Related Experiment Video
Updated: Jun 13, 2026

Automated 3D Optical Coherence Tomography to Elucidate Biofilm Morphogenesis Over Large Spatial Scales
Published on: August 21, 2019
Shape and spatially-varying BRDFs from photometric stereo
Dan B Goldman1, Brian Curless, Aaron Hertzmann
1Adobe Systems, Inc., Seattle, WA 98103, USA. dgoldman@adobe.com
This study presents a photometric stereo technique for surfaces with varying Bidirectional Reflectance Distribution Functions (BRDFs). The method accurately recovers shape and material properties, enabling realistic renderings and interactive editing.
Area of Science:
- Computer Vision
- Computer Graphics
- Material Science
Background:
- Photometric stereo typically assumes uniform surface properties.
- Spatially-varying Bidirectional Reflectance Distribution Functions (BRDFs) present a significant challenge for existing methods.
- Accurate material and shape recovery is crucial for realistic rendering and analysis.
Purpose of the Study:
- To develop a photometric stereo method for surfaces with spatially-varying BRDFs.
- To recover both surface geometry and detailed material properties (BRDFs and weight maps).
- To enable accurate re-rendering under novel lighting and facilitate interactive editing.
Main Methods:
- An optimization-based photometric stereo approach.
- Constraining pixels to be representable by a combination of at most two fundamental materials.
- Leveraging observations about the composition of most objects from a limited set of materials.
Main Results:
- Successful recovery of surface shape and spatially-varying BRDFs.
- Generation of accurate weight maps for material composition.
- Demonstration of high-fidelity re-renderings under new lighting conditions.
- Validation across a variety of object types.
Conclusions:
- The proposed method effectively handles complex surfaces with spatially-varying diffuse and specular properties.
- The approach provides a robust framework for shape and material estimation.
- Enables new possibilities for interactive manipulation and re-rendering of 3D objects.
Related Concept Videos
Depth Perception and Spatial Vision
Curvilinear Motion: Rectangular Components
As the car advances, its position evolves over time. Quantifying the car's velocity involves computing the time...
Divergence Theorem in 3D Space
Methods of Obtaining Topography
Cylinders in Three-Dimensional Space
Quadric Surfaces
