Static and dynamic myoelectrical signal compression using embedded zero-tree wavelets.

摘要

Recent progress in the diagnostic use of the myoelectric signal (MES) for neuromuscular diseases, and increasing interests in telemedicine applications, mandate the need for an efficient MES compression technique. Consequently, the objective of this work was to investigate wavelet-based MES compression.; MES data were compressed using the one-dimensional wavelet transform (WT) in conjunction with the embedded zero-tree wavelet (EZW) compression algorithm. Three factors were varied in this investigation to assess their effects on MES compression: the wavelet type, the muscle site, and the contraction type. It is shown in this work that compression performance is dependent on the type of wavelet used in the WT, while the effects of different muscle sites and contraction types on compression performance were inconclusive.; A comparison between EZW encoding and conventional MES compression methods revealed that adaptive differential pulse code modulation (ADPCM) is computationally more efficient than EZW encoding, and produces reconstructed signals with less distortion.