Data repository of Project Coswara

Code for the Cough Against COVID-19 (CAC) Project by Wadhwani AI

Collection of cough sounds for cough classification

Official implementation of Hierarchical Spectrogram Transformers (HST)

Sickness Detection with 80% Val Accuracy. Trained on (25% of available train data~5GB for 'Pfizer Digital Medicine Challenge'). Reached to accuracy of 80% in this Demo/Prototype

Dataset of recordings of induced cough

Another project for classifying Covid and non-covid patients through cough sound. Using CRNN-Attention model with the sound data converted into image data

Flask app that converts cough sound into Mel-Spectrogram and utilizes a CNN model to classify them into Healthy and Covid.

This project was made for classifying Covid and non-covid patients through cough sound. This is a non-invasive, inexpensive, and easy method so we can use that as the first filter

Automated Detection model of COVID-19 via Cough record

CovidDetector is a webapp where a user can join, record their cough and get an estimate on the likelihood of them having COVID-19 using machine learning. The users are also able to contribute and provide training data to help retrain the CNN model.

2641-Videos-Cough-Behavior-Data