Estimating Material Parameters Using Light Scattering Model and Polarization

Abstract

When rendering the 3D graphics object in the computer vision and graphics fields, the parameter values, such as the surface roughness, refractive index, and light absorption are used for modeling realistic light scattering rendering behavior of a material. Most of these values were usually obtained from different literature or online sources. In general, a specific measuring tool is needed when acquiring the measurement, but, what if we could use an alternative method of measuring when the tools are not available? The idea is to use the light scattering model to estimate the parameter measurement values of material by doing the inverse rendering of the capture polarization image of the object’s light scattering. In this paper, we investigate whether we can estimate four different measurements, using ARLLS model. We captured the object’s degree of polarization (DOP) to be fitted and compare with the model to investigate the relationship between the two by doing the correlation test between the object measurement DOP and the model parameters to see whether the estimations are in agreement with each other (e.g. by reference or by the practical characteristics of the object). Furthermore, the test will be conducted using materials with varying degree of properties (e.g. very rough surface; highly chromatic).