Related Experiment Videos
Wavelet and wavelet packet compression of electrocardiograms
1Department of Computer Science, University of South Carolina, Columbia 29208, USA. hilton@cs.sc.edu
IEEE Transactions on Bio-Medical Engineering
|May 1, 1997
Summary
Wavelet-based compression effectively reduces electrocardiogram (ECG) data size. Cardiologists found that 8:1 and 16:1 compression ratios preserve clinically useful ECG information.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Medical Informatics
Background:
- Wavelets and wavelet packets are advanced techniques for signal compression.
- Electrocardiogram (ECG) signal analysis is crucial for cardiac diagnostics.
- Efficient compression of ECG data is needed for storage and transmission.
Purpose of the Study:
- To develop and evaluate wavelet and wavelet packet-based compression algorithms for ECG signals.
- To assess the effectiveness of embedded zerotree wavelet (EZW) coding for ECG compression.
- To determine the impact of compression ratios on the clinical utility of ECG data.
Main Methods:
- Development of ECG compression algorithms using wavelet and wavelet packet transforms.
- Implementation of embedded zerotree wavelet (EZW) coding.
- Evaluation of eight different wavelet functions for Holter ECG data compression.
- Blind evaluation of compressed ECG signals by cardiologists.
Main Results:
- Wavelet and wavelet packet-based methods show promise for ECG signal compression.
- Embedded zerotree wavelet (EZW) coding is applied to ECG data.
- Eight wavelets were assessed for their compression performance on Holter ECG data.
- Pilot study indicates 8:1 compression preserves essential clinical information.
Conclusions:
- Wavelet-based compression algorithms, particularly EZW coding, are effective for ECG signals.
- An 8:1 compression ratio generally preserves clinically relevant ECG information.
- In many cases, 16:1 compressed ECGs remain clinically useful, demonstrating high compression efficiency.