Enhanced color image security via rough set partitioning and chaotic map assignment
Hemang Sonvarsha, Sanjay Kumar, Deepmala Sharma · Journal of Cyber Security Technology · 2026
In light of the increasing usage of digital multimedia, there is a growing need for highly secure means of safeguarding color images against potential attacks. In this research, a novel approach of hybrid cryptography that combines rough set-based image segmentation with chaos encryption is introduced. In the proposed approach, the color image would be first segmented into its RGB components. Applying rough set theory to each of the components using Otsu Thresholding and morphological operators, including Erosion and Dilation operations would result in segmentation of each channel into Sure Regions, Possibility Regions and Uncertainty Regions. To improve encryption efficiency and performance, different mapping functions are applied to the segmented images as follows: the logistic map to Sure Regions, the Tent map to Possibility Regions and the 2D Henon map to the Uncertainty Regions. Experimental results confirm our encryption framework is highly effective. The ciphertext passes all NIST randomness tests, demonstrating strong resistance to statistical and differential attacks with an information entropy of 7.99952 (near 8), a correlation coefficient close to 0, and NPCR and UACI values up to 99.685% and 32.98%. The proposed approach provides a highly secure and robust solution for visual data protection in an evolving cyber-threat landscape.