Photometric stereo is a computer vision technique that estimates the surface normals of objects by observing them under different lighting conditions. It relies on the fact that the amount of light reflected by a surface depends on its orientation. The technique has been expanded to handle various situations, such as extended light sources and non-Lambertian surface finishes, and current research focuses on improving its performance in the presence of shadows, highlights, and non-uniform lighting.
Carnegie Mellon University
Spring 2022
This course gives an expansive introduction to computer vision, focusing on image processing, recognition, geometry-based and physics-based vision, and video analysis. Students will gain practical experience solving real-life vision problems. It requires a good understanding of linear algebra, calculus, and programming.
No concepts data
+ 19 more concepts