Cryptanalysis and security evaluation of optimized algorithms for image encryption in deep optimal network

Cryptanalysis and security evaluation of optimized algorithms for image encryption in deep optimal network

연구 분야: Networking

학회: International Journal of Information Technology

Due to the enhancement of multimedia technology, numerous electronic devices are utilized and the most significant is cryptography which is responsible for securing the information in the cloud network. Traditional methods for image encryption in cryptography face several challenges such as limited resources and encrypted traffic. To tackle these limitations, this paper proposed the Deep Convolution-based Local Tasmanian Devil (DC-LTD) algorithm for efficient image encryption in cryptography. We utilize the Tasmanian Devil Optimization (TDO) with a local search strategy for enhancing the image encryption performance. The encrypted data is identified accurately by designing the image of the original data determined in cipher text. The efficiency of the DC-LTD method is validated by conducting experiments using Canadian Institute for Advanced Research 10 (CIFAR 10) dataset. The evaluation of the DC-LTD model is performed using various measures, including key generation time, encryption time, decryption time, F1-score, recall, precision, accuracy, and specificity as 4.51 s, 4.48 s, 0.0137 s, 97.86%, 97.81%, 97.93%, 98.65% as well as 97.87% respectively. The results demonstrated that the DC-LTD model outperforms other methods. Therefore, the DC-LTD method attains better results in image encryption. Furthermore, the optimized algorithms are integrated into the Adaptive Elliptic Curve Cryptography scheme and evaluate their impact on image encryption efficiency and security.