FPGA-Based Image Encryption Using ANN-Driven Lorenz Chaotic System
Year: 2026
Type: conference-paper
Abstract: Chaotic systems have gained significant attention in cryptography due to their inherent randomness, sensitivity to initial conditions, and deterministic unpredictability. Among these, the Lorenz attractor is widely used for pseudo-random number generation and secure communication. However, conventional hardware implementations rely on numerical solvers such as Runge Kutta (RK4), which introduce high computational complexity, increased power consumption, and latency.This paper presents an FPGA-based image encryption system using an Artificial Neural Network (ANN) driven Lorenz chaotic generator. The ANN model, trained offline on numerically generated Lorenz data, eliminates the need for solving differential equations in real time. The trained model is implemented on FPGA to generate chaotic sequences efficiently.The proposed encryption scheme integrates chaotic sequences into a two-stage process: confusion using the Arnold Cat Map (ACM) and diffusion using XOR based transformation. The system ensures strong encryption performance while maintaining low hardware resource utilization. Experimental results demonstrate that the proposed method achieves high randomness, low pixel correlation, and improved hardware efficiency compared to traditional approaches. The work highlights a novel integration of machine learning and hardware cryptography for real time secure image processing.
Institution: Delhi Technological University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7169874591
