ER 03: Vision and Machine-learning

From the 24th to the 28th of January 2011 — ENS Lyon

The lectures will be delivered in English.

Goals

Some of the notable recent successes in automated visual recognition are results of combining advanced visual representations together with powerful machine learning techniques. The objective of this winter school is to provide an overview of basic tools and some latest advances in visual recognition together with the related machine learning algorithms. A particular emphasis will be given on topics related to recognition of objects and human actions in video and still images. Lectures will be given by experts in visual recognition and machine learning. Lectures will be complemented by practical sessions, where participants will obtain hands-on experience with the discussed material.

Speakers :

  1. Zaid Harchaoui (INRIA, LEAR, INRIA Rhône-Alpes)
  2. Ivan Laptev (INRIA, WILLOW, INRIA Rocquencourt)
  3. Cordelia Schmid (INRIA, LEAR, INRIA Rhône-Alpes)
  4. Josef Sivic (INRIA, WILLOW, INRIA Rocquencourt)

Tentative outline and schedule of the course

[L] means a lecture, and [E] means an exercise session.

If possible, please bring a laptop with you for the exercise sessions.

  • Monday, January 24

    • [L] 8:15 AM – 9:15 AM. Unsupervised learning (Zaid Harchaoui). Slides available here, here and here.
    • [L] 9:30 AM – 10:30 AM. Unsupervised learning (Zaid Harchaoui)
    • [E] 10:45 AM – 11:45 AM. Unsupervised learning (Zaid Harchaoui)
    • [L] 1:30 PM – 2:30 PM. Image features (Cordelia Schmid). Slides available here and here (ppt files) or here and here (pdf files).
    • [L] 2:45 PM – 3:45 PM. Image features (Cordelia Schmid)
    • [E] 4:00 PM – 5:00 PM. Image features (Cordelia Schmid, Josef Sivic)
  • Tuesday, January 25

    • [L] 8:15 AM – 9:15 AM. Camera geometry and Image Alignment (Josef Sivic). Slides available here.
    • [L] 9:30 AM – 10:30 AM. Camera geometry and Image Alignment (Josef Sivic)
    • [E] 10:45 AM – 11:45 AM. Camera geometry and Image Alignment (Ivan Laptev and Josef Sivic).
    • [L] 1:30 PM – 3:00 PM. Efficient visual search (Josef Sivic). Slides available here.
    • [L] 3:15 PM – 4:45 PM. Efficient visual search (Cordelia Schmid). Slides available here (ppt file) and here (pdf file).
  • Wednesday, January 26

    • [L] 8:15 AM – 9:15 AM. Bag-of-features models for category-level classification (Cordelia Schmid). Slides available here (ppt file) and here (pdf file).
    • [L] 9:30 AM – 10:30 AM. Bag-of-features models for category-level
      classification (Cordelia Schmid)
    • [E] 10:45 AM – 11:45 AM. Bag-of-features models for category-level classification (Ivan Laptev, Cordelia Schmid and Josef Sivic)
    • [L] 1:30 PM – 2:30 PM. Category-level localization (Ivan Laptev). Slides available here.
    • [L] 2:45 PM – 3:45 PM. Category-level localization (Ivan Laptev)
    • [E] 4:00 PM – 5:00 PM. Category-level localization (Ivan Laptev and Josef Sivic)
  • Thursday, January 27

    • [L] 8:15 AM – 9:15 AM. Motion and human actions (Ivan Laptev). Slides available here.
    • [L] 9:30 AM – 10:30 AM. Motion and human actions (Ivan Laptev)
    • [L] 10:45 AM – 11:45 AM. Motion and human actions (Ivan Laptev)
  • Friday, January 28

    • [L] 8:15 AM – 9:15 AM. Supervised learning (Zaid Harchaoui)
    • [L] 9:30 AM – 10:30 AM. Supervised learning (Zaid Harchaoui)
    • [E] 10:45 AM – 11:45 AM. Supervised learning (Zaid Harchaoui)

Total hours: 18 hours of lectures + 6 hours of practical exercises.

Prerequisites

The course is self-contained.

Bibliography

  • D.A. Forsyth and J. Ponce, “Computer Vision: A Modern Approach, Prentice-Hall, 2003.
  • J. Ponce, M. Hebert, C. Schmid, and A. Zisserman, “Toward category-level object recognition”, Springer LNCS, 2007.
  • R. Szeliski, “Computer Vision: Algorithms and Applications”, Springer, 2010.
  • O. Faugeras, Q.T. Luong, and T. Papadopoulo, “Geometry of Multiple Images,” MIT Press, 2001.
  • R. Hartley and A. Zisserman, “Multiple View Geometry in Computer Vision”, Cambridge University Press, 2004.
  • J. Koenderink, “Solid Shape”, MIT Press, 1990.

Registration

There are no registration fees. For organization reasons it is however necessary to register online; see the registration page.

Location and schedule

All the lectures will take place in “amphithéatre B”, at the 3rd floor of the GN1 building (main building of the “exact sciences” part of ENS Lyon, a.k.a. Campus Jacques Monod).

Local organizers

There are several useful informations on the main page of the research schools in Computer Science at ENS Lyon.

2 comments to ER 03: Vision and Machine-learning