Noise Suppression Method Based on CBAM-DnCNN in EM Leakage Signals

The inadvertent electromagnetic (EM) emissions from computers and electronic devices pose a significant threat to information security. In the field of EM leakage detection., extracting useful information from weak EM signals while eliminating the interference of noise is crucial. This paper focuses on the image information leaked through EM emissions from displays., which are often incomplete and blurred due to low SNR and environmental noise interference. Various denoising methods have been proposed., but they tend to blur useful information. In this study., we propose the CBAM-DnCNN for denoising EM leakage signals. CBAM-DnCNN combines the network structures of CBAM and DnCNN to improve the denoising performance of EM signals through attention mechanisms and residual connection. Experimental results demonstrate that CBAM-DnCNN effectively suppresses EM noise while preserving effective information in the signal., compared to the original DnCNN and traditional noise suppression methods. The paper also describes the network structure and working principles of CBAM-DnCNN and quantitatively evaluates its performance using metrics such as SNR and RMSE. The experimental results highlight the significant advantages of CBAM-DnCNN in denoising EM leakage signals., offering promising prospects for applications in EM information leakage detection.