Artificial intelligence in care and health sector

The Hamburg-based Beyond Emotion, for instance, has come up with AI-assisted emotion recognition systems to make life easier for caregivers. "In Germany, 3.1 million people in need of care live in their own homes. Around 80 per cent are cared for by their relatives," said Hanne Butting, joint founder of the start-up. However, relatives worry about their loved ones constantly as 24/7 care is almost impossible. Beyond Emotion, a spin-off from the Smart Systems Research and Transfer Centre (FTZ) in the Faculty of Technology and Computer Science at the Hamburg University of Applied Sciences, offers a remedy for such worries. "Relatives receive feedback in real time about the state of the person being cared for," Butting explains. The self-learning AI recognises 17 different feelings including fear, terror and boredom based on facial expressions. The results and a digital image are sent to relatives per app. The recipient can react immediately - in person or per video call. "We want to reduce the everyday stress triggered by uncertainty about how the person being cared for is feeling," said Butting.

The DigiRehab start-up is keen to enable the elderly to lead self-determined lives for as long as possible. To this end, the start-up has come up with a customized digital exercise workout. Lars Jessen, Germany Manager of DigiRehab GmbH, said: "We want to strengthen their physical and mental health and increase their quality of life." The concept has been tried and tested in Denmark since 2017 and in Norway, Austria and Germany (Hamburg) for over two years. "Studies show that our exercise scheme has reduced the need for care by 74 per cent," said Jessen. Given the ongoing shortage of skilled nurses and carers, the scheme is gaining importance. DigiRehab's workout includes 2,500 exercises which are tailored to individuals and adapted based on regular AI screenings.

Philip Mertes and Dirk Simon, founders of Evocal Health GmbH, highlight the importance of the human voice as it can predict a person's state of health. "A large part of the body is involved in speaking," said Mertes. More than 100 muscles are used including the diaphragm to inhale and exhale. Vocal folds in the larynx produce the sound and the voice varies depending on the tongue, lips and facial muscles. "The voice becomes a prognostic and diagnostic biomarker." Evocal uses AI-assisted and machine learning to identify recurring patterns that are typical of certain diseases. "We don't analyse semantics, but focus on dynamic speech parameters, so our approach can be applied to very different languages."

ys/kk/pb