R-peak detection and ECG data compression scheme based on empirical mode decomposition and wavelet transform

As a crucial foundation for diagnosing cardiovascular conditions, electrocardiogram (ECG) signals play a pivotal role in clinical practice. Given the exponential growth of medical data and the importance of accurate clinical diagnosis, lossy compression of ECG signals has become a key area of focus. Despite notable advancements in lossy compression technology, there is still room for improvement in the balance between its efficiency and distortion.

This study focuses on a novel signal compression scheme that integrates empirical mode decomposition, and discrete wavelet transform. The composition of the ECG signal under empirical mode decomposition determines the appropriate recombination method. Subsequently, threshold processing of the discrete wavelet transform is utilized for the initial quantization of the coefficients. Following this, the small coefficients near zero are replaced by dead-zone quantizer. Moreover, the Pan-Tompkins algorithm is employed to analyze the QRS complexes both prior to and following the reconstruction of the ECG signal, thereby assessing the efficacy of the compression scheme.

Extensive simulation results corroborate that this scheme achieves superior compression ratios while preserving the integrity of the ECG signal to a significant extent. In the context of the MIT-BIH Arrhythmia Database, the scheme demonstrated average performance metrics of 23.50 for compression ratio, 5.66 for quality score, and 22.38 for signal-to-noise ratio.

Our scheme is an optimization of the ECG compression scheme based on empirical mode decomposition. It is believed that this compression scheme can enhance overall performance and provide valuable insights for practical applications.