Our prototype is already able to generate state of the art recordings, but you guessed it, the bottle neck is how to make sense of all the data. We have tested several methods, from recurrent neural networks on the raw data to simple classifiers on processed data (sliding RMS of the signals, or sliding covariance matrix of the different channels) etc...
Do you have any idea on how to interpret sEMG signals into useful computer commands ? Contact us to get your hands on a prototype and implement your ideas.
Made an account just to say this is really cool!
I'm a data scientist but would love to make one, can we get the details?