SensiML COVID-19 Cough Screening Dataset
Data Collection Page
NOTE: This site is tested to work best using the Chrome web browser on Windows 10 PCs.
Given the worldwide pandemic, SensiML is offering its edge AI solution to the development of a rapid new pre-diagnostic screening technology. This technology, building on recent academic research from University of Oklahoma, Michigan State University, and others*, analyzes recorded cough noises to predict positive COVID cases with over 90% accuracy. While not a replacement for clinical testing, it can play a key role in improving back-to-work employee wellness checks and public facility screening to help protect the health of everyone.
If you are a U.S. resident of legal age, you can contribute to this project by sharing your anonymous, non-personally identifiable cough sound samples.
SensiML will aggregate fully anonymized results as an open source public dataset and for use in our own development efforts.
We need a broad sampling including healthy people, individuals with other respiratory conditions, and those having or suspecting they have COVID-19.
The process is intended to be simple and take less than 3 minutes:
- Create an audio recording of your coughing sounds from your device’s microphone (only .WAV files allowed)
- Answer all non-identifiable health questions you are comfortable sharing
- Attach your audio file using the “Choose File” button
- Read the consent form, confirm via the checkbox, and press “Submit Data”
All answers are used solely for aggregated AI analysis of the sound files. We thank you for contributing to this effort!
Click here for additional FAQs
Record a sound file to be saved locally to your browser download directory.
You will select this file at the end of the page using the “Choose File” button to submit it.
Trouble recording audio?
H5P Audio Recorder is MIT licenced , Copyright (c) 2020 Joubel
* Reference: "AI4COVID-19: AI Enabled Preliminary Diagnosis for COVID-19 from Cough Samples via an App"; IEEE Access, Apr 02, 2020;
Authors: Ali Imran, Iryna Posokhova, Haneya N. Qureshi, Usama Masood, Sajid Riaz, Kamran Ali, Charles N. John, Muhammad Nabeel