Related Experiment Videos
Myoelectric signal compression using zero-trees of wavelet coefficients.
Jason A Norris1, Kevin B Englehart, Dennis F Lovely
1Institute of Biomedical Engineering, University of New Brunswick, 25 Dineen Drive, NB, E3B 5A3, Fredericton, Canada.
Medical Engineering & Physics
|October 2, 2003
Summary
The embedded zero-tree wavelet (EZW) algorithm effectively compresses myoelectric signals for neuromuscular disease diagnosis in telemedicine. Compression improves with segment length and specific wavelet types, offering online compression factor control.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Medical Informatics
Background:
- Advancements in myoelectric signal diagnostics for neuromuscular diseases necessitate efficient data compression.
- Growing telemedicine applications require robust techniques for transmitting physiological data.
Purpose of the Study:
- To evaluate the efficacy of the embedded zero-tree wavelet (EZW) compression algorithm for myoelectric signals.
- To analyze the impact of analysis parameters and measurement conditions on EZW performance.
- To compare EZW with other lossy compression techniques.
Main Methods:
- The embedded zero-tree wavelet (EZW) algorithm was applied to myoelectric signals.
- Analysis parameters included segment length and wavelet type (Meyer, biorthogonal).
- Measurement conditions involved different muscle types and contraction types.
Main Results:
- Compression performance of EZW improved with increasing analysis segment length.
- Meyer and fifth-order biorthogonal wavelets demonstrated good compression performance.
- The impact of muscle and contraction types on compression was inconclusive.
- EZW outperformed hard thresholding wavelet compression but was less effective than adaptive differential pulse code modulation.
Conclusions:
- EZW is a promising compression technique for myoelectric signals in telemedicine applications.
- Its ability to specify compression factor online makes it suitable for streaming data.
- Further research may be needed to clarify the influence of muscle and contraction types.