Abstract:Convolution is conventionally defined as a linear operation on functions of one or more variables which commutes with shifts. Group convolution generalizes the concept to linear operations on functions of group elements representing more general geometric transformations and which commute with those transformations. Since similarity transformation is the most general geometric transformation on images that preserves shape, the group convolution that is equivariant to similarity transformation is the most general shape preserving linear operator. Because similarity transformations have four free parameters, group convolutions are defined on four-dimensional, joint orientation-scale spaces. Although prior work on equivariant linear operators has been limited to discrete groups, the similarity group is continuous. In this paper, we describe linear operators on discrete representations that are equivariant to continuous similarity transformation. This is achieved by using a basis of functions that is it joint shiftable-twistable-scalable. These pinwheel functions use Fourier series in the orientation dimension and Laplace transform in the log-scale dimension to form a basis of spatially localized functions that can be continuously interpolated in position, orientation and scale. Although this result is potentially significant with respect to visual computation generally, we present an initial demonstration of its utility by using it to compute a shape equivariant distribution of closed contours traced by particles undergoing Brownian motion in velocity. The contours are constrained by sets of points and line endings representing well known bistable illusory contour inducing patterns.

Similarity Equivariant Linear Transformation of Joint Orientation-Scale Space Representations

Automating Image Morphing Using Structural Similarity on a Halfway Domain

Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data

Empowering Networks With Scale and Rotation Equivariance Using A Similarity Convolution

A Method of Image Recognition with Rotation and Scale Invariability

On Similarity Transformation Problems: Globally Optimal Results and Applications

Similarity Transformation Operators As The Images Of Classical Symplectic Transformations In Coherent State Representation

The Compression and Stretching of Similaritons with Nonlinear Tunneling in Optical Fibers

Similarity-Invariant Subspaces and Similarity-Preserving Linear Maps

A similarity computing algorithm for volumetric data sets

A Linear Operator and Associated Families of Meromorphically Multivalent Functions

Convolution theorems for the linear canonical transforms

Exact Spatial Similaritons for the Generalized (2+1)-Dimensional Nonlinear Schrödinger Equation with Distributed Coefficients

Positive Transversality Via Transfer Operators and Holomorphic Motions with Applications to Monotonicity for Interval Maps

Method of similar operators in harmonious Banach spaces

Similarity Transformations in One- and Two-Mode Fock Space

New Convolution Theorem for the Linear Canonical Transform and Its Translation Invariance Property

Affine Invariance in Continuous-Domain Convolutional Neural Networks

A Tour of Convolutional Networks Guided by Linear Interpreters

Linear Canonical Transform Related Operators and Their Applications to Signal Analysis — Part I: Fundamentals

Linear Transformations of Vertex Operators Presentations of Coupled Hall–Littlewood Functions