Gauge Equivariant Mesh CNNs: Anisotropic convolutions on geometric graphs

03/11/2020
by   Pim de Haan, et al.
116

A common approach to define convolutions on meshes is to interpret them as a graph and apply graph convolutional networks (GCNs). Such GCNs utilize isotropic kernels and are therefore insensitive to the relative orientation of vertices and thus to the geometry of the mesh as a whole. We propose Gauge Equivariant Mesh CNNs which generalize GCNs to apply anisotropic gauge equivariant kernels. Since the resulting features carry orientation information, we introduce a geometric message passing scheme defined by parallel transporting features over mesh edges. Our experiments validate the significantly improved expressivity of the proposed model over conventional GCNs and other methods.

READ FULL TEXT
research
07/06/2020

Distance-Geometric Graph Convolutional Network (DG-GCN)

The distance-geometric graph representation adopts a unified scheme (dis...
research
10/23/2020

Primal-Dual Mesh Convolutional Neural Networks

Recent works in geometric deep learning have introduced neural networks ...
research
01/31/2019

GEOMetrics: Exploiting Geometric Structure for Graph-Encoded Objects

Mesh models are a promising approach for encoding the structure of 3D ob...
research
04/02/2020

DualConvMesh-Net: Joint Geodesic and Euclidean Convolutions on 3D Meshes

We propose DualConvMesh-Nets (DCM-Net) a family of deep hierarchical con...
research
03/28/2021

Picasso: A CUDA-based Library for Deep Learning over 3D Meshes

We present Picasso, a CUDA-based library comprising novel modules for de...
research
07/30/2019

Orientation-aware Semantic Segmentation on Icosahedron Spheres

We address semantic segmentation on omnidirectional images, to leverage ...
research
11/24/2020

An Optimal Triangle Projector with Prescribed Area and Orientation, Application to Position-Based Dynamics

The vast majority of mesh-based modelling applications iteratively trans...

Please sign up or login with your details

Forgot password? Click here to reset