Pornchai Phukpattaranont. Feature evaluation for SNR estimate in EMG contaminated with power line interference. (). King Mongkut's University of Technology North Bangkok. Central Library. : , 2020.
Feature evaluation for SNR estimate in EMG contaminated with power line interference
Abstract:
The electromyography signal (EMG) recorded from a muscle contraction has a
variety of uses and applications. However, it can be contaminated by power line
interference, which may cause a severe reduction in the signal-to-noise ratio (SNR).
We have developed the SNR estimation algorithm of the EMG signal contaminated
with power line interference for monitoring signal quality. The feature calculated from the
EMG signals was input to the neural network (NN). Then, the output from the NN, which is
the SNR, was estimated. Six features were evaluated, i.e., skewness, kurtosis, mean
absolute value, wavelength, zero crossing, and mean frequency. The best average correlation
coefficient at 0.9928 was obtained from kurtosis
King Mongkut's University of Technology North Bangkok. Central Library