Automatic rhythm analysis using SPART method

Author

Kowalski, HA ; Skorupski, A. ; Jedrasik, P. ; Wojciechowski, D. ; Kacprzak, P. ; Zych, R.

Author_Institution

Inst. of Comput. Sci., Warsaw Univ. of Technol., Poland

fYear

1999

fDate

6/21/1905 12:00:00 AM

Firstpage

515

Lastpage

518

Abstract

Accurate identification of ectopic beats is important in frequency HRV analysis. It may be done by classification of QRS planar loop morphology. The objective of this research was to compare results of QRS morphology analysis in different spatial planes: frontal, horizontal, and preferential. The preferential plane of a QRS loop was found with modified SPART method. Fourier descriptors of plane loop were used for classification of QRS loops morphology analysis. The study was performed using ICAR System and Multilead CSE Database. Identification of optimal spatial orientation of QRS loop using SPART method was accurate in all analyzed beats. Statistical indices showed low values for the QRS classification in the frontal plane XY. The results estimated for XZ and UV planes were comparable. Our results confirm usefulness of ICAR System with implemented SPART method for automatic rhythm analysis

Keywords

electrocardiography; medical signal processing; patient monitoring; pattern classification; statistical analysis; Fourier descriptors; ICAR System; Multilead CSE Database; QRS classification; QRS morphology analysis; QRS planar loop morphology classification; SPART method; UV plane; XZ plane; accurate identification; automatic rhythm analysis; classification; ectopic beats; frequency HRV analysis; frontal plane; frontal plane XY; horizontal plane; modified SPART method; optimal spatial orientation; plane loop; preferential plane; spatial planes; statistical indices; Biomedical engineering; Cardiology; Computer science; Electrocardiography; Heart rate variability; Hospitals; Image segmentation; Morphology; Rhythm; Spatial databases;

fLanguage

English

Publisher

ieee

Conference_Titel

Computers in Cardiology, 1999

Conference_Location

Hannover

ISSN

0276-6547

Print_ISBN

0-7803-5614-4

Type

conf

DOI

10.1109/CIC.1999.826021

Filename

826021