GAMesh: Guided and Augmented Meshing
for Deep Point Networks
Nitin Agarwal
Gopi Meenakshisundaram
Interactive Graphics & Visualization Lab



Abstract

We present a new meshing algorithm called guided and augmented meshing, GAMesh, which uses a mesh prior to generate a surface for the output points of a point network. By projecting the output points onto this prior and simplifying the resulting mesh, GAMesh ensures a surface with the same topology as the mesh prior but whose geometric fidelity is controlled by the point network. This makes GAMesh independent of both the density and distribution of the output points, a common artifact in traditional surface reconstruction algorithms. We show that such a separation of geometry from topology can have several advantages especially in single-view shape prediction, fair evaluation of point networks and reconstructing surfaces for networks which output sparse point clouds. We further show that by training point networks with GAMesh, we can directly optimize the vertex positions to generate adaptive meshes with arbitrary topologies.

Summary Video



Code

[Code]


Paper

N. Agarwal, M. Gopi
GAMesh: Guided and Augmented Meshing for Deep Point Networks
3DV, 2020 [arXiv Preprint]
Supplementary Document
[Bibtex]


Application 1: Single-View 3D Reconstruction

GAMesh combines geometry from point networks and topology from implicit networks to reconstruct high fidelity meshes with correct topology for single-view reconstruction.


Application 2: Training with GAMesh

Point reconstruction networks can be trained with GAMesh, to generate adaptive meshes with arbritary topology.


Application 3: Fair Evaluation of Point Networks

We should evaluate both the output points and the output surfaces (from GAMesh) for evaluating point networks.


Application 4: Mesh Sparse Point Clouds

GAMesh can mesh points from various point networks and loss functions.


Related Work

Nitin Agarwal, Sung-eui Yoon, M Gopi Learning Embedding of 3D models with Quadric Loss. In BMVC, 2019. [PDF]
Nicholas Sharp, Maks Ovsjanikov PointTriNet: Learned Triangulation of 3D Point Sets. In ECCV, 2020. [PDF]
Rana Hanocka, Gal Metzer, Raja Giryes, Daniel Cohen-Or Point2Mesh: A Self-Prior for Deformable Meshes. In SIGGRAPH, 2020. [PDF]


Project template was borrowed from Richard Zhang