SECURITY ENHANCEMENT OF AN IMAGE ENCRYPTION SCHEME BASED ON CHAOTICALLY COUPLED CHAOTIC MAPS

Abstract

This research aims to enhance the security of image encryption scheme using the application of chaos theory. Due to some intrinsic features of the images, such as bulk data capacity and high correlation among pixels, the traditional encryption techniques such as AES, DES, RSA, etc. are not suitable for image cryptography. So chaos based image cryptography is introduced to overcome these drawbacks associated with traditional encryption techniques. This image encryption scheme encompasses confusion and diffusion stages. During confusion stage, three levels of shuffling using different chaotic maps are used. In the first level, cubic map is used and Arnold cat map in the second level followed by tent map. Finally the shuffled image is diffused using Henon map to produce the ciphered image for transmission. Sensitivity analysis is performed with respect to key to its security against brute-force attacks. The average entropy value of 7.97 provides vigorous upon entropy attack. The Number of Pixel Change Ratio (NPCR) of above 99% obtained from the diffusion stage signify the use of multiple chaotic maps. The image cryptosystem has 2128 different combinations of the secret keys. The maximum correlation values in vertical, horizontal and diagonal directions of encrypted images are 0.0826, 0.0772 and 0.0721 respectively and structured similarity index measurement of not more than 0.0152 between original image and encrypted image validates the robustness of this algorithm.