CNN-based BRDF parameter estimation

The proposed network architecture and loss design.

Abstract

The behavior of surfaces is an essential field in computer games and movies. An exact representation of a real-world surface allows for a higher degree of realism. Capturing or artistically creating these materials is a time-consuming process. Thus, in this thesis a method which utilizes an encoder-decoder Convolutional Neural Networks (CNN) to extract information of the Bidirectional Reflectance Distribution Function (BRDF) automatically is proposed. Opposed to previous means this method retrieves information of the whole surface as spatially varying BRDF-parameters with a sufficiently high resolution for real-world usage. The capture process for materials only requires five known light positions with a fixed camera position and thus can be acquired even in a mobile setup. This reduces the scanning time drastically and a material sample can be obtained in seconds with an automated system.

Type
Avatar
Mark Boss
PhD. Student

My research interests lie at the intersection of machine learning and computer graphics.

Related