Stereo Matching from the Topological Viewpoint,
Abstract
A unifying abstract mathematical structure is presented for a number of vision problems, notably stereo, motion stereo, optic flow, and matching. Ideas from modern differential topology are presented and applied to the general matching problem, a common approach to stereo matching. Analogously to the Erlanger Programm, the task of computer vision can be viewed as finding invariants of irradiance functions under the rigid motion group of R sub 3. This paper describes this structure in the language of modern abstract mathematics, providing a framework for understanding and the possibility of applying powerful methods to resolve fundamental questions. The author proves a theorem which says that occlusion-free stereo matching requires at least 2 color dimensions or some knowledge of the imaging geometry.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1983
- Accession Number
- ADP001217
Entities
People
- A. Peter Blicher
Organizations
- Stanford University