Loading...

A Dynamical Model for Generating Synthetic Phonocardiogram Signals and Model-based Processing

Almasi, Ali | 2012

777 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 42636 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Shamsollahi, Mohammad
  7. Abstract:
  8. In this thesis a dynamical model is introduced for Phonocardiogram based on its morphology which is capable of generating synthetic signals with realistic morphology. The model is for normal Phonocardiograms which includes the two dominant heart sounds, namely the first and the second sounds, and is inspired by the Electrocardiogram dynamical model. In the proposed dynamical model each heart sound is modeled with several Gabor kernels. The ultimate goal of such dynamical model is to establish a model-based processing framework for Phonocardiogram. This framework is devised by employing the dynamical model equations within an Extended Kalman Filter structure, and the simultaneously recorded Electrocardiogram signal is used as an input too. The introduced model-based framework is then further discussed and is employed for the applications such as denoising, heart sound detection and localization, as well as compression. Finally, the proposed method is evaluated on a database of PCG signlas from healthy subjects and the results are reported qualitatively and quantitatively. As for denoising, the proposed model-based framework demonstrates better performance than Wavelet Denoising (WD) method, over a wide range of PCG SNRs. Especially in low SNR near 0dB, the results of the proposed method show about 1dB more SNR improvement with respect to WD. The quantitative results of PCG compression show that using the proposed model-based framework, we can reach a compression rate (CR) of 24.8, about two and a half times more than CR of wavelet-based PCG compression method, while the reconstruction error of both methods are nearly %12. The results show that the model-based framework is comparable to other methods in such contexts. Therefore, the proposed framework can serve as a novel and effective tool for Phonocardiogram processing applications
  9. Keywords:
  10. Dynamic Modeling ; Kalman Filters ; Compression ; Noise Removing ; Detection ; Heart Sound Localization ; Phonocardiogram ; Model-Based Processing

 Digital Object List

 Bookmark

...see more