CS 281b/ECE 281b: Advanced Topics in Computer Vision
Instructor
B. S. Manjunath, manj at ece. Rm 3157, Engineering I. Tel: 805.893.7112
Office hours: M 11am-12 noon, W 10-11am, or by appointment (or drop by and see if I'm available)
Meeting Times and Locations
Lecture: Tues/Thurs 2:00pm - 3:50pm, PHELPS 1431
About the course:
This is an advanced graduate level course in computer vision. If you are new to computer vision, you may want to attend my morning lectures on computer vision for the undergraduate students (CS/ECE 181b). During this quarter, we will explore the following topics:
Discussions will be based on selected articles (recently published journal and conference articles). Grading will be based on class participation (10%), paper presentations (40%), final project (report+presentation) 50%. Students are encouraged to come up with project ideas that go beyond a simple implementation of an existing method. Evaluation of the projects will be based on novelty and final results.
READERS: Will present a brief report (10 mins) at the next class meeting and a brief written critique (suggested length: 3 pages) that will be posted on the web.
_____________
| Date | Paper | Presenters | Readers/Critique |
|---|---|---|---|
| April 13, 2006 | Baker & Matthew, Lucs-Kanade 20 years on: A Unifying framework | Marco Zuliani | |
| April 18, 2006 | Geman & Geman, Stochastic Relaxation, Gibbs Distributions and the Bayesian Restoration of Images. | ||
| April 20, 2006 | Segmentation: Overview | ||
| April 25, 2006 | Segmentation (contd | ||
| April 27, 2006 | Spectral methods for Image Segmentation (3 papers: Normalized Cuts, segmentation using eigenvectors, factorization approach to grouping ) |
Luca Bertelli | Ibrahim Khalil and Mike Quinn, critique |
| April 28, 2006 (FRIDAY CLASS, PHELPS 1425, 330PM) |
Illusory/Salient Contour Shapes (3 papers: Segmentation of multiple salient closed contours, a model for illusory contours, emergent properties of illusory contour shape.) | Nhat Vu | Thomas Kuo and Ibrahim Khalil, critique |
| May 02, 2006 | Markerless tracking | Taehee Lee | Chris Coffin and Luca Bertelli critique |
| May 02, 2006 | Factorization method for computing structure from motion | Jonathon Ventura | Mike Quinn and Nhat Vu critique |
| May 04, 2006 | Scale Invariant Features | Mei-Chen Yeh | Juriz Guzman and Jim Kleban |
| May 04, 2006 | Efficient matching with local features | Emily Moxley | Jon Ventura and Nhat Vu |
| May 05, 2006 (FRIDAY) | flexible method for camera calibration, aligning sequences from multiple cameras, Collaborative approach to sensor calibration |
Michael Quinn | Taehee Lee and Anindya Sarkar critique |
| May 09, 2006 | Active contours without edges | Mei-Hsuan Wu | Jon Ventura and Joriz Guzman critique |
| MAY 09, 2006 | Matching widely separated views | Jim Kleban | |
| May 11, 2006 | Appearence cloning (IJCV) | Ariane Gravel | Taehee Lee and Thomas Kuo Critique |
| May 11, 2006 | Video Epitomes (CVPR'05) | Chris Coffin | Taehee Lee and Thomas Kuo |
| May 12, 2006 | RANSAC | Dr. Charles Kenney | |
| May 18, 2006 | 3-D Object modeling (IJCV'06) | Thomas Kuo | Ariane Gravel and Amir Rahimi critique |
| May 18, 2006 | 3-D Projective reconstructions Provably convergent iterative methods, Self calibration of cameras, Recovering 3-D structure and motion |
Ibrahim Khalil | Chris Coffin and Emily Moxley |
| May 19, 2006 | GUEST LECTURE: Level sets in Computer vision | Professor Frederic Gibou | |
| May 23, 2006 |
Rapid object detection using a boosted cascade of simple features, |
Anindya Sarkar and Amir R | Mei-Hsuan Wu and Jim Kleban |
| May 23, 2006 | Multiscale segmentation | Joriz Guzman | Ariane Gravel and Mei-Hsuan Wu critique |
| May 25, 2006 | GUEST LECTURE: Visual control of action: Importance of optic flow and 3-D representation |
Professor Jack Loomis, Psychology Department, UCSB | |
| June 1, 2006 | MPEG-7 (5MB, PPT file) | Powerpoint presentation |
Anindya Sarkar and Amir Rehami: " Object class recognition by unsupervised scale-invariant learning," by Fergus, Perona and Zisserman, CVPR 2003. Objective: to implement and evaluate the basic method(s) described in the paper. Test data: obtained from the web.
Emily Moxley: Improved homogeneous texture descriptor
Mei-Chen Yeh: Face recognition
Taehee Lee: 3D reconstruction from videos for augmented reality applications.
Joriz Guzman and Jim Kleban: Automatic stereo background subtraction for interactive applications.
Michael Quinn: Calibration issues in a distributed camera networks
Luca Bartelli and Ibrahim Khalil: 3D reconstruction
Thomas Kuo: EMD for similarity search and retrieval evaluations
Nhat Vu: Salient contour extraction, Segmentation of multiple salient closed contours.
Ariane Gravel and Christopher Coffin, Camera pose estimation and object reconstruction.
Jon Ventura, Virtual walk-through using video