We developed an AI-based cough sound analysis technology to help identify at-risk Covid patients before administering lab-based tests, even if they’re asymptomatic.
To the best of our knowledge, this is currently the largest Covid cough dataset with verified ground truth labels from RT-PCR test results. At the current performance and at an assumed disease prevalence of 5%, our model can potentially increase testing capacity by 43%.
Given recent developments in the fight against Covid, we have open sourced our codebase, trained models and datasets. We believe that open innovation is the fastest and most reliable way to take our work to the next level and deliver screening and triaging solutions that can benefit millions of people in a short time frame.
We aim to make our work available to researchers and healthtech developers who can use the resources to build solutions that meet clinical and regulatory requirements, and can get these solutions to the health system faster than we are able to. We also believe that our work can be beneficial to those developing screening and triaging solutions for other respiratory illnesses. We also believe that our work can be beneficial to those developing screening and triaging solutions for other respiratory illnesses. The various pathogens that cause respiratory diseases affect the lungs and airways in different ways. The cough sound contains vital information about the state of a patient’s respiratory tract. Doctors use stethoscopes to listen to the lungs when they suspect a respiratory problem. Experienced doctors can even recognise the acoustics of the cough and use it to inform their assessment of the patient.