An efficient image encryption using deep neural network and chaotic map

Inspite of progressive growth of cryptography, encrypting sensitive information of an image is still a computationally complex task. After reviewing existing literature, it is now known that security problems are yet not solved and there is an open scope of further research. In most recent times, it has been noticed that neural network has proven cost effective optimization mechanism in offering security towards images. However, such implementation are computationally expensive process and do not solve various diversified attacks on image. Hence, the prime purpose of proposed system is to introduce an analytical research methodology for presenting a sophisticated framework where deep neural network has been used for optimizing the performance of simple encryption approaches. The robustness of optimization principle is further added with chaotic map concept for enhanced security performance. The study outcome shows that proposed implementation offers much better security performance without any negative effect on image quality. (c) 2020 Elsevier B.V. All rights reserved.