This paper presents a hierarchical estimation framework for computing various representations of motion information in image sequences. The framework consists of a global model that constraints the overall structure of the motion, a local model used in the estimation process, and a coarse-fine refinement strategy. Four specific motion models—affine flow, planar surface flow, rigid body motion, and general optical flow—are described and applied to specific examples. The key features of the framework include the use of a global model like rigidity constraints and a local model such as constant displacement over a patch. The coarse-fine refinement strategy is employed to improve accuracy and efficiency, addressing issues like aliasing and computational efficiency. The paper also discusses the advantages of hierarchical estimation techniques and provides detailed descriptions of each motion model, including their estimation algorithms and experimental results. The framework unifies diverse model-based estimation algorithms and supports the combined use of parametric global models and local models, making it applicable to a wide range of image processing applications.This paper presents a hierarchical estimation framework for computing various representations of motion information in image sequences. The framework consists of a global model that constraints the overall structure of the motion, a local model used in the estimation process, and a coarse-fine refinement strategy. Four specific motion models—affine flow, planar surface flow, rigid body motion, and general optical flow—are described and applied to specific examples. The key features of the framework include the use of a global model like rigidity constraints and a local model such as constant displacement over a patch. The coarse-fine refinement strategy is employed to improve accuracy and efficiency, addressing issues like aliasing and computational efficiency. The paper also discusses the advantages of hierarchical estimation techniques and provides detailed descriptions of each motion model, including their estimation algorithms and experimental results. The framework unifies diverse model-based estimation algorithms and supports the combined use of parametric global models and local models, making it applicable to a wide range of image processing applications.