Open patientzero opened 5 years ago
Neither classification or regression are going to be performed on the phone.
The speech features suitable to compute on the phone are: jitter, voice rate and standard deviation of f0.
Word error rate and phonological features are going to be computed on the server.
This means checking if features can be properly extracted on Android and processed.