Reconnaissance d’objets et vision artificielle 2012/2013
Object recognition and computer vision 2012/2013

Jean Ponce, Ivan Laptev, Cordelia Schmid and Josef Sivic


Course Information

Room: ENS Ulm  Salle UV aile Rataud, 45 rue d'Ulm

Class time: Tuesday 16:15-19:15

News:

  1. Internship proposals at Willow
  2. Schedule of the final project presentations. (Tue Dec 11 and Wed Dec 12. See the link for details.)
  3. Assignment 3 “Simple face detector” is out.
  4. Final project topics are out.
  5. Assignment 2 “Image classification” is out!
  6. Assignment 1 “Instace-level recognition” is out.
  7. Matlab tutorial will be organized on September 27, 15:00-17:00 at the INRIA/Willow lab, Salle Orange1. Materials for the tutorial are here.

List of received reports:

https://docs.google.com/spreadsheet/pub?key=0Aszo_00aoj7SdFpiVGxCd2lSSGdkQmhuODhhZHQ2Y3c&output=html

Course description

Automated  object  recognition -- and  more  generally  scene  analysis -- from  photographs  and videos  is  the  grand  challenge  of  computer  vision.  This  course  presents  the  image,  object,  and scene models, as well as the methods and algorithms, used today to address this challenge.

Assignments

There will be three programming assignments representing 50% (10% + 20% + 20%) of the grade. The supporting materials for the programming assignments and final projects will be in Matlab.

Final project

The final project will represent 50% of the grade.

Computer vision and machine learning talks

You are welcome to attend seminars in the Willow group. Please see the current seminar schedule. Typically, these are one hour research talks given by visiting speakers. The talks are at 23 avenue d'Italie. Ring the bell to get into the building, then take the elevator to the 5th floor.

Course schedule (subject to change):

Lecture

Date

Topic and reading materials.

Slides

1

Sep 25

Introduction (J. Ponce);

PDF1
PDF2

2

Oct 2

Instance-level recognition I. - Local invariant features (C. Schmid)

Materials: Mikolajczyk & Schmid, Scale and affine invariant interest point detectors, IJCV 2004; D. Lowe, Distinctive image features from scale-invariant keypoints, IJCV 2004, R. Szeliski (pdf), Sections 4.1, 4.1.1 and 4.1.2 from Chapter 4: Feature detection and matching.

Assignments:
Assignment 1 out

PDF1

PDF2

Oct 9

NO LECTURE

3

Oct 16

Instance-level recognition II. - Correspondence, efficient visual search (I. Laptev)

Materials: R. Szeliski (pdf), Sections 4.1.3 (feature matching) and 6.1 (feature-based alignment); Muja & Lowe, Fast approx. nearest neighbors with automatic algorithm configuration, VISAPP'09; Sivic & Zisserman, Video Google: Efficient visual search of videos (chapter from this book)

Philbin et al., Object retrieval with large vocabularies and fast spatial matching, CVPR'07.

Background materials (geometry): History: J. Mundy - Object recognition in the geometric era: A retrospective.; Camera geometry: Forsyth&Ponce Ch.1-2. Hartley&Zisserman - Ch.6

PDF1

PDF2

4

Oct 23

Instance-level recognition III. - Very large scale image indexing (C. Schmid)

Materials: Jegou et al., Improving bag-of-features for large scale image search, IJCV 2010; Jegou et al., Aggregating local image descriptors into compact codes, PAMI 2011;

Bag-of-feature models for category-level recognition (C. Schmid)

Materials: Csurka et al., Visual categorization with bags of keypoints, 2004

Assignments:
Assignment 1 due

Assignment 2 out

PDF1

PDF2

5

Oct 30

Category-level localization I. (J. Sivic)

Materials: Fergus et al., A Sparse Object Category Model for Efficient Learning and Complete Recognition (constellation model) (chapter from this book); Leibe et al., An Implicit Shape Model for Combined Object Categorization and Segmentation (chapter from this book); Dalal&Triggs, A histogram of oriented gradients (HOG) for human detection, CVPR'05; Felzenszwalb et al., A Discriminatively Trained, Multiscale, Deformable Part Model, CVPR’08; Pascal VOC Challenge

Assignments:
Topic suggestions for the final project are out

PDF

6

Nov 6

Category-level localization II. - Efficient fitting of pictorial structures; Human pose estimation (J. Sivic)

Motion and human actions I. (I. Laptev)

Assignments:

Assignment 2 due

Final project proposal due (Nov 9).

Assignment 3 out

PDF

PDF

7

Nov 13

Neural networks; Optimization methods (N. Le Roux)

 

PDF

8

Nov 20

Motion and human actions II. (I. Laptev)

Sparse coding and dictionary learning for image analysis (J. Ponce)

Materials: Bach,Mairal,Ponce,Sapiro, Tutorial on sparse coding and dictionary learning for image analysis, at CVPR'10

Assignments:
Assignment 3 due

PDF1

PDF2

9

Nov 27

Motion and human actions III.

Face detection and recognition. (C. Schmid)

Materials:

P. Viola, M. Jones: Robust Real-Time Face Detection. International Journal of Computer Vision 57(2): 137-154 (2004)

M. Guillaumin, T. Mensink, J. Verbeek, C. Schmid

Face recognition from caption-based supervision

International Journal of Computer Vision, Springer, 2012, 96 (1), pp. 64-82

PDF1

PDF2

10

Dec 4

Scenes and objects (J. Sivic and I. Laptev)

PDF1
PDF2

11

Dec 11

Dec 12

Final project presentations and evaluation (I. Laptev, J. Sivic)

Presentation schedule. 

Tue Dec 11 presentations are at the standard class location and time (16:15-19:15 ENS Ulm).

Wed Dec 12 presentations are at INRIA, 23 Av. d’Italie, 75013. See the link above for directions and exact time schedule.

Relevant literature:

[1]

D.A. Forsyth and J. Ponce, "Computer Vision: A Modern Approach", Prentice-Hall, 2003

[2]

J. Ponce, M. Hebert, C. Schmid and A. Zisserman "Toward Category-Level Object Recognition", Lecture Notes in Computer Science 4170, Springer-Verlag, 2007

[3]

O. Faugeras, Q.T. Luong, and T. Papadopoulo, "Geometry of Multiple Images", MIT Press, 2001.

[4]

R. Hartley and A. Zisserman, "Multiple View Geometry in Computer Vision", Cambridge University Press, 2004.

[5]

J. Koenderink, "Solid Shape", MIT Press, 1990

[6]

R. Szeliski, "Computer Vision: Algorithms and Applications", 2009. A draft of a new book, which can be downloaded online.