Abstract

Info

Author(s):: HU Danqin; CAI Wenjie; School of Medical Instrument and Food Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China

Abstract:: Electrocardiogram (ECG) is an important way to diagnose various heart diseases, and the accurate identification of QRS complex is required for automatic electrocardiogram analysis. The traditional methods for detecting QRS complex mainly include differential threshold method, detection algorithm based on double-threshold, empirical mode decomposition method, wavelet transform algorithm, etc. The main steps of these algorithms contain preprocessing, feature extraction and detection. The traditional methods have a poor generality and a high requirement of ECG signal quality. Compared with traditional methods for detecting QRS complex, the development of artificial intelligence, especially the emergence of deep learning, provides a new method for QRS complex detection. Deep learning can be used to independently extract QRS complex feature information, thereby realizing precise positioning. Compared with traditional methods, deep learning has a better robustness and a better detection effect for the data with poor signal quality. Herein the technologies used in the preprocessing and detection of QRS complex detection are reviewed, and the future developments are discussed.