The last decade has witnessed rapid progress in computer vision and statistical methods are successfully applied to all aspects of vision. These advances, combined with increasing computer power, make us optimistic about tackling difficult and exciting problems such as building general purpose visual inference machines for segmentation, recognition, and tracking.

This workshop will provide an international forum for discussing recent advances in the statistical and computational theories of vision and for studying solid theoretic foundation for integrated vision systems. The workshop is aimed at brings together highly respected researchers from computer vision, mathematics, statistics, information theory, and psychology to address key issues in Bayesian image analysis. Original papers are solicited on topics including, but are not limited to:

•  Advanced theories for statistical modeling of realistic visual patterns,
•  Advanced theories for visual learning from observed visual patterns,
•  Advanced techniques for efficient computation and search,
•  Advanced techniques for Markov chain Monte Carlo Sampling.
•  Advanced theory for performance bounds on the accuracy of models and speed of algorithms.
• Relations of the above theories to biologic vision at the neural, psychophysical, and cognitive levels.
• Applications and Systems for image segmentation, perceptual organization, object recognition, tracking, motion analysis.

9:00AM-10:30AM Session 1: Statistics of Natural Imagery and Robust Estimation
(session chair: Alan Yuille)

• Exploiting Spatial and Spectral Image Regularities for Color Constancy ---- B. Singh, W.T. Freeman, and D. Brainard
• Whitening for Photometric Comparison of Smooth Surfaces under Varying Illumination---- M. Osadchy, M. Lindenbaum, and D. Jacobs
• On Incremental and Robust Subspace learning---- Y. Li, L-Q Xu, J. Morphett and R. Jacobs

10:30AM-10:45AM Coffee Break

10:45AM-12:15PM Session 2: Statistical Models for Motion
(session chair: Song-Chun Zhu)

• Adaptive Estimation of Motion Using Generalized Cross Validation---- H. Foroosh
• Deterministic and Statistical Properties of Multi-Resolution 3D Modeling---- A. RoyChowdhury, H. Liu, and R. Chellappa
• An Information Fusion Framework for Robust Shape Tracking---- X.S. Zhou, D. Comaniciu, and S. Krishnan

12:15PM-2:30PM Lunch Break

2:30PM-4:15PM Session 3: Poster Session

• Bayesian Articulated tracking Using Single Frame Pose Sampling-- L. Taycher and T. Darrell
• How are Statistical Methods for Geometric Inference Justified?-- K. Kanatani
• A Model-Based Range Image Segmentation Algorithm Using a Novel Robust Estimator-- H. Wang and D. Suter
• Kurtosis and the Phase Structure of Images-- J. Gluckman
• A Probabilistic Network Model for Integrating Visual Cues and Inferring Intermediate-level Representations-- K. Baek and P. Sajda
• Uncertainty in 3D Shape Estimation-- H. Ji and C. Fermuller
• A Particle Filtering Approach to Abnormality Detection in Nonlinear Systems and its Application to Abnormal Activity Detection-- N. Vaswani and R. Chellappa
• How Heuristics Expedite Markov Chain Search?-- R. Maciuca and S.C. Zhu
• Efficient Graph Cuts for Unsupervised Image Segmentation using Probabilistic Sampling and SVD-based Approximation-- J. Keuchel and C. Schnörr
• Parameter Estimation from Optical Flow without Nuisance Parameters-- N. Ohta
• Surround Statistics and the Perception of Intensity and Color-- J.J. Clark and D. Hernadez
• Unsupervised Clustering of Images using their Joint Segmentation-- Y. Seldin, S. Starik, and M. Werman
• Exploiting the Sparse Derivative Prior for Super-Resolution and Image Demosaicing-- M.F. Tappen, B.C. Russell, and W.T. Freeman
• The Kalman-EM Contour Tracker-- A.E.C. Pece

4:15PM-4:30PM Coffee Break

4:30PM-6:00PM Session 4: Discriminative Methods and Algorithms 
(session chair: Bill Freeman)

• Scale-Invariance of Support Vector Machines based on the Triangular Kernel---- F. Fleuret and H. Sahbi
• Fast Color Image Retrieval via Embeddings---- P. Indyk and N. Thaper
• Learning a Rare Event Detection Cascade by Direct Feature Selection---- J. Wu, J.M. Rehg, M.D. Mullin

Alan Yuille www.stat.ucla.edu/~yuille , Song Chun Zhu www.stat.ucla.edu/~sczhu, Zhuowen Tu www.stat.ucla.edu/~ztu

Ted Adelson, www.ai.mit.edu/people/adelson
Michael Black, www.cs.brown.edu/people/black/
Andrew Blake, research.microsoft.com/~ablake/
Rama Chellappa, www.cfar.umd.edu/~rama/main.html

Frank Dellaert, www.cc.gatech.edu/~dellaert/
David Forsyth, www.cs.berkeley.edu/~daf/
Bill Freeman, www.ai.mit.edu/people/wtf/
Davi Geiger, www.cs.nyu.edu/cs/faculty/geiger/home.html
Alan Jepson, www.cs.toronto.edu/~jepson/
Daniel Kersten, gandalf.psych.umn.edu/~kersten/ kersten-lab/kersten-lab.html
Jitendra Malik, www.cs.berkeley.edu/~malik/
Lionel Moisan, www.cmla.ens-cachan.fr/Utilisateurs/moisan/
Olivier Faugeras, www-sop.inria.fr/robotvis/personnel/ faugeras/faugeras-eng.html

James Rehg, www.cc.gatech.edu/~rehg/
Harry Shum, research.microsoft.com/users/hshum/
Eero Simoncelli, www.cns.nyu.edu/~eero/
Stefano Soatto, www.cs.ucla.edu/~soatto/

Phil Torr, research.microsoft.com/~philtorr/
Yair Weiss, www.cs.huji.ac.il/~yweiss/
Ying-Nian Wu, www.stat.ucla.edu/~ywu/
Laurent Younes, www.cmla.ens-cachan.fr/Utilisateurs/younes/

•  Final version of accepted paper. 20/September/2003
•  Workshop 12/October/2003