Shape and spatially-varying BRDFs from photometric stereo

This paper describes a photometric stereo method designed for surfaces with spatially-varying BRDFs, including surfaces with both varying diffuse and specular properties. Our method builds on the observation that most objects are composed of a small number of fundamental materials. This approach recovers not only the shape but also material BRDFs and weight maps, yielding compelling results for a wide variety of objects. We also show examples of interactive lighting and editing operations made possible by our method.

1 Introduction

Reconstructing scenes with realistic materials from images is a challenging, open research problem. In many real-world scenarios, both the shape and the materials are unknown and need to be recovered. Although the problem is very difficult in this general setting, good solutions exist when either the materials or shape are known. In particular, if the shape is known, several BRDF estimation methods can be applied from the graphics literature [12], [14], [15], [18], [27] to estimate material properties. On the other hand, when the BRDF is arbitrary but known or can be measured using reference objects, example-based photometric stereo methods [7], [19] enable reconstructing shape models. Figure 1 From a) photographs of an object taken under varying illumination (one of ten photographs is shown here), we reconstruct b) its normals and materials, represented as c) a material weight map controlling a mixture of d, e) fundamental materials. Using this representation we can f) re-render the object under novel lighting.