vital sign ai

Vital Sign AI’s app detects vital signs via smartphones and AI to aid the fight against this global pandemic, and easing pressure on our massively strained medical professionals.

Prognosis, monitoring and screening for the patients infected with COVID-19 is suggested to be based on vital signs such as respiration and blood oxygen saturation. Machine Learning (ML) methods have already shown promise and convincing results in the detection of vital signs.

PFA prototype link:

Our application is helpful especially for people living remotely, elderly, and vulnerable and immuno-compromised people as well as people in under-resourced communities.

Breathing.AI detection of heart rate using smartphone and laptop cameras and AI

The project was started March 22, 2020, via the website, and currently has 130 volunteering healthcare practitioners (HCP), data scientists, engineers and researchers from Stanford, Cornell Tech, and from various parts of the globe with institutional support. 

Please contact us (email: team@ our website’s name) if you want to help.

Crucially, the app will be released for free across the globe, and the data collected for this purpose will also be released for research purposes, thereby promoting further work to fight against pandemics. The app provides FDA-endorsed remote monitoring. It will help to reduce the intake of patients with mild symptoms for the doctors and hospitals, and limit the direct interaction of people. 

The app facilitates non-invasive and remote detection of vital signs such as heart rate variability, breathing patterns, and hemoglobin. This process relies on smart-phones and laptops to capture audio and video data, use state-of-the-art Machine Learning and Deep Learning algorithms to draw inferences, and subsequently make predictions. 

Prognosis, monitoring and screening for the patients infected with COVID-19 is suggested to be based on breathing characteristics (Wang et al. 2020.) as well as other vital signs such as blood oxygen saturation/hemoglobin levels.

We create an open-source database of vital signs, which would potentially be invaluable to the medical and data-science communities. 

The first weeks of COVID-19 have shown a serious lack of testing abilities. The detection of COVID-19, as well as some of its major symptoms, requires visits to a medical facility or a designated test site. This results in possible transmission or infection of COVID-19, and puts a massive strain on available medical resources. Indeed, there is empirical evidence regarding heightened risk of transmission in close proximity to treatment and testing facilities.

This is particularly true for the elderly and medically immuno-compromised people, as well as for the large population groups in under-resourced communities. This stems from relatively poor health conditions of the comorbid population (which often includes the elderly), as well as limited access to healthcare providers and  wearables to track health data. 

Naturally, this makes it an imperative to investigate methods to remotely detect COVID-19. However, lack of labeled COVID-19 data and medical expertise hinder the development of such algorithmic approaches at this time.

Also, we firmly believe that it would be dangerous to try to bypass medical professionals and remotely detect COVID-19 at such an early stage of the disease.

In fact, while there are already claims regarding remote detection of COVID-19 using smartphone sensors and additional clinical data, it is unsafe to claim results on diagnosis without clinical trials, and such approaches might turn out to be detrimental in this fight.  

Our app, Vital Sign AI, provides vital sign detection via data collected from smartphones and laptops, and includes access to healthcare practitioners and therapeutic methods.

While detection of necessary vital signs is an active area of research within machine learning, most approaches treat different signs as standalone problems. Multiple vital signs have not been detected in an app in combination with ML, with added access to HCP and potential therapeutic benefits. The wide breadth and specialized skills of our interdisciplinary team make the implementation of the app well within our reach. 

The app can also be used to practice suggested therapeutic exercises proposed for health at home, and ongoing mental and physical health monitoring. The vital sign detection shows progress and offers a motivating way to practice aerobic and mindfulness exercises as therapeutic methods.

A well-documented pipeline involving research and development provides long-term value to the medical, scientific and technical fields to detect other diseases, and provide therapeutic support at home.