Electromyographic signal compression based on preprocessing techniques

Recently, electromyographic records have been rearranged into two-dimensional arrays and encoded with image compressors, in the same way as image data. However, as a consequence of this reshaping, the correlation among signal segments is generally lost, which reduces the compression efficiency. In the present work, new preprocessing techniques for encoding electromyographic signals as two-dimensional matrices are presented, namely percentage difference sorting and relative complexity sorting, which have the potential to favor the exploitation of the intersegment dependencies. The experiments were carried out with real isometric records acquired in laboratory, that were first preprocessed and then compressed with a JPEG2000 encoder, showing that the proposed framework is effective and outperforms even state-of-the-art schemes present in the literature, in terms of PRD × Compression Ratio.