Influence of different visualization techniques of 3D free-form surfaces on perceptual performance in a CAVE

Computer-aided design methods are extensively applied in the design of aircraft, automobiles, buildings, computer and many other products. In the next future, CAD methods will be employed in conjunction with 3D immersive displays, which may dramatically improve the possibilities of visualization and interaction offered by common 2D display CAD workstations. Thus, careful evaluation needs to be made as to how the computer-generated object is represented on 3D display. It is commonly recognized that a computer-generated 3D object should be a) an accurate description of the model being designed b) presented in a realistic and integrated format, so that it can be visualized and interpreted without introducting uncertainty regarding to the represented proprieties (Brown, 1995). To ensure that the displayed 3D image will satisfy these requirements, it is important focusing on which perceptual/cognitive operation is performed by the user during the design process. We performed one experiment to examine the effects of different visualization techniques of 3D computer-generated free-form surfaces on depth perception, using Steven's stimulus magnitude estimation paradigm. The factors we investigated were presence versus absence of binocular disparity, four different monocular coding techniques (wireframe, flat shading, Gouraud shading, Gouraud shading and normals) and two levels of shape complexity. We found that a stereoscopic image provides subjects with a visual cue that not only enhances perceived relief of 3D forms but also improves accuracy of activities necessary for 3D design, such as depth estimation of free-form surfaces. Furthermore, the nature of the experimental task implies that stereoscopic displays can be most useful when information is presented from a vertical viewpoint.