Methods for Automatic Heart Sound Identification

• Author / Creator

  Joya, Michael

• This thesis provides a description of the cardiac rhythm as a latent chain of heart sound arrivals which occur over time, where each arrival generates a fixed window of observable data that can be described with arbitrary feature functions. This description of the process produces tractable procedures for inference of timing parameters and estimation of the most likely chain of arrivals. It is shown that the central obstacle for accurate estimation is that the timing of the arrivals for a particular subject will often differ substantially from those of the pooled sample, often resulting in poor estimates. One of the theoretical contributions of this work is a method for estimating the unique timing parameters of the rhythm through the use of signal filtration applied directly to the observed data. This technique is effective at modeling the distribution of these parameters for recordings with repetitious patterns in the signal.

• Subjects / Keywords

  • Auscultation
  • Machine learning
  • Signal processing

• Graduation date

  Fall 2012

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/R37S9C

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.