Equivariant Wavelets: Fast Rotation and Translation Invariant Wavelet Scattering Transforms
- 7 April 2022
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 45 (2) , 1716-1731
Abstract
Wavelet scattering networks, which are convolutional neural networks (CNNs) with fixed filters and weights, are promising tools for image analysis. Imposing symmetry on image statistics can improve human interpretability, aid in generalization, and provide dimension reduction. In this work, we introduce a fast-to-compute, translationally invariant and rotationally equivariant wavelet scattering network (EqWS) and filter bank of wavelets (triglets). We demonstrate the interpretability and quantify the invariance/equivariance of the coefficients, briefly commenting on difficulties with implementing scale equivariance. On MNIST, we show that training on a rotationally invariant reduction of the coefficients maintains rotational invariance when generalized to test data and visualize residual symmetry breaking terms. Rotation equivariance is leveraged to estimate the rotation angle of digits and reconstruct the full rotation dependence of each coefficient from a single angle. We benchmark EqWS with linear classifiers on EMNIST and CIFAR-10/100, introducing a new second-order, cross-color channel coupling for the color images. We conclude by comparing the performance of an isotropic reduction of the scattering coefficients and RWST, a previous coefficient reduction, on an isotropic classification of magnetohydrodynamic simulations with astrophysical relevance.Keywords
All Related Versions
Funding Information
- National Science Foundation (DGE-1745303)
- National Science Foundation (AST-1614941)
This publication has 47 references indexed in Scilit:
- Deep roto-translation scattering for object classification2015
- scikit-image: image processing in PythonPeerJ, 2014
- Rotation, Scaling and Deformation Invariant Scattering for Texture Discrimination2013
- Group Invariant ScatteringCommunications on Pure and Applied Mathematics, 2012
- DENSITY STUDIES OF MHD INTERSTELLAR TURBULENCE: STATISTICAL MOMENTS, CORRELATIONS AND BISPECTRUMThe Astrophysical Journal, 2009
- Matplotlib: A 2D Graphics EnvironmentComputing in Science & Engineering, 2007
- IPython: A System for Interactive Scientific ComputingComputing in Science & Engineering, 2007
- The Design and Implementation of FFTW3Proceedings of the IEEE, 2005
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998
- Shiftable multiscale transformsIEEE Transactions on Information Theory, 1992