Save lives and
monitor health of
one billion people

Donate your voice to help researchers understand if a person is suffering from COVID-19.
Record your voice

How it works

Donate your voice

After a short survey our app will record your breathing and coughing.

Contribute

The more voices we have the more accurate our algorithm will be.

Your data is secure

Your voice will only be used for scientific research. We won't track you.

Record your voice

The science
behind VoiceMed

According to the latest scientific studies, vocal biomarkers reveal a lot about your health; and can help to detect serious diseases and health risks. The analysis of vocal biomarkers has a great potential to transform diagnosis through their accuracy, speed and cost-effectiveness.

This enabled us to create our COVID-19 pre-diagnostic tool, that’s aimed at rationalizing the use of lab tests and reducing the need for visiting primary care providers in case of flu-like symptoms.

Read more

Awards and prizes

Our Technology

Our artificial-intelligence combines real-world data collected from any phone with clinically validated data to achieve the highest possible level of accuracy in detecting COVID-19.  

We are leading the way to healthcare transformation while abiding by the highest standards in terms of privacy and safety for our patients.

About us and
our milestones

We’re an international, highly skilled team made of data scientists, doctors, software engineers and so much more.

We are a fast-growing start-up currently collaborating with two hospitals and we have achieved high visibility by receiving media coverage and multiple awards. 

We’re looking for more hospitals and other entities willing to collaborate in research studies on vocal biomarkers and clinical trials.

About us

Articles about VoiceMed

FAQ

What is the correlation between voice and Covid-19?

From the various literature that have been published till date on clinical symptoms of COVID19, cough and shortness of breath is among the main clinical characteristics, indicating involvement of respiratory pathophysiology. The pathophysiological changes caused by different respiratory conditions modulate the sound quality, these sound streams associated with breathing difficulty & cough can be used to detect COVID19.

Can the difference be measured? Can it be heard?

Machine learning (ML) algorithms will be able to pick the latent sound features that are typical to sick patients by comparing the features between healthy subjects' voices and sick patients' voices. By using some statistical modeling through the data from healthy and sick people, ML algorithm can measure the difference in sound features from both. Though abnormal sounds can be heard by physicians, it would be difficult for them to catch the typical latent sound features unless complemented by technology.

How does artificial intelligence help in the diagnosis and what changes in the voice of a healthy and a sick person?

Sick patients with respiratory infection have symptoms that are normally absent in healthy individuals. The unique sound signatures (either absence of certain sounds or additive abnormal sounds) from the infected patients and healthy subjects can be analysed by machine learning algorithms and classify them as pathological sound and normal sound. There is a systematic review paper published in Plos One journal (one of reputed journals) by researchers from Imperial College London that elaborately describe these features.

Do you have any medical references backing the concept?

Most of the literature on clinical characteristics of COVID-19 patients have reported the involvement of respiratory pathophysiology (with cough, sore throat and breathing difficulty) as the main hallmark [1] [2] [3] [4] [5] [6] [7]. Acoustic analysis has the potential to expedite detection and diagnosis of voice disorders. Applying artificial intelligence using deep neural networks may provide an alternative to the single or multidimensional parameter approach to acoustic analysis.Cough and sore throat is a common symptom with characteristic sounds and movements. Machine learning can help to identify unique sound features related to characteristics of different conditions. Previous researchers have deployed a machine learning algorithm based cough analysis for different respiratory diseases like pneumonia [8] [9] pertussis [10] and tuberculosis [11]. Moreover, in collaborative efforts from both medical and computer scientists there have been deep dive to disinter better machine-learning algorithms to enhance voice detection capability through features like cough classification [12]. In a recent study [13] Kvapilova et.al developed a smartphone application to collect audio-data and measure cough using a machine learning algorithm that is optimized for clinical research.

Get in touch

If you have questions, suggestions or you’re willing to collaborate, get in touch!

You can also find VoiceMed on social media:

Partners