Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Stationary nonstationary signal separation with use of wavelet packets transform and fractal dimension.


I.A. Fotiou, L. Hadjileontiadis

Aristotle University of Thessaloniki, vol. AUT03-19, Dept. of Electrical and Computer Engineering

In this paper, we present a new method for the separation of the stationary and non-stationary signal part with application to the enhancement of explosive lung sounds (LS) and bowel sounds (BS). This new method is based on wavelet packet analysis and fractal dimension metrics. These two elements are combined to form a WP-FD filter for the enhancement and separation of the explosive LS (ELS) and explosive BS (EBS) form the background stationary signal. Analysis of the numerical results obtained by applying the aforementioned method to real signals from patients with pulmonary or gastrointestinal dysfunction indicate that the WP-FD filter performs very efficiently in avoiding the adverse effect of the presence of the background stationary signal (noise). Additionally, the WP-FD filter has no need of any noise reference signal or template and is independent of the subjectivity, intrinsic to any human-made judgment. The robustness of the WP-FD filter is established through noise stress test of various noise contamination levels. This certifies the applicability of the filter to everyday clinical medicine, where the environments are usually characterized by an increased amount of noise. Data-volume reduction applications are also made possible with the use of WP-FD filter, since it separates the desired, information-bearing signal, form the underlying unwanted one. This is particularly useful in intensive care units or sleep laboratories, where long-term recordings are employed. Moreover, the accurate performance of the WP-FD filter enables it to be used as an objective tool, assisting the physician in the diagnosis procedure and in the interpretation of the auscultation findings. It is simple enough to be implemented in a real time context and can be used for continuous ELS and EBS screening. Keywords: Lung sounds, bowel sounds, explosive character, noise reduction, structure extraction, wavelet packets transform, fractal dimension thresholding.


Type of Publication:

(04)Technical Report

File Download:

Request a copy of this publication.
(Uses JavaScript)
% Autogenerated BibTeX entry
@TechReport { FotHad:2003:IFA_2095,
    author={I.A. Fotiou and L. Hadjileontiadis},
    title={{Stationary  nonstationary signal separation with use of
	  wavelet packets transform and fractal dimension.}},
    address={Aristotle University of Thessaloniki},
Permanent link