Discrimination between Myopathy and normal EMG signals using intrinsic mode functions

In this paper, an efficient technique for the analysis of Myopathy and normal electromyography (EMG) signals is proposed. Empirical mode decomposition (EMD) technique has been used to decompose non-stationary EMG signals in some set of narrow band intrinsic mode functions (IMFs). Hilbert transform of IMFs provides analytic signals, which are used for extraction of instantaneous frequency and instantaneous amplitude. Area feature has been estimated by the trace of analytic IMFs in the complex plane. Instantaneous frequency determined from the analytic signal is involved for the computation of mean of first derivative of instantaneous frequency feature (MFD_IF). These two features are used for discrimination between the Myopathy and normal EMG signals. It is shown in results that both the features are providing valuable discrimination performance.