Vector Neurons: A General Framework for SO(3)-Equivariant Networks
Authors:
Top Cited Papers
- 1 October 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 12180-12189
Abstract
Invariance and equivariance to the rotation group have been widely discussed in the 3D deep learning community for pointclouds. Yet most proposed methods either use complex mathematical tools that may limit their accessibility, or are tied to specific input data types and network architectures. In this paper, we introduce a general framework built on top of what we call Vector Neuron representations for creating SO (3) -equivariant neural networks for pointcloud processing. Extending neurons from 1D scalars to 3D vectors, our vector neurons enable a simple mapping of SO (3) actions to latent spaces thereby providing a framework for building equivariance in common neural operations – including linear layers, non-linearities, pooling, and normalizations. Due to their simplicity, vector neurons are versatile and, as we demonstrate, can be incorporated into diverse network architecture backbones, allowing them to process geometry inputs in arbitrary poses. Despite its simplicity, our method performs comparably well in accuracy and generalization with other more complex and specialized state-of-the-art methods on classification and segmentation tasks. We also show for the first time a rotation equivariant reconstruction network. Source code is available at https://github.com/FlyingGiraffe/vnn.Keywords
All Related Versions
This publication has 19 references indexed in Scilit:
- Rotation Invariant Convolutions for 3D Point Clouds Deep Learning2019
- Effective Rotation-Invariant Point CNN with Spherical Harmonics Kernels2019
- ClusterNet: Deep Hierarchical Cluster Network With Rigorously Rotation-Invariant Representation for Point Cloud Analysis2019
- Occupancy Networks: Learning 3D Reconstruction in Function Space2019
- Learning SO(3) Equivariant Representations with Spherical CNNs2018
- Point convolutional neural networks by extension operatorsACM Transactions on Graphics, 2018
- Interpretable Transformations with Encoder-Decoder Networks2017
- A scalable active framework for region annotation in 3D shape collectionsACM Transactions on Graphics, 2016
- Learning 6D Object Pose Estimation Using 3D Object Coordinates2014
- Transforming Auto-Encoders2011