Computer vision is a dual to computer graphics.
It starts with an image or animation and deduces
the model representation for the computer. A classical
example here is classifying and sorting objects
based on camera input. Taken together, this dual
relationship looks like Figure 1. Computer graphics
goes down from model to image, and computer
vision goes up from image to model. This is, of
course, a simplification, but it serves as a good basis
to understand recent developments intersecting
the two fields