Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Classification of prosthetic heart valve sounds. A parametric approach

Journal Article · · Journal of the Acoustical Society of America
DOI:https://doi.org/10.1121/1.412385· OSTI ID:72386
;  [1]
  1. Univ. of California, Livermore, CA (United States)
People with heart problems have had their lives extended considerably with the development of the prosthetic heart valve. Great strides have been made in the development of the valves through the use of improved materials as well as efficient mechanical designs. However, since the valves operate continuously over a long period, structural failures can occur-even though they are relatively uncommon. Here the development of techniques to classify the valve either as having intact struts or as having a separated strut, commonly called single leg separation, is discussed. In this paper the signal processing techniques employed to extract the required signals/parameters are briefly reviewed and then it is shown how they can be used to simulate a synthetic heart valve database for eventual Monte Carlo testing. Next, the optimal classifier is developed under assumed conditions and its performance is compared to that of an adpative-type classifier implemented with a probabilistic neural network. Finally, the adaptive classifier is applied to a data set and its performance is analyzed. Based on synthetic data it is shown that excellent performance of the classifiers can be achieved implying a potentially robust solution to this classification problem. 21 refs., 11 figs., 1 tab.
Sponsoring Organization:
USDOE
DOE Contract Number:
W-7405-ENG-48
OSTI ID:
72386
Journal Information:
Journal of the Acoustical Society of America, Journal Name: Journal of the Acoustical Society of America Journal Issue: 6 Vol. 97; ISSN 0001-4966; ISSN JASMAN
Country of Publication:
United States
Language:
English