Cough is an important symptom of numerous respiratory diseases, including COVID-19. While different cough phases (i.e., inhalation, compression, and expulsion) have been shown to be related to different pathological origins, existing cough-based COVID-19 detection systems rely on the entire cough recording, thus such phase-related characteristics are overlooked. In this study, our aim is two-fold. First, we have annotated over 1,250 cough recordings from two publicly-available cough sound databases, thus providing the research community with fine-grained cough phase labels. Next, we extract a number of temporal and acoustic features from each cough phase and test their usefulness and complementarity for COVID-19 detection. Experiments show the importance of cough phase segmentation, not only for improved COVID-19 detection, but also for the development of models that are interpretable and can better generalize across datasets.
This repo contains:
- Cough phase annotation files of ComParE and DiCOVA2 cough recordings.
- Scripts to reproduce results shown in our paper "On the importance of different cough phases for COVID-19 detection", which can be found here: https://www.techrxiv.org/articles/preprint/On_the_importance_of_different_cough_phases_for_COVID-19_detection/21382176/1
You can find a detailed instruction in Instructions.md. We have included an example showing you how to import files, visualize them with PRAAT, and segment recordings into phases.
For data privacy purposes, we cannot share the cough sound data in this repository. But the sound files as well as the corresponding COVID-19 labels can be obtained upon approval from the organizers of ComParE and DiCOVA2 challenges. The contact info as well as details about these challenges can be found in the following papers:
A summary of the ComParE challenge: https://arxiv.org/abs/2202.08981
The parental dataset of ComParE (Cambridge sound database): https://openreview.net/forum?id=9KArJb4r5ZQ
DiCOVA2 challenge: https://arxiv.org/abs/2110.01177
If you are only interested in reproducing the results shown in this paper, we encourage you to simply run the covid_prediction.py from inside of the scripts folder. Try different features stored in the feature folder as well as their combinations, you should be able to get the same results.
For generating the acoustic features, check acoustic_feature_extraction.py.
For generating the temporal features (and visualize histograms, etc.), check temporal_feature_extraction.py.
For more details, you can refer to our paper https://www.techrxiv.org/articles/preprint/On_the_importance_of_different_cough_phases_for_COVID-19_detection/21382176/1 . If you are using the annotation files or the scripts in this repo, please make a reference to this paper.
We would like to thank the organizers of DiCOVA2 and ComParE challenges for collecting and sharing the sound data. The aforementioned organizations do not bear any responsibility for the analysis and results presented in our paper. All results and interpretation only represent the view of the authors. We also thank our annotators Zack and Hong for spending their time on the annotation!
This repo is created by Yi Zhu in Oct,2022. If you have any questions, feel free to contact at [email protected].