Adaptive image encryption based on twin chaotic maps

The information stored or shared via the internet is growing massively and includes images primarily. The images are vulnerable to attacks when transferred over the internet as they contain confidential information about a person. We propose an adaptive image encryption scheme for the security of images based on twin chaotic maps; Quadratic map and 2-dimensional chaotic Henon map. In the proposed encryption scheme, the Pseudo-Random Number (PRN) sequence for shuffling the pixels has been generated from the Henon map. The PRN sequence for the diffusion process has been produced from the classical Quadratic map. The significant contribution of the presented work is that both the chaotic sequences (for confusion and diffusion) are plain image dependent, making the encryption scheme adaptive and hence highly resilient to brute force attack. Image quality analysis, histogram analysis, correlation coefficient analysis, entropy analysis, key sensitivity analysis, and differential attack analysis have been carried out on eight natural images of size 512 × 512 to validate the performance of the proposed encryption scheme. The proposed encryption algorithm has correlation coefficient values that are very close to zero, entropy value of 7.9993 (average), Net Pixels Change Rate (NPCR), and Unified Average Changing Intensity (UACI) values of 99.62% (average) and 33.3% (average) respectively. The histogram of the encrypted images is flat, meaning no information can be deduced about the plain image. The comparison of the proposed encryption technique with other recent encryption schemes validates the performance of the proposed cryptosystem. The computational time taken to encrypt Lena image of size 512 × 512 using the proposed encryption scheme is 0.25 s.