Accès direct au contenu


Version anglaise


Accueil > Formations > Master MVA > Présentation des cours

Sparse wavelet representations and classification

Lecturer : Stéphane MALLAT, (ENS Ulm)

Objective of the course :

The course introduces sparse wavelet representation techniques, for compression, noise removal and for audio and image classification.

Topics :

  • Fourier transform, linear approximations and sampling theorems
  • Time-frequency representations
  • Wavelet orthogonal bases
  • Adaptive and non-linear wavelet approximations
  • Information theory for image and audio compression.
  • Linear and non-linear noise removal
  • Linear classifiers and curse of dimensionality
  • Invariants for classification
  • Deep Neural Networks
  • Image and audio signal recognition

    References :

               Une explorations des signaux en ondelettes", S. Mallat, Éditions de l'Ecole Polytechnique.