Selective BRDFs for High Fidelity Rendering

Abstract

High fidelity rendering systems rely on accurate material representations to produce a realistic visual appearance. However, these accurate models can be slow to evaluate. This work presents an approach for approximating these high accuracy reflectance models with faster, less complicated functions in regions of an image which possess low visual importance. A subjective rating experiment was conducted in which thirty participants were asked to assess the similarity of scenes rendered with low quality reflectance models, a high quality data-driven model and saliency based hybrids of those images. In two out of the three scenes that were evaluated significant differences were not found between the hybrid and reference images. This implies that in less visually salient regions of an image computational gains can be achieved by approximating computationally expensive materials with simpler analytic models.