An efficient hybrid encryption model based on deep convolutional neural networks, deoxyribonucleic acid computing and chaotic system

Abstract

In this research, we present a substantial image encryption system constructed on deep convolutional neural networks for key generation to reduce processing power, complexity, and time consumption, as well as a creative method of DNA sequence operations and scrambling to encrypt digital images. A two-dimensional Logistic map is used to perform a chaotic scrambling. The results of the experiments, as well as security analysis such as statistical, differential, and key space analysis, demonstrated that the suggested method may achieve a high level of security while remaining efficient. With UACI and NPCR of 33.4 and 99.6, respectively, the method achieves an average entropy of 7.999 and a near-zero correlation. The algorithm’s efficiency is also evaluated to the state of the art in encryption techniques.