Fig. 3
From: Long short-term memory (LSTM) recurrent neural network for muscle activity detection
Example of post-processing applied to the standard approach output (\({\widehat{y}}_{TKEO}\)). A Sample realization of simulated sEMG signal (blue line) with the superimposition of the truncated Gaussian function used to modulate the muscle activity (black line). The SNR is set equal to 20 dB, the standard deviation of the truncated Gaussian (σ) is 100 ms, and the multiplicative constant (α) of the time support is 1.5. B The output of the standard approach (\({\widehat{y}}_{TKEO}\)) without any post-processing. C The output of the standard approach after rejecting all the activations shorter than 30 ms (post-processing step)