The perception of 3D is done by exaggerating the size of the foreground and depreciating the size of the background. If you look at the classic windows 98 3D maze screensaver, it shows a first person view of navigating a maze. The illusion of the camera moving through the maze is done by "stretching" the walls, floor, and ceiling towards and past the camera. Slowly exaggerating the height of furthest section of the walls and moving them away from the center until they equal the height of the screen.
The advantage of 3D over 2D graphics is the automated illusion of depth that is present. 2D can achieve an illusion of depth, but it's existence and magnitude has to be implemented entirely by the artist, whereas the computer itself can take out the guesswork in 3D. However, now the artist has to input the coordinates for a model in 3D instead of flattening their visualization for a 2D environment.
