Cardiac problems lead to significant health risks for nowadays’ society. Newly developed technologies, such as recording of the magnetic heart signal (magnetocardiograms), enable the passive monitoring of the heart activity. Still, the method is subjected to high amplitude magnetic interferences and reliable signal processing algorithms have to be developed. The present paper focuses on developing and enhancing wavelet based algorithms for the processing of magnetocardiograms, in order to offer accurate results for the diagnosis of cardiac problems. Different threshold values have been assigned for each wavelet decomposition level for performing the denoising better. The method is adaptive and based on minimizing Stein’s unbiased risk estimate for each level, combined with a wavelet thresholding method. The results have been tested on acquired data and are compared both graphically and statistically for a good evaluation of the algorithms’ performances.

Keywords: Biomedical signal processing, Wavelets, Stein’s unbiased estimate of risk, Adaptive filtering