Virtual reality has much potential and many challenges. We investigate geometric image distortion issues arising from human factors concerns in systems using stereoscopic head-tracked displays. These displays are stationary and attached to a desk, tabletop or wall. The user perceives a true 3D image and can examine the virtual scene from different viewpoints by physically moving around the display. Stereoscopic displays raise concerns beyond those found in simpler monoscopic display systems. To maximize viewing comfort and user interaction, viewing parameters must be automatically and dynamically adjusted. This thesis contributes the following: (1) a framework for understanding, classifying, and comparing software techniques that help the viewer fuse the two stereoscopic 2D images into a single 3D perception; (2) analytic descriptions of the distortions induced by the following fusion control techniques: (a) false eye separation; (b) image shifting; (c) image scaling; (3) a comparison of the geometric properties of the above three techniques with each other and with the other fusion control techniques; (4) a fusion control technique with fewer dynamic, geometric distortion components than prior methods; (5) a technique that balances multiple stereo viewing issues when traveling through an extensive, global terrain virtual environment while maintaining an exo-centric, or orbital, view; (6) geometric guidelines for matching an application's geometric requirements to a set of appropriate fusion control techniques.
展开▼
