2nd Workshop on


Understanding, Modeling, Capture and Animation


Rio de Janeiro, Brazil

October 20, 2007

In Conjunction with ICCV 2007





Workshop program:


8:45-8:50    Opening

8:50-10:00  Oral Session I 


Marker-less 3D Feature Tracking for Mesh-based Human Motion Capture

Edilson de Aguiar, Christian Theobalt, Carsten Stoll and Hans-Peter Seidel


Nonparametric Density Estimation with Adaptive, Anisotropic Kernels for Human Motion Tracking

Thomas Brox, Bodo Rosenhahn, Daniel Cremers, Hans-Peter Seidel


Boosted Multiple Deformable Trees for Parsing Human Poses

Yang Wang and Greg Mori


10:00-11:00  Poster Session I


11:00 12:10 Oral Session II 


Multi-Activity Tracking in LLE Body Pose Space

Tobias Jaeggli, Esther Koller-Meier and Luc Van Gool


Modeling Human Locomotion with Topologically Constrained Latent Variable Models

Raquel Urtasun, David J. Fleet and Neil D. Lawrence


Exploiting Spatio-Temporal Constraints for Robust 2D Pose Tracking

Grgory Rogez, Ignasi Rius, Jesús Martínez-del-Rincón and Carlos Orrite


12:10-1:30 Lunch Break


1:30-2:30 Invited talk Prof. Demetri Terzopoulos (UCLA)


2:30-3:30 Poster Session II


3:30-5:00  Oral Session III 


Semi-Latent Dirichlet Allocation: A Hierarchical Model for Human Action Recognition

Yang Wang, Payam Sabzmeydani, and Greg Mori


Real-time and Markerless 3D Human Motion Capture using Multiple Views

Brice Michoud, Erwan Guillou and Sada Bouakaz


Silhouette Based Generic Model Adaptation for Marker-less Motion Capturing

Martin Sunkel, Bodo Rosenhahn, and Hans-Peter Seidel


Robust Spectral 3D-bodypart Segmentation along Time

Fabio Cuzzolin, Diana Mateus, Edmond Boyer and Radu Horaud


5:00-5:15  Closing Remarks




Poster Session I: Motion Capture, Body Tracking and Segmentation


Gradient-Enhanced Particle Filter for Vision-Based Motion Capture

Daniel Grest and Volker Krger


Effcient Upper Body Pose Estimation from a Single Image or a Sequence

Matheen Siddiqui and Grard Medioni


3D Hand Tracking in a Stochastic Approximation Setting

Desmond Chik, Jochen Trumpf, Nicol N. Schraudolph


Multi Person Tracking within Crowded Scenes

Andrew Gilbert and Richard Bowden


Joint appearance and deformable shape for nonparametric segmentation

S. Boltz, É. Debreuve and M. Barlaud


Articulated Object Registration Using Simulated Physical Force/Moment For 3D Human Motion Tracking

Bingbing Ni, Stefan Winkler, and Ashraf Kassim


An Ease-Of-Use Stereo-based Particle Filter for Tracking Under Occlusion

Ser-Nam Lim and Larry Davis


Poster Session II: Activity Recognition


Recognizing Activities with Multiple Cues

Rahul Biswas, Sebastian Thrun and Kikuo Fujimura


Human Action Recognition Using Distribution of Oriented Rectangular Patches

Nazlı İkizler and Pınar Duygulu


Human Motion Recognition Using Isomap and Dynamic Time Warping

Jaron Blackburn and Eraldo Ribeiro


Behavior Histograms for Action Recognition and Human Detection

Christian Thurau


Learning Actions Using Robust String Kernels

Changjiang Yang, Yanlin Guo, Harpreet Sawhney, and Rakesh Kumar






Ahmed Elgammal, Rutgers University, USA elgammal (at) cs.rutgers.edu

Bodo Rosenhahn, MPI Informatics, Germany rosenhahn (at) mpi-inf.mpg.de

Reinhard Klette, Auckland University, New Zealand r.klette (at) auckland.ac.nz




Paper Submission: June 29th, 2007

Notification: end of July 2007    





The workshop proceeding will be published as volume number 4814 in Springer's Lecture Notes in Computer Science (LNCS) at the time of the workshop.




The 1st issue of this workshop took place in June 2006 at Dagstuhl/Germany, co-chaired by B. Rosenhahn, D. Metaxas, and R. Klette. This 2nd issue will accompany ICCV 2007.


Modeling, tracking and understanding of human motion based on image sequences (such as video) is a field of research of increasing importance,  with applications in sports sciences, medicine, biomechanics, animation (avatars), surveillance, and so forth. Progress in human motion analysis depends on research in computer graphics, computer vision and biomechanics.  Though these fields of research are often treated separately, human motion analysis requires an interaction of computer graphics with computer vision, which also benefits from an understanding of biomechanic constraints.


Eadweard Muybridge (1830-1904) is known as the pioneer in motion capturing with his famous experiments in 1887 called ``Animal Locomotion'' (Do all feet leave the ground during the gallop of a horse? He used photography to answer the question.) The field of animal or human motion analysis has developed into many directions since then. However, human-like animation and recovery of motion is still far from being satisfactory. Various groups are dealing with different aspects of modeling, estimation and animation of human motions. Motivations differ, and define directions of research. Examples of motivations are the analysis of movements for disease detection (hip dislocations, knee injuries etc.),  sports movement optimization (ski or high jumping, golf playing, swimming, etc.), the animation of avatars in movies  (e.g. Gollum in Lord of the Rings), or the realistic character animation in computer games.


The goal of this workshop is to encourage interaction and to post collaboration between researches in computer vision, animation, and biomechanics.  New results and specific research strategies will be discussed at the workshop to approach this highly complex field. The intention is to discuss theoretical fundamentals related to those issues and to specify open problems and major directions of further development in the field of human motion related to computer vision, computer graphics or biomechanics. The workshop encourages interdisciplinary (vision + graphics, biomechanics + vision, etc.) contributions.





Chandrajit Bajaj

Michael Black

Richard Bowden

Thomas Brox

Stefan Carlsson

Rama Chellappa

Larry Davis

Leo Dorst

Pascal Fua

David Fleet

Vaclav Hlavac

Jessica K. Hodgins

Atsushi Imiya

Reinhard Koch

Volker Krueger

Nadia Magnenat-Thalmann

Markus Magnor

Dimitris Metaxas

Meinard Muller

Lars Muendermann

Michael Neff

Ramakant Nevatia

Vladimir Pavlovic

Zoran Popovic

Fatih Porikli

Dimitris Samaras

Stan Sclaroff

Hans-Peter Seidel

Cristian Sminchisescu

Gerald Sommer

Demetri Terzopoulos

Matthias Teschner

Christian Theobalt

Matthew Turk

Thomas Vetter

Jian J. Zhang







Paper  submissions are solicited in the following topics


  • Accurate and anatomically correct modeling of human beings (this includes surface representations, deformable modeling, models of joints, surface morphing, rendering etc.)
  • Image or video processing for human motion (tracking and pose estimation, etc.)
  • Biomechanic constraints during motion modeling and estimation (self occlusion handling, self collision avoidance, joint stiffness modeling, etc.)
  • Sensor fusion for human motion estimation (e.g., using cameras, structured lighting, or a laser range scanner).
  • Learning human motion patterns.


Further subjects are possible as long as in the general field of the workshop.