Title :
Hybrid hidden Markov models for ECG segmentation
Author :
Shi, Wu ; Kheidorov, Igor
Author_Institution :
Coll. of Mech. & Power Eng., Harbin Univ. of Sci. & Technol., Harbin, China
Abstract :
The new and reliable method of electrocardiogram segmentation was developed. This method has an important application to diagnostics of critical heart diseases and investigation of new drugs effect on heart. It uses powerful mathematical techniques including wavelet transforms, neural networks and hidden Markov models. The method was tested on signals from freely available QT database and showed the results, which are very close to electrocardiogram segmentation performed by heart specialists.
Keywords :
electrocardiography; hidden Markov models; medical signal processing; neural nets; wavelet transforms; ECG segmentation; critical heart diseases; electrocardiogram segmentation; hybrid hidden Markov models; neural networks; wavelet transforms; Databases; Electrocardiography; Heart; Hidden Markov models; Training; Wavelet transforms; Electrocardiogram; hidden Markov models; neural networks; segmentation; wavelet transforms;
Conference_Titel :
Natural Computation (ICNC), 2010 Sixth International Conference on
Conference_Location :
Yantai, Shandong
Print_ISBN :
978-1-4244-5958-2
DOI :
10.1109/ICNC.2010.5583618
