Generic Deep Networks with Wavelet Scattering

Edouard Oyallon, St├ęphane Mallat, Laurent Sifre

Dec 23, 2013 (modified: Dec 23, 2013) ICLR 2014 workshop submission readers: everyone
  • Decision: submitted, no decision
  • Abstract: We introduce a two-layer wavelet scattering network, which involves no learning, for object classification. This scattering transform computes a spatial wavelet transform on the first layer and a joint wavelet transform along spatial, angular and scale variables in the second layer. Image classification results are given on Caltech databases.