A Practical Framework for Measuring and Modeling the Appearance of Strongly Anisotropic Materials
We present a practical sparse measurement technique and a novel parameter fitting approach for the appearance of strongly anisotropic materials, with application to finished wood. Our approach makes use of bilateral symmetry arguments to reduce the amount of input data required to capture a spatially varying BRDF. This significantly decreases the necessary acquisition and computation time to recover the model parameters, with an observed speedup close to an order of magnitude over previous work, while achieving significantly improved results. We validate the quality of the rendered results from the new approach using additional dense ground truth measurements obtained using a 4-DoF spherical gantry. We also demonstrate a field measurement system using a portable hoop with individually addressable LEDs. The device is inexpensive, simple to build, fast in operation, and fully compatible with the proposed acquisition technique.
Computer science; Appearance Model; BRDF; Computer Graphics; Material Acquisition; SV-BRDF
Marschner, Stephen Robert
Bala, Kavita; Bindel, David S.
MS of Computer Science
Master of Science
dissertation or thesis