Recent developments in computer graphics, and particularly within predictive rendering, enable highly realistic simulations of object appearances. Though physically-based reflectance (PBR) models offer widespread utility, measured material reflectance data yields significantly superior accuracy through the direct empirical observation of complex light-scattering interactions. Nevertheless, acquiring and modeling reflectance data causes substantial computational overhead. This work explores learning-based methods to facilitate the acquisition, representation, and rendering of reflectance data for predictive rendering purposes. We present a compressed sensing framework to optimize gonioreflectometer-based measurements, proposing a novel sampling strategy for surface reflectance acquisition. Furthermore, we employ sparse representation techniques upon the existing reflectance datasets, ensuring representational fidelity while allowing for real-time rendering. This research aims to balance accuracy and efficiency, contributing to the domains of photo-realistic image synthesis and predictive rendering.
Zoom link: https://cam-ac-uk.zoom.us/j/83107754095?pwd=Y2ietFlkaTqqWhlZ4PUC6cSSUkJ2Vl.1
