Doscendo Discimus A trek from circuits to code

Deep Neural Networks In Cough Detection

DNN - Application Notes


Automatic Differentiation


This page is to give a cursory overview of automatic differentiation. Most of the content on here was lifted from the following resources:


A Coarse Cough Detection


- Statistical Tests: the idea here is to capture the salient feature of the cough waveform by looking at the ratios of novel peaks and dips in the signal. We could have fixed divisions which correspond to distinct portions of the cough waveform. Then a hypothesis test can be performed on these ratios (maybe F-test?) to obtain some p-value (probability of the data being cough).