Robust and Accurate Non-Parametric Estimation of Reflectance using Basis Decomposition and Correction Functions Supplementary Material

In the first part of this additional evaluation we show that our algorithm achieves a significantly lower perceptual error (error of 0.19) for a set of 16 newly measured materials when compared to other state-of-the-art basis decomposition methods (LC, error of 0.50; RBF, error of 0.38). Additionally, we show that the BRDF estimates computed by RBF and the scattered data interpolation technique PP are often physically implausible, in contrast to the results provided by our method. In the second part of this additional evaluation we analyze how the size of the generated basis affects the fitting accuracy of the final solution. 1 Experiments with Measured Data