SearchPublisher and Institutes
You are here: Home :: Area NEM :: Computer science
Andrés Bruhn, Andrés Bruhn
Motion Estimation in Real-Time with Optimisation Methods
Keywords: image processing, scene analysis, numerical analysis, partial differential equations, numerical linear algebra
The estimation of motion information from image sequences is one of the key problems in computer vision. In this context, global optimisation methods play an important role, since they allow for both a precise and dense estimation of the displacement field between consecutive frames. This thesis contributes in two ways to the research in motion analysis: (I) On one hand, novel models for optimisation techniques are presented that yield the most accurate results in the entire literature. (II) On the other hand, specifically adapted multigrid algorithms are developed, that enable the computation of the results in real-time. Thus, for the first time, it becomes possible to apply such high accuracy techniques also in practice.
it – Information Technology (vormals it+ti), Oldenbourg Wissenschaftsverlag
Print ISSN: 1611-2776
Volume: 50, 01/2008
Pages: 066 - 069