Invariant Scattering Convolution Networks

TL;DR

This paper introduces wavelet scattering networks that produce translation-invariant, deformation-stable image representations, improving classification by capturing high-frequency information and higher-order moments.

Contribution

It presents a mathematically grounded deep convolution network architecture that enhances image classification and texture discrimination, with state-of-the-art results.

Findings

Achieves state-of-the-art results on handwritten digit classification.

Effectively discriminates textures with identical Fourier spectra.

Provides a theoretical analysis explaining properties of deep convolution networks.

Abstract

A wavelet scattering network computes a translation invariant image representation, which is stable to deformations and preserves high frequency information for classification. It cascades wavelet transform convolutions with non-linear modulus and averaging operators. The first network layer outputs SIFT-type descriptors whereas the next layers provide complementary invariant information which improves classification. The mathematical analysis of wavelet scattering networks explains important properties of deep convolution networks for classification. A scattering representation of stationary processes incorporates higher order moments and can thus discriminate textures having the same Fourier power spectrum. State of the art classification results are obtained for handwritten digits and texture discrimination, using a Gaussian kernel SVM and a generative PCA classifier.