This demonstration puts together simple robotics with our bottom-up model of visual attention to provide the audience with an intuitive understanding of the saliency model.  A stripped-down version of our saliency-based visual attention code runs in real-time on a portable workstation. Video inputs are digitized from a camera mounted on the robot and are processed at 30 frames/second. Saliency maps are computed, highlighting in real-time which locations in the robot's surroundings would most strongly attract the attention of a typical human or monkey observer. At any given time, the coordinates of the most active locations are then transmitted to the robot head's motor control system which executes a combination of rapid (saccadic) eye movements and slower (smooth pursuit) eye and head movements. Click here to see a movie of some of the monkey's basic movements before augmentation.  

These are some shots of the final construction and the monkey performing at the Vision Science Society Demo Night(2006).  Some video clips below.