Designing Tomorrow's Category-Level 3D Object Recognition Systems:
An International Workshop

September 8 to September 10, 2003
Taormina, Sicily

Objectives

The ability to recognize living creatures and inanimate objects in photographs or video clips is a critical enabling technology for a wide range of applications including defense, health care,  human-computer interaction, image retrieval and data mining, industrial and personal robotics, manufacturing, scientific image analysis, space exploration, surveillance and security, and transportation. In fact, with the ever expanding array of imagery sources, some form of automatic object recognition technology must eventually be an integral part of every information system.  Despite 40 years of research, however, today's recognition systems are still largely unable to handle the extraordinarily wide range of appearances assumed by common objects in typical images.
The tenet of this workshop is that fundamental new advances in automated object recognition can be achieved by integrating the sophisticated geometric and physical image formation models developed in the computer vision community with the effective models of data distribution and classification procedures developed in the statistical learning theory and theoretical computer science communities.
This three-day workshop will bring together prominent computer vision, machine learning, and computational geometry researchers interested in the fundamental and applicative aspects of object recognition, as well as representative of industry and funding agencies. Its goals are (1) to promote the creation of an international object recognition community, with common datasets and evaluation procedures, (2) to map the state of the art and identify the main open problems and opportunities for synergistic research, and (3) to articulate the industrial and societal needs and opportunities for object recognition research worldwide.

Participants
  • Kobus Barnard (University of Arizona)
  • Stab Bileschi (MIT)
  • Andrew Blake (Microsoft)
  • Chris Bishop (Microsoft)
  • Stefan Carlsson (Royal Intitute of Technology)
  • Guillaume Charpiat (ENS/INRIA)
  • Jeff Erickson (University of Illinois)
  • Mark Everingham (Oxford)
  • Rob Fergus (Oxford)
  • Patrick Gallinari (Universite Pierre et Marie Curie)
  • Martial Hebert (Carnegie Mellon University)
  • Bernd Heisele (MIT)
  • Yutaka Hirano (Toyota)
  • Anthony Hoogs (GE)
  • Katsushi Ikeuchi (University of Tokyo)
  • David Kriegman (U.C. San Diego)
  • Sanjiv Kumar (Carnegie Mellon University)
  • Svetlana Lazebnik (University of Illinois)
  • Yann LeCun (NYU)
  • David Lowe (UBC)
  • Jiri Matas (Czech Technical University)
  • Steve Maybank (University of Reading)
  • Daniel Morris (Northrop Grumman)
  • Joe Mundy (Brown)
  • Kevin Murphy (MIT)
  • Jean Ponce (University of Illinois)
  • Jim Rehg (Georgia Tech)
  • Henri Sanson (France Telecom)
  • Cordelia Schmid (INRIA)
  • Bernhard Schoelkopf (MPI for Biological Cybernetics)
  • Antonio Torralba (MIT)
  • Bill Triggs (CNRS)
  • Akihiro Tsukada (Toyota)
  • Tinne Tuytelaars (KU Leuven)
  • Shimon Ullman  (Weizmann Institute)
  • Luc Van Gool (Catholic University of Leuven)
  • Carola Wenk (University of Arizona)
  • Chris Williams (University of Edinburgh)
  • Song-Chun Zhu (UCLA)
  • Andrew Zisserman (Oxford)

  •  
    Sponsors
     

    Abstracts abstracts.pdf  

    Final NSF Report wreport.pdf  

     

    Program
     

    Sunday 7 Sept.

    7:00-8:30
                             Reception                         

     

    Monday 8 Sept.

    8:15-9:00
    Breakfast
    9:00-9:15
     Introduction
    Jean Ponce
    9:15-10:00
      Learning objects and parts in images 
    Chris Williams
    10:00-10:45
    Kernel methods and dimensionality reduction 
    Bernhard Schoelkopf 
    10:45-11:00
    Break
    11:00-11:30
    Object recognition as multimedia translation and data mining
    Kobus Barnard 
    11:30-12:00
      Component-based object recognition
    Bernd Heisele 
    12:00-12:30
    Using the forest to see the trees: a graphical model relating features, objects and scenes 
    Kevin Murphy
    12:30-4:00
    Lunch/Break
    4:00-4:45
    Invariant recognition of generic objects from shape
    Yann LeCun
    4:45-5:15
    Learning a rare event detection cascade by direct feature selection
    Jim Rehg 
    5:15-5:30
    Break
    5:30-6:15
    Machine learning in text information retrieval
    Patrick Gallinari
    6:15-7:15
    Panel I: Learning - A.Zisserman
    C. Bishop , Y. LeCun, M. Hebert,
    B. Schoelkopf, B. Triggs

    Tuesday 9 Sept.

    8:15-9:00
    Breakfast
    9:00-9:45
    Approximate nearest neighbor search
    Jeff Erickson
    9:45-10:15
    Geometric algorithms for biomedical applications 
    Carola Wenk 
    10:15-10:45
    Object recognition in the geometric era: a retrospective
    Joe Mundy
    10:45-11:00
    Break
    11:00-11:30
    Object class recognition by unsupervised scale-invariant learning 
    Rob Fergus
    11:30-12:00
      Fragment-based recognition and segmentation 
    Shimon Ullman 
    12:00-12:30
     Qualitative shape matching for object and action recognition  
    Stefan Carlsson 
    12:30-4:00
    Lunch/Break
    4:00-4:30
    Object recognition at GE
    Anthony Hoogs
    4:30-4:45
    France Telecom's expectation and research in object recognition
    Henri Sanson 
    4:45-5:00
    3D object recognition at Toyota
    Yutaka Hirano
    5:00-5:15
    Break
    5:15-6:15
    Panel II: Categories - C. Schmid
    A. Blake, S. Carlsson, J. Mundy,
    J. Ponce, S. Ullman
    6:15-8:30
    Poster & Demo Session
    8:30
    Workshop dinner

    Wednesday 10 Sept.

    8:15-9:00
    Breakfast
    9:00-10:45
    Invariant local features for recognition
    David Lowe & Cordelia Schmid & Jiri Matas & Tinne Tuytelaars
    10:45-11:00
    Break
    11:00-11:30
    Texture recognition using affine-invariant regions
    Svetlana Lazebnik
    11:30-12:00
    Exploring images for object recognition
    Luc Van Gool 
    12:00-12:30
    Recognition by parts
    Martial Hebert
    12:30-4:00
      Lunch/Break
    4:00-4:30
    Application of Fisher information to line detection 
    Steve Maybank 
    4:30-5:00
      Illumination and Reflectance Modelling, and its application to face recognition 
    David Kriegman 
    5:00-5:15
    Break
    5:15-5:45
     Markov chain method in visual computing
    Song-Chun Zhu
    5:45-6:15
     Learning from observation: object and task recognition for programing by demonstration
    Katsushi Ikeuchi
    6:15-6:30
     Conclusions

     
     
    Posters

    Advances in Component Based Face Detection
    Stab Bileschi

    Shape warping and statistics
    Guillaume Charpiat

    Scene categorization by learning repeated elements
    Mark Everingham

    SVM-based nonparametric discriminant analysis, an application to face detection
    Rik Fransens, Tinne Tuytelaars, Luc Van Gool

    Discriminative random fields for modeling spatial dependencies in images
    Sanjiv Kumar and Martial Hebert

    3D object modeling and recognition using affine-invariant patches and multi-view spatial constraints
    Fred Rothganger, Svetlana Lazebnik, Cordelia Schmid, and Jean Ponce


     
     

    Demos

    Wide-baseline matching and object recognition using extremal regions
    Jiri Matas

    Video google: a text retrieval approach to object matching in videos
    Josef Sivic and Andrew Zisserman