Learning Implicit Functions for Topology-Varying Dense 3D Shape Correspondence

2020-12-30 2020-12-30 约 131 字预计阅读 1 分钟

Learning implicit functions for topology-varying dense 3D shape correspondence

NeurIPS2020 Advances in Neural Information Processing Systems

Feng Liu, Xiaoming Liu

Michigan State University

implicit function, correspondence score, category shape correspondence

Motivation

给定一组3D shapes（点云），category-specific model 无监督地学出逐pair（source 与target之间）的correspondence和部件segmentation
即使拓扑不一样也能学到相关性
有correspondence相关性分数输出