Hybrid DWT NLM method with NOA optimization for ECG signal denoising

Despite the hybrid Discrete Wavelet Transform+Non Local Mean (DWT+NLM) method's ability to leverage the processing strengths of high - and low - frequency components, it faces issues like translation transformations, modal aliasing, patch effects, and threshold - induced distortion. These problems undermine the accuracy of electrocardiogram (ECG) - based cardiovascular disease diagnosis. This study presents a Nutcracker Optimization Algorithm (NOA) - enhanced DWT+NLM framework. Using NOA, the framework dynamically optimizes wavelet decomposition levels and basis functions for precise high/low - frequency separation. It adaptively adjusts NLM parameters to mitigate the patch effect and introduces a sigmoid - tuned threshold function to eliminate constant deviation. Experiments conducted on Physionet datasets demonstrate that when mitigating Additive White Gaussian Noise (AWGN), the proposed method achieves a maximum Signal-to-Noise Ratio (SNR) gain of 2.42 dB and an average gain of 1.73 dB over the suboptimal approach specifically for AWGN. Notably, in real-world noise scenarios (Baseline Wander (BW), Muscle Artifact (MA), and Electrode Motion Artifact (EM)), the method delivers an average SNR enhancement of 3.12 dB compared to the second-best method, underscoring its robust adaptability and practical superiority in noisy environments. This research holds promise for integration into wearable ECG sensors, potentially enhancing the diagnostic accuracy of cardiovascular diseases.