Multilevel Optimization for Dense Motion Estimation

This monograph offers design for fast and reliable technique in the dense motion estimation. This Multilevel Optimization for Dense Motion Estimation work blends both theory and applications to equip reader with an understanding of basic concepts necessary to apply in solving dense motion in a sequence of images. Illustrating well-known variation models for dealing with optical flow estimation, this monograph introduces variation models with applications. A host of variation models are outlines such as Horn-Schunck model, Contrast Invariation Models and Models for Large Displacement. Special attention is also given to multilevel optimization techniques namely multiresolution and multigrid methods to improve the convergence of the global optimum when compared to using only one level resolution in the context of computer vision. This monograph is a robust resource that provides insightful introduction to the field of image processing with its theory and applications. Overall, Multilevel Optimization for Dense Motion Estimation is highly recommended for scientists and engineers for an excellent choice for references and self-study.