Pennsylvania: People are often unaware of the dangers of stroke and it is too late by the time they get to the hospital. In this context, the American expert has created a machine learning app that can assess the condition from the voice and video of the victim.
Read More:A common drug reduces the risk of death in patients with code 19, research says
Even after arrival at the hospital, it is important to see a neurologist to assess the severity and severity of the stroke. The app was jointly developed by Pennsylvania State University and Houston Methodist Hospital. To prepare for this, the voices and videos of the first 80 or so stroke patients have been recorded and added to a database. Stroke patients have difficulty speaking while the facial muscles become stiff and they do not remain normal.
Read More:Green tea and coffee are useful drinks for diabetics
The algorithm was then trained by passing the data through machine learning. When it was tested on stroke patients, it predicted the severity of stroke with 79% accuracy, which was also confirmed by various tests. The app is claimed to work similarly to emergency room diagnostics. Most importantly, the app can detect paralysis in just four minutes.
Professor Sharon Huang, who is doing research on it, says that this is the first work that can diagnose stroke in an emergency. Facial data and voice are very helpful in this regard and in this way patients sitting at home can assess their condition even from a distance. In that sense, the app is medically important.