Font Size: a A A

Location Algorithm Design For Pacemaker Of Premature Ventricular Contraction In Eletrocardiogram Signal

Posted on:2019-06-08Degree:MasterType:Thesis
Country:ChinaCandidate:M Z LinFull Text:PDF
GTID:2404330590475464Subject:Integrated circuit engineering
Abstract/Summary:PDF Full Text Request
Cardiovascular disease is one of the leading causes of morbidity and mortality world widely,relative diagnosis techniques are highly developed,one of which is Electrocardiogram technology.Electrocardiogram technology is a kind of non-invasive cardiac disease diagnostic method.Premature contraction is a common type of cardiac disease or arrhythmia.Especially,researches on premature ventricular contraction(PVC)are gaining tremendous attentions because of its high incidence.This paper designs and implements location algorithm for pacemaker of PVC based on its ECG characteristics,it includes detecting PVC's happening and identifying PVC's pacemaker location.This research has guiding significance in diagnosis and treatment of PVC,its main work and achievements are as following:1)Implement and improve algorithm for detecting beats of PVC.This paper designs and implements PVC detecting algorithm based on template matching algorithm,which is robust to noise attack and has an excellent overall performance.The test result of MIT-BIH database shows that the algorithm's sensitivity?specificity and accuracy are 94.97%?97.87% and 97.72%.2)design and implement location algorithm for pacemaker of PVC.Based on the medical diagnosis rules of detecting the location of PVC's pacemaker,this paper extracts PV C's ECG characteristic zone,which is used to train CART decision tree on locating the pacemaker of PVC.This decision tree is used for three classification and five classification,whose accuracy are 90.63% and85.42%.The algorithm provides effective guidance and helps doctors determine the best medical schemes to cure PVC.
Keywords/Search Tags:ECG, PVC, location algorithm, template matching, convolution neural network, CART decision tree
PDF Full Text Request
Related items