Abstract: The present disclosure provides a GPU-based parallel electrocardiogram signal analysis method, comprising: performing a filtering process of electrocardiogram signals through a long interval artifact removal and a short interval artifact removal; performing a QRS detection of the filtering-processed electrocardiogram signals through an R-wave position extraction, a QRS complex start and end positions extraction and a QRS complex width extraction; performing an abnormal waveform classification of the QRS-detected electrocardiogram signals through template creation; wherein at least one of the long interval artifact removal, the short interval artifact removal, the R-wave position extraction, the QRS complex width extraction and the creation template is performed by a multiple threads at a GPU device side in parallel, any thread being read through its unique index number to process corresponding data.

Abstract: The present disclosure provides a GPU-based parallel electrocardiogram signal analysis method, comprising: performing a filtering process of electrocardiogram signals through a long interval artifact removal and a short interval artifact removal; performing a QRS detection of the filtering-processed electrocardiogram signals through an R-wave position extraction, a QRS complex start and end positions extraction and a QRS complex width extraction; performing an abnormal waveform classification of the QRS-detected electrocardiogram signals through template creation; wherein at least one of the long interval artifact removal, the short interval artifact removal, the R-wave position extraction, the QRS complex width extraction and the creation template is performed by a multiple threads at a GPU device side in parallel, any thread being read through its unique index number to process corresponding data.