| ECG examination, which is one of the clinical routine examinations, is of great significance for the early detection and treatment of heart disease. From ECG, doctors can get the required clinical information and make the diagnosis and the treatment. The automated identification of ECG clinical information that is named as ECG automated analysis techniques is the key of the transition from 12-lead ECG to dynamic ECG.Now, ECG automated analysis can not completely replace manual analysis, most of them are only used to assist experts to diagnose the patients. The main reason is that the existing techniques lack accuracy and can't fully meet the requirements of clinical application. Therefore, improving the accuracy and utility of ECG automated analysis system is of great realistic significance for automated analysis.First of all, a new ECG filtering algorithm is used based on Hilbert-Huang Transform(HHT) in this paper, after studying the current ECG algorithms. It is a better way to filter baseline wander, EMG interference and frequency interference, with good adaptability and robustness. Then, a method for QRS detection with the modified integral threshold algorithm reduces the error rate effectively. On this basis, we use extremum searching method to detect Q, S waves and P, T waves, locate the onset and offset of waves, and calculate the corresponding characteristic parameters. In the end, based on the calculated RR interval and QRS wave width, twelve kinds of common arrhythmia are classified using the logical branch judgement method. The classification algorithm is important for clinical application with high precision and fast analysis speed.In this thesis, the algorithms are designed on Matlab platform, and evaluated using MIT-BIH Arrhythmia database. It is proved that the algorithms have high reliability and utility. The technology of ECG automated analysis is a huge task, and there are many aspects needing further researchs and improvements. |
PDF Full Text Request