Important Dates

Submission deadline: July 1 (5pm your local time), 2005

Notification of acceptance: August 19, 2005

Camera ready: September 1, 2005

Workshop date: October 21, 2005

Accepted Papers Click here for Final Program

  • 30 Layered Dynamic Textures
  • 33 Robust 3D Segmentation of Multiple Moving Objects Under Weak Perspective
  • 37 Force/Vision Based Active Damping Control of Contact Transition in Dynamic Environments
  • 41 A Batch Algorithm For Implicit Non-Rigid Shape and Motion Recovery
  • 43 Tracking of Multiple Objects Using Optical Flow Based Multiscale Elastic Matching
  • 52 Bayesian tracking with auxiliary discrete processes. Application to detection and tracking of objects with occlusions
  • 55 Using a Connected Filter for Structure Estimation in Perspective Systems
  • 57 Online Video Registration of Dynamic Scenes using Frame Prediction
  • 58 Homeomorphic Manifold Analysis: Learning Decomposable Generative Models for Human Motion Analysis
  • 61 A Rao-Blackwellized Parts-Constellation Tracker
  • 64 Estimating the Pose of a 3D Sensor in a Non-Rigid Environment
  • 65 View-invariant modeling and recognition of human actions using grammars
  • 67 A Probabalistic Framework For Correspondence and Egomotion
  • 70 Articulated Motion Segmentation Using RANSAC With Priors
  • 71 Spatial Segmentation of Temporal Texture Using Mixture Linear Models
  • 78 The Space of Multibody Fundamental Matrices: Rank, Geometry and Projection

Paper submission

  • All paper submissions to ICCV Workshop on Dynamical Vision will be electronic. Submit a paper Enter reviews View reviews
  • All papers must be submitted in Acrobat PDF format.
  • All reviewing will be double blind, so the paper must not include any information which allows the authors to be identified. Papers that provide obvious identifying information will be rejected without review. See the Anonymity Guidelines for more details.
  • A complete paper must be submitted in camera-ready format with no more than 8 pages in length. Papers accepted for the conference will be allocated 6 pages in the proceedings, with the option of purchasing up to 2 extra pages. Papers should be submitted in the length and format intended for final publication. Specifically, submitted papers should have a maximum of 8 pages and should adhere to the formatting guidelines below. Papers with more than 8 pages will be rejected without review.
  • The paper format must follow the standard IEEE 2-column format of single-spaced text in 10 point Times Roman (or closely resembling), with 12 point interline space. All printed material, including illustrations, must be kept within a print area of 6-7/8 inches (17.5 cm) wide by 8-7/8 inches (22.5 cm) high. We ask that you do not deviate from these guidelines since this will be a cause for paper rejection without review. A simple LaTex Template may assist you in meeting this format.


Call for Papers

Classical multiple-view geometry studies the reconstruction of a static scene observed by a rigidly moving camerae. Many real-world applications however require the modeling and reconstruction of a scene that has much more complex dynamics. That is, the scene may consist of multiple moving objects (e.g., a traffic scene) or articulate motions (e.g., a walking human) or even non-rigid dynamics (e.g., smoke, fire, waterfall). To study the problems of reconstructing different dynamical scenes, many new algebraic, geometric, statistical, and computational tools have recently emerged in computer vision, computer graphics, image processing, and vision-based control. The goal of this workshop is to converge different aspects of the research on dynamical vision and identify common mathematical problems, models, and methods for future research in this emerging new topic. The workshop welcomes papers that fall into the following categories:


  • Segmentation, Estimation & Tracking of Multiple Rigid-body Motions.
    • Based on Optical Flows or Image Gradients
    • Based on Feature Correspondences
    • Integrated Approaches and Hybrid Motion Models
  • Segmentation, Estimation & Tracking of Articulate and/or Non-Rigid Motions.
    • Human Motions
    • Non-Rigid Motions and Deformations
  • Identification and Recognition of of Dynamical Scenes from Videos.
    • Dynamical Textures
    • Video Segmentation
  • Mathematical Tools for Model, Analysis, and Synthesis of Dynamical Scenes
    • Subspace Methods and High-Order Tensors
    • Dynamical Systems
    • Stochastic Models (e.g. Hidden Markov Models)
  • Applications of Dynamical Vision (e.g., Surveillance, Graphics, and Robotics).
    • Visual Servoing and Vision-Based Control & Navigation
    • Localization and Mapping




Prof. Anders Heyden
Department of Mathematics,

Lund University and Malmo University, Sweden

Prof. Yi Ma
Department of Electrical & Computer Engineering

University of Illinois at Urbana Champaign, USA


Prof. René Vidal
Department of Biomedical Engineering

Johns Hopkins University, USA


Program Committee

Yiannis Aloimonos
Serge Belongie
Noah Cowan
Kostas Daniilidis
Ahmed Elgammal
Ruggero Frezza
Bijoy Ghosh
Greg Hager
Richard Hartley
Kun Huang
Joao Hespanha
Rolf Johansson
Fredrik Kahl
Kenichi Kanatani
Jana Kosecka
Nemanja Petrovic
Marc Pollefeys
Shankar Sastry
Stefano Soatto
Harry Shum
Peter Sturm
Ying Wu
Jie Zhou


University of Maryland at College Park

University of California at San Diego

Johns Hopkins University

University of Pennsylvania

Rutgers University

University of Padova, Italy

Washington University at St. Louis

Johns Hopkins University

Australia National University, Australia

Ohio State University

University of California at Santa Barbara

Lund University, Sweden

Lund University, Sweden

Okayama University, Japan

George Mason University

Siemens Corporate Research, Princeton

University of North Carolina at Chapel Hill

University of California at Berkeley

University of California at Los Angeles

Microsoft Research in Asia, China

INRIA Rhône-Alpes, France

Northwestern University

Tsinghua University, China


Final Program (TBA)


Sponsors (TBA)