Title :
Computing optical flow based on the mass-conserving assumption
Author_Institution :
Sch. of Comput. Sci. & Eng., Hebrew Univ., Jerusalem, Israel
Abstract :
The image formation process involves both geometric and photometric aspects. The definition of optical flow is descriptive and the conformity of optical flow to motion flow is closely related to the algorithm, or the assumption that one takes for optical flow computation. The paper presents an algorithm for optical flow computation based on the mass-conserving assumption by first showing that this assumption is valid under appropriate conditions in both geometric and photometric views. Further discussions on these issues and preliminary experimental results are given
Keywords :
brightness; computer vision; image sequences; least squares approximations; geometric aspects; image formation process; mass-conservation assumption; optical flow; photometric aspects; Brightness; Equations; Geometrical optics; Image analysis; Image motion analysis; Layout; Optical computing; Optical variables control; Photometry; Shape;
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
Print_ISBN :
0-7695-0750-6
DOI :
10.1109/ICPR.2000.903721