Learning-based Denoising Algorithm for the Reconstructed Image using Electromagnetic Emanations from the Display Device

This paper proposes a learning-based denoising algorithm that improves the signal-to-noise ratio (SNR) of the information signal emitted from the display device. The information signal is easily degraded by noise and interference on the channel and has various SNR. In this situation, an algorithm that enhances the model's robustness is required to improve the degraded information signal into a learning-based denoising model. Therefore, this paper proposes a normalization method to enhance the robustness of the model.