This paper presents a framework for dynamic 3D shape reconstruction from multi-viewpoint images using a deformable mesh model. With our method, we can obtain 3D shape and 3D motion of the object simultaneously. We represent the shape by a surface mesh model and the motion by translations of its vertices, i.e., deformation. Thus global and local topological structure of the mesh are preserved from frame to frame. This helps us to analyse the motion of the object, to compress the 3D data, and so on. Our model deforms its shape so as to satisfy several constraints. This constraint-based deformation provides a computational framework to integrate several reconstruction cues such as surface texture, silhouette, and motion flow observed in multi-viewpoint images.
Dynamic 3D shape from multi-viewpoint images using deformable mesh model
01.01.2003
391706 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Dynamic 3D Shape from Multi-Viewpoint Images using Deformable Mesh Models
British Library Conference Proceedings | 2003
|3D anatomical shape atlas construction using mesh quality preserved deformable models
British Library Online Contents | 2013
|Deformable Model-Based Shape and Motion Analysis from Images using Motion Residual Error
British Library Conference Proceedings | 1998
|3D shape recovery using a deformable model
British Library Online Contents | 1995
|