Map (SCM) have been proposed. Furthermore, the chaotic initial values are modified using the plain image (the input color image is decomposed into Red, Green and Blue (R, G and B) planes, and bit-plane recombination is carried out), which makes it possible to defend against chosen-plaintext and known-plaintext attacks. In this scheme, the image pixels are first scrambled using chaotic sequences, and then DNA coding is applied. After DNA coding, DNA arithmetic operations are performed, and then DNA decoding is applied. In order to further diffuse the pixels, multiplication over GF(17) is applied. Security analysis, including histogram uniformity, information entropy, and inter-pixel correlation coefficients, demonstrates that the proposed scheme is secure.