Video Algorithm Makes Invisible Motion Visible
What’s the Latest Development?
Scientists at MIT’s Computer Science and Artificial Intelligence Laboratory have developed a process, Eulerian Video Magnification (EVM), that can be applied to ordinary video to display movements that are all but invisible to the naked eye. Since presenting the process at last summer’s SIGGRAPH conference in Los Angeles, the team has worked on improving its accuracy and clarity, and the resulting code is available online along with a site to which people can upload video clips and see how they look with the process applied.
What’s the Big Idea?
The original purpose of EVM was to allow monitoring of babies in neonatal wards without physically touching them. Now the team envisions EVM being used in functions ranging from search-and-rescue — checking whether a victim is still breathing, for instance — to monitoring the safety of factory equipment. MIT co-author Michael Rubinstein says that people have also asked about EVM’s effectiveness in lie detection, either in a police examining room or at a poker table: “[They] wanted to be able to analyze their opponent…and be able to know whether they’re cheating or not, just by the variation in their heart rate.” The team is working on turning the program into a smartphone app.
Photo Credit: Shutterstock.com
