A hardware-friendly Chialvo model incorporating electromagnetic flux for image encryption and neural network analysis

In this paper, a hardware-friendly version of the Chialvo neural model, named Exponential to Approximate Chialvo (E2A-Chialvo), is presented. This model simplifies computations by replacing complex nonlinear functions with approximations based on binary power functions (2 n ), resulting in a significant reduction in hardware resource consumption. The proposed model can accurately reproduce synchronized, desynchronized, and chimera behavioral patterns in ring, star, and ring-star hybrid networks. The successful implementation of the model on an Artix-7 Field-Programmable Gate Array (FPGA) chip shows that the total hardware resource usage of the proposed model (11.75%) is reduced by more than 50% compared to the original Chialvo model versions (approximately 24.7%). In addition, an image encryption algorithm based on the three-dimensional nonlinear sequences of the E2A-Chialvo model is proposed. This algorithm utilizes the complex and nonlinear features of the sequences to effectively hide information within images and provides high security against statistical analysis. Security evaluation metrics, including NPCR (Number of Pixel Change Rate) (99.6175%, 99.6088%) and UACI(Unified Average Changing Intensity) (30.6202%, 30.6202%), confirm the algorithm’s high sensitivity to input changes and its effectiveness in concealing the image’s information.