A wavelet-packets based algorithm for EEG signal compression.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Medical Informatics

Background:

• Biomedical signal transmission in clinical practice generates large data volumes, particularly electroencephalographic (EEG) signals.
• High storage capacity and channel bandwidth demands necessitate efficient data compression systems for clinical data.

Purpose of the Study:

• To develop an efficient lossy compression algorithm specifically for electroencephalographic (EEG) signals.
• To evaluate the compression rate and signal distortion of the proposed EEG compression method.

Main Methods:

• EEG signal segmentation followed by decomposition using Wavelet Packets (WP).
• Thresholding of WP coefficients, quantization, and run-length coding for compression.
• Reconstruction of the compressed EEG signal through an inverse process.

Main Results:

• The Wavelet Packet (WP) transform demonstrated high robustness in the compression-decompression process.
• Achieved compression rates (CR) typically in the range of 5-8 with reasonably low signal distortion.
• The algorithm exhibited a relatively low computational cost, indicating practical applicability.

Conclusions:

• The developed Wavelet Packet-based lossy compression algorithm is effective for EEG signals.
• The algorithm balances compression efficiency with acceptable signal fidelity for clinical use.
• Its low computational cost makes it a viable solution for real-time or resource-constrained applications.