Smartcough

Continuous intelligent cough detection and identification based on smartphones.

Background

This project aims to develop new technologies to enable real-time automatic detection, severity assessment, and characterisation of different types of cough events so a better understanding of their underlying mechanisms can be achieved.

In order to provide a low cost and easy to use telemedicine solution, we will focus on the exploitation of the functionalities of daily use devices such as current smartphones or recently released smart watches.

Special emphasis will be paid to the user experience so there is an apparent reduction of the medicalisation of the patient’s life. That is, we conceive the smartphone or watch working as a cough detector device with minor impact on daily activity, e.g. carried in the pocket or handbag with just the specific app running on it.

Method

To this end, we will investigate robust and reliable signal processing methods allowing effective cough identification on different daily situations with no need to manipulate the device. 

The use of seamless affordable add-ons will also be explored to achieve full diagnostic capabilities at high specificity and sensitivity rates.

Finally, the development of new solutions featuring advanced power management for intense computing using smartphones, as well as high quality sound processing with active noise cancellation will also be investigated.

These solutions will pave the way for new generations of smartphones ready to incorporate state-of-the-art connected health applications.

Progress

Work on this project is ongoing. Several technical obstacles have been overcome and results have been presented at international conferences.

Robust Detection of Audio-Cough Events using local Hu moments (Journal of Biomedical and Health Informatics, 2018)

Efficient k-NN Implementation for Real-Time Detection of Cough Events in Smartphones (Journal of Biomedical and Health Informatics, 2017)