CAP6411 - Computer Vision Systems - Fall 2010

T, R 03:00 PM - 04:15 PM - HEC 103

Instructor: Professor Mubarak Shah, Office 247 HEC, shah@eecs.ucf.edu

Office Hours: Tuesdays 2:00 to 3:00 PM, Thursdays 4:30 PM to 5:30 PM, and by appointment.

GTA: Imran Saleemi, Office 254 HEC, imran@knights.ucf.edu

The University Golden Rules will be observed in this class. Copying or Plagiarism is violation of the Golden Rules.

Course Goals: To prepare students for graduate research in computer vision.

Course Description: To cover in depth the fundamental concepts useful for Computer Vision Systems.

Course Prerequisites: CAP5415 or consent of instructor

Book: Computer Vision: Algorithms and Applications by Richard Szeliski, Microsoft Research:
(http://szeliski.org/Book/drafts/SzeliskiBook_20100805_draft.pdf)

Syllabus: Chapter 1, Chapter 4 (4.1), Chapter 6 (6.1, 6.2), Chapter 8 (8.1—8.5), Chapter 9 (9.1—9.3), Chapter 14 (14.1—14.6), Appendix A and Appendix B.

Grading:
Homework 10%
Programs / Projects 70%
Final Exam (comprehensive) 20%

Programming:
Programming will be in MATLAB.
You are not supposed to use MATLAB code from the web, written by someone else.
Everything has to be written by yourself except the standard Matlab functions.
There will be a tutorial on MATLAB on Thursday, August 26 in the class.

Assignments
Implementation of SIFT key point detector Due date September 27
Deliverables: Source code, Project report discussing problems encountered, where the algorithm fails, Images of intermediate steps. Classroom demo is required on an unknown set of images. Program should display computed features. A comparison with the author’s program (only executable) should show if your algorithm fails. Look at the IJCV paper by D. Lowe.

Exercise 8.1, including additional ideas, variants, and questions due October 26.

Adaboost and PCA for face detection (due December 7)
Download one or more of the labeled face detection databases in Table 14.2.
Generate your own negative examples by finding photographs that do not contain any people.
Implement following face detectors:
(a) boosting (Algorithm 14.1) based on simple area features, with an optional cascade of detectors (Viola and Jones 2004);
(b) PCA face subspace (Moghaddam and Pentland 1997);

Lecture Notes

Camera Model and Pose Estimation

Lectures 9/21 and 9/23 - Lecture Notes

CAP6411 - Computer Vision Systems - Fall 2010

Lecture Notes

Camera Model and Pose Estimation

Optical Flow and Pyramids

Global Flow and Levenberg–Marquardt

Motion Layers Extraction

Face Recognition

Kinematic Features