Smart Meeting Rooms

Computer vision-based monitoring is used for automated recognition of the activities of participants in a meeting. A head tracker, originally developed for monitoring in a home environment, was evaluated for this smart meeting application using the PETS-ICVS 2003 video data sets. The shape of each person's head was modeled as approximately elliptical whilst internal appearance was modeled using colour histograms. Gradient and color cues were combined to provide measurements for tracking using particle filters. A particle filter based on Iterated Likelihood Weighting (ILW) was used which, in conjunction with the broad likelihood responses obtained, achieved accurate tracking even when the motion model was poor. It was compared to the widely used Sampling Importance Re-sampling (SIR) algorithm. Results are reported for tracking and recognition of the actions of the six meeting participants in the PETS-ICVS data. ILW reliably tracked all participants throughout the meeting scenarios.

Video data from two cameras, one mounted on each of two opposing walls, were used. The sequence shown begins with each of six participants entering and then sitting down. Subsequently, each in turn stands up, walks to the whiteboard, writes something and then returns to his seat twice. Finally, each person in turn exits the room. The following AVI files show each person being tracked. The data sets were made available by the PETS workshop series

Tracking people entering the room:

Camera 1 avi
Camera 2 avi

Tracking people during the meeting:

Camera 1 avi
Camera 2 avi

The following images show recovered trajectories from which actions such as entering, exiting, sitting down, getting up, and going to the whiteboard can be automatically recognized.