On-line classification of arterial stenosis severity using principal component analysis applied to Doppler ultrasound signals

D. R. Prytherch, D. H. Evans, M. J. Smith, D. S. Macpherson

Research output: Contribution to journalArticlepeer-review

Abstract

Principal component analysis is a powerful method of feature extraction which can be applied to continuous-wave Doppler waveforms. A microprocessor system for the on-line calculation of the coefficients of principle components has been devised and tested in an experimental model. Doppler waveforms were obtained from positions distal to stenoses of known severity implanted in the iliac arteries of three dogs and classified into one of four groups. By reference to data from a previous series of experiments the microprocessor correctly classified 75% of stenoses. The remaining 25% were all classified as being one group more severe than they actually were.
Original languageEnglish
Pages (from-to)191-200
JournalClinical Physics and Physiological Measurement
Volume3
Issue number3
DOIs
Publication statusPublished - 1982

Fingerprint

Dive into the research topics of 'On-line classification of arterial stenosis severity using principal component analysis applied to Doppler ultrasound signals'. Together they form a unique fingerprint.

Cite this