The telemedicine industry has become an essential way for healthcare specialists to continue providing medical care while minimizing contact during the Covid-19 pandemic. However, telephone or video calls further complicate the diagnosis, mainly to obtain accurate results of vital signs such as respiratory rate or pulse. For that reason, a group of scientists from the University of Washington, developed a method that uses the camera of the phone or a computer vital signs in real time.
The system measures physiological signals, it is also less prone to errors, when it comes to having a suitable condition such as lighting or facial features and even skin color. Likewise, the machine learning tool allows images to be classified, that is, if they need a series of photographs where a particular signal appears, the system itself relates patterns to locate it. Xin Liu, main author of the work, stressed, “Machine learning is useful for telemedicine, as it is a system that can identify the region of interest through an image.”
It should be noted, the first version of the system was trained with a data set that contained so many videos and photographs of faces. In addition, the data records on the pulse and respiration rate of each person measured with conventional medical instruments. In this way, the system was able to use the spatial and temporal information from the videos to calculate vital signs.
After the first test, the scientists realized that the system should be focused on improving the search in important areas of the face within the video. The researchers found that lighting conditions, skin tone, and blood flow could be targeted to specific areas to measure pulse and respiration rate.
The new system outperformed its predecessor thanks to machine learning when given more challenging data sets, especially if the images are not in a well-lit environment. According to Xin Liu, the light that is reflected on the face is essential to have better information when capturing an image or video.
To date, researchers continue to work in a variety of collaborations with physicians to improve the functioning of the clinical system. This will allow doctors to have more tools for telemedicine including personal care, monitoring of chronic diseases or people who do not have access to a nearby clinic.