Two examples of audible percussion signals collected from healthy volunteers: (a) signal collected over the left subclavicular area (resonant character); (b) signal collected over the left abdominal area (tympanic” character).
Fourier amplitude spectra of the signals shown in Fig. 1. The EDS from Tables I and II are superimposed on the spectra. Each EDS is represented by a vertical line of height (Ai/di ) at the corresponding frequency fi .
Signal reconstruction error from all M EDS after the SVD truncation.
(a) Reconstruction of the original resonant signal from the increasing number of EDS components sorted by the amplitude. (b) Reconstruction of the original tympanic signal from the increasing number of EDS components sorted by the amplitude.
PRD error between the original signal and the signal reconstructed from a subset of EDS sorted by the value of A/d (squares) and resorted according to the above algorithm (circles). (a) Resonant signal from Fig. 1; (b) tympanic signal.
Frequency distribution of the dominant EDS for the resonant and tympanic signals.
Distribution of the quality factor of the dominant EDS for resonant and tympanic signals.
Distribution of the resonant and tympanic dominant EDS in the (f,Q) space.
MPM decomposition results for the resonant signal from Fig. 1(a).
MPM decomposition results for the tympanic signal from Fig. 1(b).
Article metrics loading...
Full text loading...