Stereoscopic pointing in desktop applications

Interactive stereoscopic applications require the addressing of objects in 3D space, i.e., stereoscopic pointing. Therefore it is necessary to feed the pointing system with data of all co-ordinates in a user-friendly way. Hence it is rational to rely on a well-known input device, i.e., the computer mouse. Co-ordinates of the 2D mouse have to be transformed into the 3D pointer co-ordinates by different means: knowledge of a 3D description of the scene can be employed as well as a multiple use of mouse co-ordinates. And, on the other hand, it has to be checked to what extent restrictions in use are tolerable. We focused on stereoscopic pointing in telecooperation and developed three different pointing systems, which were evaluated in a laboratory experiment. Results show that the pointing system which does not necessarily rely on 3D data of the scene and, therefore, is the easiest one to implement, is sufficient for pointing on static objects. Both other pointing systems enable the user to point easily at both, static and dynamic objects.