Muggle scientists develop Harry Potter 'Marauder's Map' technology
The CMU algorithm allows researchers to follow the movements of 13 people even when they slip out of view of cameras.
However, the scientists’ methods are far from magical — they used cues from the video feed, including apparel color, trajectory and facial recognition, to track individuals even when they wandered into blind spots. Previously, such complex tracking and analysis was impossible for computers and had to be done manually.
Multi-camera, multi-object tracking has been an active field of research for a decade, but it’s mostly focused on controlled environments. The CMU team, by contrast, monitored actual nursing home residents and employees using 15 cameras.
Research was complicated by a variety of factors, including long hallways, blind spots and variations in lighting. For example, something as simple as tracking colored clothing is difficult because the same color can appear different to cameras in different locations due to lighting.
Also, a camera’s view of a person can be blocked by other people, by furniture or when that person enters a blind spot, so individuals must be regularly reidentified.
Face-recognition technology is helpful in reidentification, but faces were recognized in less than 10 percent of the frames, so researchers had to develop mathematical models that used a variety of information, such as appearance and trajectory, in addition to facial features.
The performance of the CMU algorithm improved on two of the leading algorithms in multi-camera, multi-object tracking. It located people within one meter of their position 88 percent of the time, compared with 35 percent and 56 percent for the other algorithms.
Although Harry Potter and his friends could use the Marauder’s Map only by tapping it with a wand and saying, “I solemnly swear I’m up to no good,” the CMU scientists hope their research will be used in positive ways.
"The goal is not to be Big Brother, but to alert the caregivers of subtle changes in activity levels or behaviors that indicate a change of health status," said Alexander Hauptmann, principal systems scientist in CMU’s Computer Science Department.
In addition to monitoring the health of nursing home residents, the research could be used in airports and other places where security is a concern.
More Harry Potter science stories on MNN: