Shape Reconstruction from 3D and 2D Data Using PDE-Based Deformable Surfaces

摘要: In this paper, we propose a new PDE-based methodology for deformable surfaces that is capable of automatically evolving its shape to capture the geometric boundary data and simultaneously discover underlying topological structure. Our model can handle multiple types (such as volumetric data, 3D point clouds 2D image data), using common mathematical framework. The deformation behavior governed by partial differential equations (e.g. weighted minimal surface flow). Unlike level-set approach, our always has an explicit representation geometry topology. regularity stability numerical integration process are ensured powerful Laplacian tangential smoothing operator. By allowing local adaptive refinement mesh, accurately represent sharp features. We have applied reconstruction from unorganized view images. versatility robustness allow application challenging problem reconstruction. approach unique in combination simultaneous use high number arbitrary camera views with mesh intuitive easy-to-interact-with. model-based selects best reconstruction, allows visibility checking progressive more images become available. results extensive experiments on synthetic real demonstrate robustness, accuracy visual quality.