Computer Analysis of Objects' Movement in Image Sequences: Methods and Applications

Abstract

Computer analysis of objects' movement in image sequences is a very complex problem, considering that it usually involves tasks for automatic detection, matching, tracking, motion analysis and deformation estimation. In spite of its complexity, this computational analysis has a wide range of important applications; for instance, in surveillance systems, clinical analysis of human gait, objects recognition, pose estimation and deformation analysis. Due to the extent of the purposes, several difficulties arise, such as the simultaneous tracking of manifold objects, their possible temporary occlusion or definitive disappearance from the image scene, changes of the viewpoints considered in images acquisition or of the illumination conditions, or even non-rigid deformations that objects may suffer in image sequences. In this paper, we present an overview of several methods that may be considered to analyze objects' movement; namely, for their segmentation, tracking and matching in images, and for estimation of the deformation involved between images.