Steerable CNNs

TL;DR

Steerable CNNs are an efficient class of equivariant networks that leverage mathematical symmetry principles to improve image classification performance and parameter efficiency.

Contribution

This paper introduces steerable CNNs, a novel class of equivariant networks that incorporate symmetry-based representations for improved efficiency and accuracy.

Findings

Achieved state-of-the-art results on CIFAR image classification.

Demonstrated parameter efficiency through symmetry-based filter design.

Provided a mathematical framework for constructing steerable representations.

Abstract

It has long been recognized that the invariance and equivariance properties of a representation are critically important for success in many vision tasks. In this paper we present Steerable Convolutional Neural Networks, an efficient and flexible class of equivariant convolutional networks. We show that steerable CNNs achieve state of the art results on the CIFAR image classification benchmark. The mathematical theory of steerable representations reveals a type system in which any steerable representation is a composition of elementary feature types, each one associated with a particular kind of symmetry. We show how the parameter cost of a steerable filter bank depends on the types of the input and output features, and show how to use this knowledge to construct CNNs that utilize parameters effectively.