| Recent advances in computer graphics hardware allow us to generate interactive environments that are more complicated than previously possible. The potential for rich and complex sensory simulation is enormous, but with this great potential, there are a number of major challenges. First, neither total physical realism nor photo-realism is always possible. Even if physical realism or photo-realism is possible, much of the computation may be devoted to image characteristics that are not perceived by the observer. We should ask ourselves, “How can we cheat and get away with it?” Second, physically or photo-realistic rendered images may not be optimized to accomplish a given functional goal. We also need to know how information should be presented effectively for various tasks, i.e. how to achieve “functional realism”. To this end, we present the following work.; We performed an experiment that investigates how prior knowledge of object geometry is combined with shape information from binocular disparity when making reaches to virtual objects.; We explored image information that is used by the visual system to determine the material of an object—specifically, how the visual system uses shape, motion, and a combination of shape and motion to solve the ambiguous problem of distinguishing between specular reflections and diffuse reflectance variations.; To effectively study material perception, we created a number of useful computer graphics tools for rendering real-time, interactive graphics of objects with realistic material properties, under complex, real-world illuminations. We describe the theory and algorithms used, and include specific details about the implementations, and instructions for using those tools. |
