University of Tokyo develops a contact-free way to monitor blood pressure, and blood glucose levels using only an optical camera. Thus they have prototyped an optical method to detect hypertension and diabetes. The scientists used a multispectral camera. They pointed it at the faces of the trial participants. It detected variations in their blood flow. The researchers further analyzed the data by AI software.
Multispectral cameras in blood pressure measuring
Multispectral cameras can see into the infrared and ultraviolet range. That is why they can monitor minute changes in the pulses of blood flow across the body. Human eye cannot do it. The sensitivity of the multispectral cameras is much higher than that of the human eye.
The detection of such health conditions as diabetes and high blood pressure is difficult. The researchers applied AI software to process out great amount of thermal noise. Also, the researchers achieved removing factors that make the processing of image data difficult under traditional means.
They trained the AI by comparing images against actual blood measurements taken at the same time. Then the AI assessed and predicted most likely development in the blood pressure and glucose levels with a good deal of accuracy.
Blood pressure measuring by AI: 94.2% accuracy
The diabetes is important to detect. As it is a modern disease that affects a growing number worldwide. It is often silent and undetected for many years until severe health damage has already been don. Hence, early detection is the key to successfully managing the disease. Blood glucose and A1C levels are critical factors, and the AI was able to detect diabetes correctly 75.3% of the time. The AI was able to detect hypertension correctly 94.2% of the time.
Another product that employs a similar technology to assess heart health is already available in AppStore. It is the Heartery app, that assesses blood flow under the skin of your face by running a face scan. Heartery employs photoplethysmography (PPG). Read our interview with the man behind this tech.
