Although the conventional testing capacity has been increasing considerably, most people who are not in need of immediate health care are not yet being tested. As we wait for a vaccine and start returning to the new normality, testing will be essential in deciding who is safe to go, who must stay at home, who needs to see a doctor and how urgently.
Our artificial-intelligence-based medical device software analyzes your voice, breath and cough samples and gives you an immediate result, a COVID-19 pre-diagnosis.
We have an ongoing pilot study in partnership with two hospitals and our goal is to have an MVP (fully functional but not yet certified) by the end of October 2020.
Researchers at Cornell university have shown a proof of concept of COVID19 cough detection through machine learning algorithm with 90% accuracy distinguished from other non-COVID19 coughs. Voice analysis could also be utilized in absence of respiratory symptoms as well. Speech impairment was shown to predispose cognitive deterioration in hepatic encephalopathy. Recent publication in JAMA and NEJM has emphasised on neurological manifestation even in absence of typical clinical symptoms like cough in COVID19 patients. A very granular analysis of voice from these patients may disinter typical voice markers in COVID19 patients that could pave way for breakthrough technologies to detect or complement COVID19 diagnosis conveniently and effectively even in absence of biosafety compliant laboratories.
Our solution could be used to identify high-risk patients with COVID19 by implementing deep learning methods and classifying both COVID-19 positive and negative sound streams. This will be achieved by gathering cough sounds from healthy volunteers and COVID19 patients which will be then trained and validated using machine learning algorithms by our machine learning experts in collaboration with medical experts.