The technology had already been mentioned in the midst of the COVID-19 epidemic, but it seems that the American startup Hyfe is ready to take the next step. The concept is simple: every disease produces symptoms, and when it comes to a condition that affects the airways, the sound of the cough differs slightly depending on what has been caught. For a human, it would be very complex to distinguish between these different sounds.
But it is possible to train artificial intelligence specifically for this. An example of what Hyfe manager Peter Small calls “Epidemiological acoustics”. One of the main vocations put forward by Hyfe is tuberculosis screening “With a very low cost per patient”adds Grant Theron, a professor at Stellenbosch University (South Africa). In the event of a positive result on the application, patients could be referred “Towards a laboratory test” to confirm or deny the diagnosis.
This application published by Hyfe will soon make it possible to establish diagnoses by listening to your cough
This dematerialization of screening could prove to be an essential asset in the fight in developing countries of diseases virtually eradicated in developed countries. There is also COVID-19, of course – but this app also paves the way for more frequent diagnoses, allowing the detection of serious conditions and therefore advancing healthy life expectancy. But for now Hyfe continues to improve its app – the startup is feeding its AI with a massive amount of cough recordings for it.
At this stage, an encouraging clinical trial has already been conducted on 800 people with nocturnal cough. Other steps are still necessary before launching this application – which is nevertheless only a matter of time. This kind of approach has already been explored by competitors. For example ResApp already exists and makes it possible to quickly establish a diagnosis from cough recordings. It remains to be seen whether this kind of screening will one day be declined by variants using the other sensors of the smartphone.
Read also – This artificial intelligence deciphers the sound of the cough to predict epidemics
After all, they are equipped with many advanced sensors besides their microphones, such as cameras, pulse sensors and even on some models ultra-precise LiDAR sensors. It remains to be seen whether an AI can ultimately use this data to deliver such useful health advice.