Courses given by Roland Göcke

ANU ENGN8530 - Computer Vision and Image Understanding: Theories and Research

6 Units, 1st Semester 2007.
10.00 am - 12.30 pm, Wed and Fri, in the Ian Ross Graduate Teaching Room.
Lecturers: Chunhua Shen (Course Convenor) and Roland Göcke
Chunhua's Course Webpage (main course webpage!)
ANU Course Webpage (general description)
ANU WebCT Course Webpage (access only for enrolled students)

Overview
This course is designed for the Master of Information Communications Technology (MICT) program, which is a PhD pathway for excellent students who are interested in understaking a PhD in ICT. NICTA-ANU PhD students will be enrolled to MICT at their first year.

This course provides an overview of geometric, statistical and morphological methods in computer vision and image understanding. It aims at covering the fundamental principles of image processing, multi-view geometry and probabilistic techniques as related to applications in the scope of robotics and machine vision by introducing the student to classical problems found in the literature, such as segmentation and grouping, classification and recognition.

Requirements
There are no rigid prerequisites to participate in this course. Previous experience in computer vision, machine learning, image processing is desirable. Contact us if you have any concerns about whether or not you should take this course.

Syllabus (In no particular order. Slides will be available here. Copyright Chunhua Shen and Roland Goecke, unless otherwise specified.)

21 Feb 2007
Course info Slides
Introduction to computer vision and image understanding Slides
23 Feb 2007
Image processing and analysis basics: Slides
Image representation, image filtering, colour spaces, image statistics, etc.
28 Feb & 2 Mar 2007
Image Matching and Registration Slides
Reading Assignment - Matching and Registration - Review due 7 Mar 2007
Distinctive Image Features from Scale-Invariant Keypoints. David G. Lowe (2004). Download
Efficient Maximally Stable Extremal Region (MSER) Tracking. M. Donoser and H. Bischof (2006). Download
A Performance Evaluation of Local Descriptors. Krystian Mikolajczyk and Cordelia Schmid (2005). Download
Matching with PROSAC - progressive sample consensus. O. Chum and J. Matas (2005). Download
7 Mar 2007
Bayesian analysis for object recognition / detection Slides
Reading Assignment - Bayesian analysis, SVM - Review due 16 Mar 2007
Face Recognition: A Literature Survey. W. Zhao, R. Chellappa, A. Rosenfeld, P.J. Phillips (2003). Download
Face Recognition in Subspaces, Handbook of Face Recognition. G. Shakhnarovich and B. Moghaddam (2004). Download
A Tutorial on Support Vector Machines for Pattern Recognition. C. Burges (1998). Download
9 & 14 Mar 2007
Image and video segmentation:
Clustering, normalised cuts, mean shift algorithm, etc. Slides Slides
Reading Assignment - Clustering and Segmentation - Review due 23 Mar 2007
Normalized Cuts and Image Segmentation. J. Shi and J. Malik (1997). Download
Mean Shift: A Robust Approach toward Feature Space Analysis. D. Comaniciu and P. Meer (2002). Download
Kernel k-means, Spectral Clustering and Normalized Cuts. I. Dhillon, Y. Guan, B. Kulis (2004). Download
16 Mar 2007
Eigen-analysis for object recognition / detection:
Eigen-decomposition, "Eigenfaces", "Eigenlips", spectral analysis. Slides
21 Mar 2007
Object Tracking:
Kalman filter, particle filter, mean shift tracking. Slides
Reading Assignment - Object Tracking - Review due 30 Mar 2007
Color-Based Probabilistic Tracking. P. Perez, C. Hue, J. Vermaak, M. Gangnet (2002). Download
Kernel-based Object Tracking. D. Comaniciu, V. Ramesh and P. Meer (2003). Download
Rapid Object Detection Using a Boosted Cascade of Simple Features. P. Viola nad M. Jones (2001). Download
23 & 28 Mar 2007
Multiview geometry (Guest lecturer: Prof Richard Hartley)
30 Mar 2007
Physics based vision (Guest lecturer: Dr Robby Tan) Slides
4 Apr 2007
Object recognition: an overview.
6 June 2007
Project presentations (final project assignment)

Grading
Grading will follow the grading scheme of High Distinction, Distinction, Credit, Pass, and Fail.

4x Paper reviews during the course, each accounting for 10% of the final mark
No final exam, but
Final project (report, implementation codes, and presentations), accounting for 60% of the final mark

Assignments
Students will be asked to read 2~3 papers for each class/fortnight. They will be required to write a review of 500-800 words of one of these papers, due one week after handout, i.e. before we discuss these papers. The papers are listed above in the week that the assignment is due.
As part of this course, students will complete research-oriented projects. Project proposals will be due in the middle of the term, with projects (report, implementation) due on the 30 April 2007. We will reserve time to present and discuss each project in the class after this date.
The final project can be either of the following:

An original implementation of a new idea or recently published TPAMI/IJCV/ICCV/CVPR/ECCV/SIGGRAPH paper;
A detailed empirical evaluation of an existing implementation of one or more methods;
A paper comparing three or more papers not covered in class, or surveying recent literature in a particular area;
A project proposal not longer than 2 pages must be submitted and approved in the middle of the semester.

Textbook
You might find this book is useful (but not required):
D. Forsyth and J. Ponce. Computer Vision: A Modern Approach. Prentice Hall, 2002.
For further reading:
R. Hartley, A. Zisserman. Multiple View Geometry in Computer Vision. 2nd Edition, Cambridge University Press, 2004. and
J. C. Russ. The Image Processing Handbook. 4th Edition, CRC, 2002.

Resources online

Back to Homepage