Researchers at Carnegie Mellon University have developed a complex algorithm which could be used to help facilities monitor residents. The multi-camera, multi-object tracking system was tested in a nursing home. This system seemed destined to meet the same low percentage as other algorithms, as it has relatively few cameras, which was a problem for the other algorithms.  To achieve comprehensive coverage of residents, it was thought that there had to be an extensive number of cameras, covering every hallway, and every area.  However, this system utilizes the complex algorithm to render full camera coverage unnecessary.

Using the algorithm, this system can track residents to a meter of their actual location with an 88% success rate. The system, which could help facilities monitor residents for changes in activity and could help identify patterns indicating a mental status change, uses facial recognition to pinpoint residents. The system isn’t perfect, as the researchers are looking for ways to increase privacy protection, but it is a step forward in the journey for an indoor monitoring system. The researchers will present their findings June 27 at the Computer Vision and Pattern Recognition Conference in Portland, OR.

See full article at Mcknight’s.