Deep learning project on signal processing. This research is on human motion monitoring data from wearable sensor. Building a convolutional Autoencoder to decode the input signal. The encoder output can be used as compressed representations to determine the correctness of execution. After choosing the model, The correlation coefficients of features in the trained encoder between test data and all data are shown are highly correlated, showing it can statically apply the model furtherly to testing windows and other patients.
