Title :
Classification of QRS morphology in Holter monitoring
Author :
Dickhaus, H. ; Gittinger, J. ; Maier, C.
Author_Institution :
Dept. of Med. Inf., Heidelberg Univ., Germany
fDate :
6/21/1905 12:00:00 AM
Abstract :
Different strategies for classification of QRS morphology are developed and compared in respect to discriminatory power and computational effort on the MIT-BIH arrhythmia database. A classification procedure based on iterative correlation analysis yields better results than hierarchical clustering techniques on features that are derived from FFT-, DCT- and Hermite-coefficients. The advantage of the correlation approach is its good runtime performance and its possible applicability in cardiological routine analysis of 24 h ECG recordings
Keywords :
band-pass filters; electrocardiography; matched filters; medical signal detection; medical signal processing; patient diagnosis; pattern classification; pattern clustering; signal classification; DCT-coefficients; ECG recordings; FFT-coefficients; Hermite-coefficients; Holter monitoring; MIT-BIH arrhythmia database; QRS morphology; arrhythmia; cardiological routine analysis; classification procedure; computational effort; correlation approach; discriminatory power; hierarchical clustering techniques; iterative correlation analysis; runtime performance; Biomedical informatics; Biomedical monitoring; Cardiology; Clustering algorithms; Electrocardiography; Iterative algorithms; Morphology; Performance analysis; Signal analysis; Spatial databases;
Conference_Titel :
[Engineering in Medicine and Biology, 1999. 21st Annual Conference and the 1999 Annual Fall Meetring of the Biomedical Engineering Society] BMES/EMBS Conference, 1999. Proceedings of the First Joint
Print_ISBN :
0-7803-5674-8
DOI :
10.1109/IEMBS.1999.802339
