Hardware implementation of a robust image cryptosystem using reversible cellular-automata rules and 3-D chaotic systems

Abstract

In this contribution, a robust image cryptosystem built over a diffusion-confusion architecture is established. The design of a hardware pseudorandom number generator is achieved using two different chaotic systems: Lorenz chaotic system and Rössler chaotic system, exploited to build a single generic architecture. The proposed PRNG is designed using the Xilinx system Generator (XSG) tool and implemented onto an FPGA Zynq ZC 020. It showed high statistical properties as it passes all NIST statistical tests and possesses the best hardware performance. The generated key sequences from the PRNG were used in diffusion processes using XOR operation. Moreover, two extra generated bits were used in the permutation process that precedes the confusion. To further increase the complexity and the randomness of the proposed scheme the confusion process is established over reversible cellular automata. The hardware design of the suggested approach is set forth using XSG tool and implemented on FPGA Zynq ZC 020 working at high-level frequency of 184 MHz and a 23 552 mbps throughput value. Then, the hardware cryptosystem is tested over different plain images. To demonstrate the efficiency of this hardware cryptosystem, security analyses such as key sensitivity analysis and statistical analysis among others were performed to validate the security of proposed hardware-cryptosystem showing an entropy value of 7.9998. A comparative study with some recently suggested cryptosystems is established. Experimental results state that the suggested hardware cryptosystem performs better in terms of security and hardware efficiency such as power consumption, resources utilization and throughput.