A fiber-level model for predictive cloth rendering

Abstract

Rendering realistic fabrics is an active research area with many applications in computer graphics and other fields like textile design. Reproducing the appearance of cloth remains challenging due to the micro-structures found in textiles, and the complex light scattering patterns exhibited at such scales. Recent approaches have reached very realistic results, either by directly modeling the arrangement of the fibers [Schröder et al. 2011], or capturing the structure of small pieces of cloth using Computed Tomography scanners (CT) [Zhao et al. 2011]. However, there is still a need for predictive modeling of cloth appearance; existing methods either rely on manually-set parameter values, or use photographs of real pieces of cloth to guide appearance matching algorithms, often assuming certain simplifications such as considering circular or elliptical cross sections, or assuming an homogeneous volume density, that lead to very different appearances.