Performance evaluation of wavelet-based ECG compression algorithms for telecardiology application over CDMA network.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wireless Communications

Background:

• Wireless transmission of electrocardiogram (ECG) signals is crucial for remote patient monitoring.
• Scalable and reliable ECG compression algorithms are needed for tele-cardiology applications, especially over wireless networks like CDMA.
• Evaluating existing wavelet-based algorithms is essential to determine optimal performance in diverse network conditions.

Purpose of the Study:

• To compare the performance of three wavelet-based ECG compression algorithms: Rajoub (RA), Embedded Zerotree Wavelet (EZ), and Wavelet Transform Higher-Order Statistics Coding (WH).
• To identify the most suitable ECG compression algorithm for wireless tele-cardiology applications, particularly over CDMA networks.
• To analyze algorithm performance under various signal conditions (normal, abnormal, noisy) and channel models (noise-free, random noise, CDMA).

Main Methods:

• Performance analysis of RA, EZ, and WH algorithms using ECG signals from the MIT-BIH database.
• Evaluation of compression efficiency, reconstruction error sensitivity, and processing delay.
• Simulation across noise-free, random noise, and CDMA channel models, including testing on an actual CDMA network.

Main Results:

• The EZ algorithm demonstrated superior compression efficiency in low-noise environments.
• The WH algorithm showed competitive performance in high-error environments, even with degraded signal quality.
• Algorithm performance varied significantly based on signal type (normal, abnormal, noisy) and channel conditions.

Conclusions:

• The choice of ECG compression algorithm for wireless tele-cardiology depends on the expected network environment and signal quality.
• EZ is recommended for reliable, low-noise wireless ECG transmission.
• WH offers a viable alternative for applications where network errors or signal degradation are anticipated.