A novel representation method for electromyogram (EMG) signal with predefined signature and envelope functional bank

In this paper, a new method to model EMG signals by means of "Predefined Signature and Envelope Functional Banks (PSEB)" is presented. Since EMG signals present quasi-stationary behavior, any EMG signal X/sub i/ is modeled by the form of X/sub i//spl ap/C/sub i//spl alpha//sub K//spl phi//sub R/ on frame bases in this work. In this model, /spl phi//sub R/ is defined as the Predefined Signature Vector (PSV); /spl alpha//sub K/ is referred to as the Predefined Envelope Vector (PEV) and C/sub i/ is called the Frame-Scaling Coefficient (FSC). The EMG signal for each frame is described in terms of the two indices "R" and "K" of PSEB and the frame-scaling coefficient C/sub i/. Furthermore, it has been shown that the new method of modeling provides significant data compression while preserving the clinical information in the reconstructed signal.