SURFACE CLASSIFICATION VIA BRDF PARAMETERS, USING POPULATION MONTE CARLO FOR MRF PARAMETER ESTIMATION

摘要

A new MRF parameter estimation method is proposed, based on the Population Monte Carlo algorithm. This method is used to estimate MRF parameters which correspond to the Bidirectional Reflectance Distribution Function of a material surface given its reconstructed surface geometry, camera matrices and lighting data. The posterior distributions on the MRF/BRDF parameters are then used as features for material classification.