Sample-Driven Segmented Mesh Deformation

Firstly in this paper, we brief history of mesh deform methods shortly. Together with a few basics and often applied techniques. Specifically, a detailed introduction about a recent technique will be made. This method provides deformation based on the deformation examples, which is Mesh Inverse Kinematics. Its uniqueness separates it from traditional methods by introducing deformation under sample-based subspace. Based on observation of results and the theory behind Mesh Inverse Kinematics, we introduce the idea to segment mesh, theoretically this extends the original subspace by introducing mesh structure correlation information, and gives better usage over the given examples. This method offers its users an adjustable mechanism, which is the mesh segmentation, to tune a constraint of correlation between structures of mesh itself before mesh deformation, resulting in more variant deformation result without the loss of the merits of example-based deformation's nature. By carefully choosing the segmentation method or manually segment the mesh in our method, the users can more adaptively control the space and specific parts of the mesh they want to deform with usually less redundant deformation samples required.