Light-Crypt: A Lightweight Image Encryption Algorithm Using Chaotic Maps and Ai-Generated Substitution Boxes

The proposed energy-efficient image encryption algorithm, Light-Crypt, utilizes a chaos-based cryptography approach integrated with artificial intelligence to opti-mize S-Boxes and P-Boxes. It employs a modified grassroots structure for key generation through the Lorenz chaotic system, enhancing confusion and diffusion processes via neural networks and reinforcement learning. Performance testing demonstrated Light-Crypt's robustness, achieving an entropy increase of 0.97 bits, NPCR around 99.64%, and UACI approximately 28.65%. Correlation anal-ysis indicated effective diffusion with coefficients near zero, while histogram analysis showed uniform pixel intensity distribution, enhancing resistance to sta-tistical attacks. The algorithm passed 8 out of 12 NIST SP800-22 randomness tests, confirming its security. Sensitivity analysis revealed that minor key changes resulted in 99.39% output variation, underscoring its strong avalanche effect. Overall, Light-Crypt offers a high level of optimization for AI-enhanced S-Boxes and chaos-based encryption, making it suitable for resource-constrained and real-time applications without significant trade-offs between security and computa-tional efficiency.