9 January 2008 A novel feature extracting method of QRS complex classification for mobile ECG signals
Author Affiliations +
Proceedings Volume 6794, ICMIT 2007: Mechatronics, MEMS, and Smart Materials; 67943M (2008) https://doi.org/10.1117/12.783965
Event: ICMIT 2007: Mechatronics, MEMS, and Smart Materials, 2007, Gifu, Japan
Abstract
The conventional classification parameters of QRS complex suffer from larger activity rang of patients and lower signal to noise ratio in mobile cardiac telemonitoring system and can not meet the identification needs of ECG signal. Based on individual sinus heart rhythm template built with mobile ECG signals in time window, we present semblance index to extract the classification features of QRS complex precisely and expeditiously. Relative approximation r2 and absolute error r3 are used as estimating parameters of semblance between testing QRS complex and template. The evaluate parameters corresponding to QRS width and types are demonstrated to choose the proper index. The results show that 99.99 percent of the QRS complex for sinus and superventricular ECG signals can be distinguished through r2 but its average accurate ratio is only 46.16%. More than 97.84 percent of QRS complexes are identified using r3 but its accurate ratio to the sinus and superventricular is not better than r2. By the feature parameter of width, only 42.65 percent of QRS complexes are classified correctly, but its accurate ratio to the ventricular is superior to r2. To combine the respective superiority of three parameters, a nonlinear weighing computation of QRS width, r2 and r3 is introduced and the total classification accuracy up to 99.48% by combing indexes.
© (2008) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Lingyun Zhu, Lingyun Zhu, Dong Wang, Dong Wang, Xianying Huang, Xianying Huang, Yue Wang, Yue Wang, } "A novel feature extracting method of QRS complex classification for mobile ECG signals", Proc. SPIE 6794, ICMIT 2007: Mechatronics, MEMS, and Smart Materials, 67943M (9 January 2008); doi: 10.1117/12.783965; https://doi.org/10.1117/12.783965
PROCEEDINGS
5 PAGES


SHARE
Back to Top