The paper is devoted to the ECG-dedicated compression algorithm based on the event-driven variable quantization level in three upper octaves of the time-frequency signal representation. The algorithm uses an integer-to-integer reversible wavelet transform and the segmentation procedure developed for diagnostic purpose. Our method was implemented in Matlab and tested against the world-standard databases. Although the global compression efficiency and distortion ratio are not outstanding comparing to other compression methods, the main advantage of our method is the concentration of distortions out of the medically most important areas. For this reason, from the medical point of view, our method guarantees high fidelity of reconstructed signal and, in consequence, high reliability of signal-derived diagnostic parameters. The other advantage is that the algorithm uses integer-represented values only, that simplifies the implementation in a clinical-use real-time recorder.
[1]
Piotr Augustyniak.
The dynamic range o fan ECG in the time-frequency domain used for the lossless signal compression
,
2000
.
[2]
M Demeester,et al.
Assessment of the performance of electrocardiographic computer programs with the use of a reference data base.
,
1985,
Circulation.
[3]
W. A. Coberly,et al.
ECG data compression techniques-a unified approach
,
1990,
IEEE Transactions on Biomedical Engineering.
[4]
I. Daubechies,et al.
Wavelet Transforms That Map Integers to Integers
,
1998
.
[5]
Fayn Jocelyne.
L'Analyse séquentielle des électrocardiogrammes
,
1990
.
[7]
A. Cohen,et al.
ECG compression using long-term prediction
,
1993,
IEEE Transactions on Biomedical Engineering.