Performance characterization of fundamental matrix estimation under image degradation.

Autor: Torr, P.H.S., Zissermann, A.
Zdroj: Machine Vision & Applications; Mar1997, Vol. 9 Issue 5/6, p321-333, 13p
Abstrakt: The fundamental matrix represents the epipolar geometry between two images. We describe an algorithm for simultaneously estimating the fundamental matrix and corresponding points automatically from the two images. The performance of this algorithm is then assessed as the images are degraded by JPEG lossy compression. A number of performance measures are proposed and evaluated over image pairs corresponding to different camera motions and scene types. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index