Abstract
An affifine invariant representation is constructed with a cascade of invariants, which preserves information for classifification. A joint translation and rotation invariant representation of image patches is calculated with a scattering transform. It is implemented with a deep convolution network, which computes successive wavelet transforms and modulus non-linearities. Invariants to scaling, shearing and small deformations are calculated with linear operators in the scattering domain. State-of-the-art classifification results are obtained over texture databases with uncontrolled viewing conditions.