Poll

The ITinCVPR book ...
 
Advanced ITinCVPR in a Nutshell (Tutorial) Print
Wednesday, 27 January 2010 11:22

Tutorial: Advanced ITinCVPR in a Nutshell

 

Francisco Escolano
PhD, Associate Professor
Department of Computer Science and Artificial Intelligence,
University of Alicante (Spain)
This e-mail address is being protected from spambots. You need JavaScript enabled to view it

Anand Rangarajan
PhD, Associate Professor
Department of Computer & Information Science and Engineering,
University of Florida, Gainesville
This e-mail address is being protected from spambots. You need JavaScript enabled to view it

 

COURSE SLIDES

1. Introduction

2. Interest Points

3. Feature Selection

4. Isocontours, Registration

Additional Slides (pdf, pptx)

5. Shape, Matching, Divergences

6. Mixtures

7. Future Trends (pdf, pptx)

COURSE DESCRIPTION

ADVANCED INFORMATION THEORY IN COMPUTER VISION AND PATTERN RECOGNITION  IN A  NUTSHELL

(Advanced ITinCVPR in a Nutshell)
Half-day Tutorial in CVPR'2010
Date to be determined (between June 13 - 18)


Information Theory (IT) plays a key role in formulating and designing algorithmic solutions to many problems in computer vision and pattern recognition (CVPR): image matching, clustering and segmentation, salient point detection, feature selection and dimensionality reduction, projection pursuit, optimal classifier design, and many others. Nowadays, researchers are routinely bringing IT elements to the CVPR arena. Among  these elements are "measures" (entropy, mutual information, Kullback-Leibler and Jensen-Shannon divergence, Bregman divergence etc.), "principles" (maximum entropy, minimax entropy, minimum description length etc.) and "theories" (rate distortion theory, coding, the method of types etc.). Recently,  alternative definitions for the latter measures and new methods of estimation which bypass the often cumbersome process of pdf estimation have been proposed. The tutorial will encompass (a) practical solutions for feature selection in very high dimensional patterns and images, (b) reformulation of shape  matching problems using information-theoretic tools, (c) more efficient and faster clustering in high dimensions, (d) bypassing the estimation of scale-saliency points at many locations in images, (e) better methods for image registration, and many more.

This tutorial addresses the need for a unified presentation of the application of Information Theory in CVPR to the attendants. This is done through a novel perspective (exploring measures, principles, theories and estimators in key problems in Computer Vision). We will concentrate on several topics in order to give good examples of the application of this new perspective. Advanced theoretical insights will be complemented and illustrated by applications. Also software/implementations will be available in the webpage for evaluation and testing of the main presented algorithms. Besides the slides of the tutorial, the slides of the book "Information Theory in CVPR" (http://www.rvg.ua.es/ITinCVPR) will be available for the students. Finally, we will also present some open questions and hot topics for current and future research. Our goal is to present information-theoretic tools to a growing CVPR field in a coherent way, by first introducing the main theory and then providing algorithms and applications.

TOPICS


1. Introduction: ITinCVPR (measures, principles, theories and estimators)
2. Interest Points and the Method of Types: Application to Visual Localization.
3. Information Theoretic Image Matching with Isocontours: Applications to Medical Imaging.
4. Shape Registration with I-Divergences: Applications to 2D and 3D Registration.
5. Fast Gaussian Mixtures-based Clustering with Bypass Entropy Estimators and MDL: Application to Color Image Segmentation.
6. Feature Selection in High-Dimensional Patterns: Applications to both Images and MicroArray Data Classification.

 

RELEVANT PUBLICATIONS

 

BOOKS:


F. Escolano, P. Suau, B. Bonev. "Information Theory in Computer Vision and Pattern Recognition", Springer-London, July 2009,
ISBN: 978-1-84882-296-2. (2009)

PAPERS:


A. Peñalver, F. Escolano, J.M. Sáez. Learning Gaussian Mixture Models With Entropy-Based Criteria.
IEEE Trans. Neural Networks 20(11) 1756-1771  (2009)

P. Suau, F. Escolano. Bayesian Optimization of the Scale Saliency Filter.
Image and Vision Computing, 26(9), pp. 1207-1218 (2008)

Fei Wang, Baba C. Vemuri, Anand Rangarajan, Stephan J. Eisenschenk:
Simultaneous Nonrigid Registration of Multiple Point Sets and Atlas
Construction. IEEE Trans. Pattern Anal. Mach. Intell. 30(11): 2011-2022 (2008)

Ajit Rajwade, Arunava Banerjee, Anand Rangarajan: Probability Density
Estimation Using Isocontours and Isosurfaces: Applications to
Information-Theoretic Image Registration. IEEE Trans. Pattern
Anal. Mach. Intell. 31(3): 475-491 (2009)

SHORT BIOGRAPHIES OF THE INSTRUCTORS

Francisco Escolano

obtained his Bachelors Degree in Computer Science at the Polytechnical University of Valencia (Spain) in 1992 and his Ph Degree in Computer Science at the University of Alicante in 1997. Since 1998, he has been an Associate Professor in the Computer Science and Artificial Intelligence Department at the University of Alicante. He has been post-doctoral fellow with Dr. Norberto M. Grzywacz at the Biomedical Engineering Department at the University of South California in Los Angeles, and he has also collaborated with Dr. Alan L. Yuille at the Smith-Kettlewell Eye Research Institute of San Francisco. He also visited the Liisa Holm's Bioinformatics Lab at the University of Helsinki. His research interests are focused on the development of efficient and reliable pattern recognition and computer vision algorithms for biomedical applications, bioinformatics, robotics and applications for the visually impaired. He is the head of the Robot Vision Group. He co-chaired and organized the 6th TC15-IAPR Workshop on Graph-based Representations in Pattern Recognition (Gbr), and co-chaired the 7th Gbr workshop.  He has recently published the book "Information Theory in Computer Vision and Pattern Recognition", Springer-London, July 2009, ISBN: 978-1-84882-296-2.

 

Anand Rangarajan

 

received the B.Tech. degree in electronics engineering from the Indian Institute of Technology, Madras, in 1984, and the Ph.D. degree in electrical engineering from the University of Southern California, Los Angeles, in 1991. From 1990 to 1992, he was a postdoctoral associate in the Departments of Diagnostic Radiology and Computer Science, Yale University, New Haven, CT. From 1992 to 1995, he held a joint research faculty position in both departments. From 1995 to 2000, he was an Assistant Professor in the Image Processing and Analysis Group (IPAG), Departments of Diagnostic Radiology and Electrical Engineering, Yale University. He is now an Associate Professor in the Department of Computer and Information Science and Engineering, University of Florida, Gainesville. His current research interests are best summarized as the application of machine learning to image analysis. He is also interested in the scientific study of consciousness. Dr. Rangarajan has co-chaired EMMCVPR (2003 and 2005) and was an area
chair for the IEEE International Conference on Computer Vision (ICCV) 2007.



Last Updated on Tuesday, 15 June 2010 18:19