Adaptive wavelet EMG compression based on local optimization of filter banks

This paper presents an adaptive wavelet technique for compression of surface electromyographic signals. The technique employs an optimization algorithm to adjust the wavelet filter bank in order to minimize the distortion of the compressed signal. Orthogonality of the transform is ensured by using a restriction-free parametrization described elsewhere. A case study involving real-life isotonic and isometric electromyographic signals is presented for illustration. The results show that the proposed approach outperforms the standard non-optimized wavelet technique in terms of the percent residual difference for a given compression factor.