Scale-Equivariant Steerable Networks

Abstract

The effectiveness of Convolutional Neural Networks (CNNs) has been substantially attributed to their built-in property of translation equivariance. However, CNNs do not have embedded mechanisms to handle other types of transformations. In this work, we pay attention to scale changes, which regularly appear in various tasks due to the changing distances between the objects and the camera. First, we introduce the general theory for building scale-equivariant convolutional networks with steerable filters. We develop scale-convolution and generalize other common blocks to be scale-equivariant. We demonstrate the computational efficiency and numerical stability of the proposed method. We compare the proposed models to the previously developed methods for scale equivariance and local scale invariance. We demonstrate state-of-the-art results on MNIST-scale dataset and on STL-10 dataset in the supervised learning setting.

References in corpus (16)

• Adam: A Method for Stochastic Optimization
• Improved Regularization of Convolutional Neural Networks with Cutout
• Wide Residual Networks
• Group Equivariant Convolutional Networks
• Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point clouds
• On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups
• Gauge Equivariant Convolutional Networks and the Icosahedral CNN
• Locally Scale-Invariant Convolutional Neural Networks
• Scale-Invariant Convolutional Neural Networks
• General $E(2)$-Equivariant Steerable CNNs
• N-body Networks: a Covariant Hierarchical Neural Network Architecture for Learning Atomic Potentials
• Convolutional Networks for Spherical Signals
• Deep Scale-spaces: Equivariance Over Scale
• Scale equivariance in CNNs with vector fields
• Scale Steerable Filters for Locally Scale-Invariant Convolutional Neural Networks
• Equivariant Transformer Networks

Cited by in corpus (10)

• Learning Invariances in Neural Networks
• Scale-invariant scale-channel networks: Deep networks that generalise to previously unseen scales
• Dilated convolution with learnable spacings
• SE(3)-Equivariant and Noise-Invariant 3D Rigid Motion Tracking in Brain MRI
• ESP: Exploiting Symmetry Prior for Multi-Agent Reinforcement Learning
• Covariance properties under natural image transformations for the generalized Gaussian derivative model for visual receptive fields
• ACDC: Weight Sharing in Atom-Coefficient Decomposed Convolution
• Do CNNs Encode Data Augmentations?
• Exploiting Redundancy: Separable Group Convolutional Networks on Lie Groups
• Built-in Elastic Transformations for Improved Robustness