Integrated three-dimensional reconstruction using reflectance fields
This addresses the challenge of accurate 3D reconstruction for objects with complex material properties, though it appears incremental as it builds on reflectance field methods.
The paper tackles the problem of 3D reconstruction for objects with non-Lambertian surfaces by integrating material properties captured as reflectance fields, achieving an error of less than 0.3% for objects with anisotropic BRDF.
A method to obtain three-dimensional data of real-world objects by integrating their material properties is presented. The material properties are defined by capturing the Reflectance Fields of the real-world objects. It is shown, unlike conventional reconstruction methods, the method is able to use the reflectance information to recover surface depth for objects having a non-Lambertian surface reflectance. It is, for recovering 3D data of objects exhibiting an anisotropic BRDF with an error less than 0.3%.