Computer vision systems decipher outdoor scenes.

The article offers information on a computer vision system devised by computer scientists at Carnegie Mellon University's Robotics Institute which help computers understand an image by deciphering outdoor scenes. Virtual blocks are used by the computer to reason about the physical constraints of the scene by building a 3-D approximation of the image that makes sense based on volume and mass. According to Abhinav Gupta, a post-doctoral fellow in the Robotics Institute, the method can achieve more than 70 percent accuracy.