Michael Greenspan of Queen's Univerisity has developed a pool-playing robot called Deep Green. According to Greenspan, pool experts predicted it was unlikely a robot could play pool as well as a human, while fellow robot researchers underestimated the complexity of doing so. But after five years of work, Deep Green is equipped with a vision system that can identify striped and solid balls, track sphere dipoles (PDF format), and achieve other feats of pool-related image processing. Greenspan's vision research resulted in at least one image recognition patent. The robot also has an understanding of physics that allows it rival any human player at making a shot. Even expert human players miss shots frequently while Deep Green has an almost perfect record. Current work is being done at the strategy level and it is expected that Deep Green will be able to easily beat any human at the game of pool.