Publication
Overview Cough audio signal classification has been successfully used to diagnose a variety of respiratory conditions, and there has been significant interest in leveraging Machine Learning (ML) to provide widespread COVID-19 screening. The COUGHVID dataset provides over 30,000 crowdsourced cough recordings representing a wide range of subject ages, genders, geographic locations, and COVID-19 statuses. Furthermore, experienced pulmonologists labeled more than 2,000 recordings to diagnose medical abnormalities present in the coughs, thereby contributing one of the largest expert-labeled cough datasets in existence that can be used for a plethora of cough audio classification tasks. As a result, the COUGHVID dataset contributes a wealth of cough recordings for training ML models to address the world’s most urgent health crises. Private Set and Testing Protocol Researchers interested in testing their models on the private test dataset should contact us at [email protected], briefly explaining the type of validation they wish to make, and their obtained results obtained through cross-validation with the public data. Then, access to the unlabeled recordings will be provided, and the researchers should send the predictions of their models on these recordings. Finally, the performance metrics of the predictions will be sent to the researchers. The private testing data is not included in any file within our Zenodo record, and it can only be accessed by contacting the COUGHVID team at the aforementioned e-mail address. New Semi-Supervised Labeling The third version of the COUGHVID dataset contains thousands of additional recordings obtained through October 2021. Additionally, the recordings containing coughs were re-labeled according to a semi-supervised learning algorithm that combined the user labels with those of the expert physicians, which were modeled using ML and expanded on the previously unlabeled data. These labels can be found in the "status_SSL" column of the "metadata_compiled.csv" file. For more information about the data collection, pre-processing, validation, and data structure, please refer to the following publication: https://www.nature.com/articles/s41597-021-00937-4 The cough pre-processing and feature extraction code is available from the following c4science repository: https://c4science.ch/diffusion/10770/