A technique for precise depth representation in stereoscopic display

In observing a virtual 3D object displayed stereoscopically on a large screen, there is often a difference between the calculated depth and the perceived depth. This paper presents a method for reducing such differences of depth. To do this, we modify both the viewing position and the screen position in the stereoscopic calculation. The optimal amount of modification was determined from sample values of depth differences. The effectiveness of the proposed method is discussed on the merits of the experimental results. This technique decreased the average difference from 4.3 mm to 1.3 mm.