Comparative Analysis of Image Encryption Techniques: Hyper Chaotic Maps, Jacobean Elliptic Maps, Reversible Cellular Automata, DNA-based Encryption, and XGBoost

Abstract

The fast-growing digital communication landscape demands secure image encryption research to protect sensitive visual information from potential cyber threats. Our research experimentally assesses the effectiveness of five sophisticated encryption methods—Hyper Chaotic Maps, Jacobean Elliptic Maps, Reversible Cellular Automata, DNA-based Encryption, and XGBoost (ML-based Encryption)—for securing images. The evaluation of each algorithm includes thirteen essential parameters which cover Key Sensitivity, Histogram Analysis, Correlation Coefficient (CC), Entropy Analysis, NPCR, UACI, PSNR, Encryption Time, Computational Efficiency, Accuracy and Reliability, Security and Robustness, Strengths, and Weaknesses. The tests show that all algorithms maintain strong security features but XGBoost leads in encryption efficiency and decryption quality while resisting cryptographic attacks better than the others. These findings showcase how machine learning-based encryption methods can propel the development of cutting-edge cryptographic technologies. The findings from this research offer valuable guidance for creating image encryption methods that balance enhanced security with computational efficiency in practical applications.