Title :
Classification of Vectorcardiograms Using Walsh And Cosine Orthogonal Transforms
Author :
Dyer, Stephen A. ; Ahmed, Nasir ; Hummels, Donald R.
Author_Institution :
Department of Electrical and Computer Engineering, Kansas State University, Manhattan, KS 66506
Abstract :
In this study, the usefulness of two-dimensional transforms in classifying human vectorcardiograms is investigated. The transforms used are the Walsh-Hadamard (WHT) and the discrete cosine (DCT). Experimental results included in the paper demonstrate that about 80-85 percent correct classification may be achieved.
Keywords :
Covariance matrix; Discrete cosine transforms; Discrete transforms; Electrocardiography; Feature extraction; Humans; Karhunen-Loeve transforms; Pattern recognition; Signal processing algorithms; Signal representations; Key Words¿Feature extraction; nearest-neighbor classifier; orthogonal transforms; pattern recognition; vectorcardiography;
Journal_Title :
Electromagnetic Compatibility, IEEE Transactions on
DOI :
10.1109/TEMC.1985.304243
