Profiles of Leading Women Scientists on AcademiaNet.
Search among the members of the Leopoldina for experts in specific fields or research topics.
Bild: Wolfram Scheible
Year of election: | 2021 |
Section: | Informatics |
City: | Tübingen |
Country: | Germany |
Research Priorities: Computer vision, human motion, virtual avatars, computer graphics, machine learning
Michael Black is an American computer scientist specialising in motion estimation from videos, with a particular focus on 3D human motion. He combines computer vision, computer graphics and machine learning to teach machines to see and understand humans and their behaviour. His methods are widely used in industry as well as in social sciences, psychology and healthcare.
Humans use their bodies, facial expressions and gestures to help shape the world. In order for computers to act as fully fledged partners to humans, they must be able to see and understand people’s behaviour and recognise their facial expressions, motions and actions.
Although computers can already “see”, their powers of perception are not as strong or flexible as human perception. Certain things that humans can easily see and understand still pose a challenge to machines. For example, if computers are to fully comprehend humans and their behaviour, they need to be able to recognise situations when a person wishes to pick up something heavy and needs help to do so. They must also be able to perceive when a person is distracted or when changes in their behaviour may suggest medical problems or cognitive decline.
To overcome such challenges, Michael Black’s team uses machine learning to train computers so that they can reproduce human behaviour, motions and facial expressions in great detail and in a 3D world. Black develops robust algorithms and 3D models of the human body on a level never achieved before. He takes individual images and videos and uses them to recreate three-dimensional human figures, their movements and the objects around them. These technologies have numerous potential uses, especially in the area of human health. For instance, Black has found a way of capturing the 3D movements of infants so that cerebral palsy can be automatically diagnosed at an early stage when intervention may still be beneficial. He also uses statistical 3D body shape models and virtual reality to conduct research into the causes of eating disorders.