Surface Science for Vision and Graphics

Research Overview : The term surface science is typically used for the branch of physics and chemistry that seeks to understand the physical, chemical, electrical and structural properties of surfaces in order to predict behavior under various conditions and stimuli. In computer vision and graphics, the interaction of surfaces with light is of fundamental interest in order to predict and simulate appearance. Appearance depends on view, illumination and the scale at which the surface is observed. Real world surfaces are seldom homogeneous but instead exhibit some type of texture which may be a color or albedo variation as in a checkerboard, a paisley print or zebra stripes. Very often in real-world scenes, texture is instead due to a surface height variation, e.g. pebbles, gravel, foliage and any rough surface. This type of texture is termed 3D texture. When viewing a 3D textured surface from a distance, local surface variations are subpixel and local intensity is uniform so appearance is characterized by the BRDF (bidirectional reflectance distribution function). At fine scale, where the surface variations give rise to local intensity variations, appearance can be characterized by the BTF (bidirectional texture function). Current representations of surfaces including are often overly simplified and cannot fully capture the complexities of real world appearance. The current research in this group is centered on four main tasks: (1) surface modeling, (2) surface measurements, (3) integration of surface models into synthesis and recognition algorithms, and (4) advancement of surface models to obtain object and scene models that utilize a large range of cues such as depth and motion. Related past work includes the CUReT database which contains BTF and BRDF measurements from over 60 different samples, each observed with over 200 different combinations of viewing and illumination directions.

Publications

People

Kristin J. Dana

Oana G. Cula

Jing Wang

Raghunand Makonahalli

Alex Wu

Alumni

Jayant Susikumar
Jasvinder Singh