Anisotropic Hierarchical Motion Estimation Method Based on Decomposition of the Functional Domain

✍ Scribed by Dong-Gyu Sim; Rae-Hong Park

• Publisher: Elsevier Science
• Year: 1996
• Tongue: English
• Weight: 455 KB
• Volume: 7
• Category: Article
• ISSN: 1047-3203
• DOI: 10.1006/jvci.1996.0023

✦ Synopsis

Conventional hierarchical motion estimation methods have been proposed for fast convergence and detection of large motions. However, they have a common drawback that large errors in motion estimation are propagated across motion discontinuities to neighboring regions, resulting in noticeable artifacts. These artifacts are caused by the constraint of motion continuity and the linear interpolation of motion vectors between two successive levels. Thus, in this paper we propose an effective hierarchical motion estimation method for robust estimation at motion discontinuities. The proposed intra-level algorithm is based on decomposition of the functional domain for optimizing the nonlinear intra-level motion estimation functional. Also we propose an anisotropic inter-level motion estimation equation rather than using the conventional linear projection scheme of motion fields. Computer simulations with several noiseless/ noisy sequences show that the proposed hierarchical motion estimation algorithm performs better than the conventional ones.