Summary

• Semester: 1
• Points: 6
• Formal Contact: 4 hours per week
• Total Workload: 10 hours per week

The aim of this course is to provide an overview of digital image processing and computer vision analysis. Both areas are large enough to justify a course on their own; this 1-semester course covers the fundamentals of image formation, low-level image processing and enhancement in both the spatial and Fourier domains, and gives a flavour of higher level vision tasks by studying some of the Structure-from-X algorithms, such as Structure from Stereo, and Structure from Motion. Finally, we briefy cover some aspects of Shape from Shading and Photometric Stereo.

The course consists of 13 two-hour lectures and a weekly 2 hour laboratory session. The assessment for this course is via a 2 hour written examination at the end of the semester (for 70%) and a portfolio of practical laboratory exercises done throughout the semester (for 30%). Almost all the laboratory work is done using MATLAB.

Unit Coordinator, Lecturer, and Lab Demonstrator

Lectures

With the absence of a suitable text a considerable amount of supplementary material has been prepared for the course over the years by Associate Professor Peter Kovesi and Winthrop Professor Robyn Owens. This material is not intended to replace the notes presented in lectures. Material presented in lectures may include areas not covered by these supplementary notes.

Note: Not all of the links below under Lectures and Laboratories are currently pointing to files that exist. As we progress through the semester, these links will start to point to the right place. It is your responsibility to constantly check this website to download the supplementary notes yourself.

Note: Lecture notes can be purchased from the University Bookshop in week 1 or 2 of the semester.

Lecture note coverpage ( pdf file )

1. Introduction, image formation, camera models and perspective geometry.   ( pdf file )
2. Binary images: thresholding, moments, topology.   ( pdf file )
3. Connected components, and morphological operations.   ( pdf file )
   Sample Matlab code
4. Fourier transform theory   ( pdf file )
   Supplementary materials: DFT note, test.m
   Matlab demo.
   Web demo: demo1.
5. Image enhancement   ( pdf file )
   Convolution and Correlation   ( pdf file )
   Some useful demos on the web: demo1, demo2.
6. Edge detection: classical methods   ( pdf file )
7. Camera calibration   ( pdf file )
   
8. Projective Geometry: Single View Reconstruction   ( pdf file )
9. Stereo and structured light   ( pdf file )
   Self-calibration and fundamental matrix   ( pdf file )
10. Motion   ( pdf file )
11. (optional) Tracking and the Kalman Filter  ( pdf file )
    
12. Colour   ( pdf file )

The lecture slides for each of the above topics can be found in the /cslinux/examples/CITS4240/LectureSlides2011 directory. The easiest way to access them is via a Linux machine in a CSSE Laboratory.

Lectopia (recording of lectures)

Laboratories and Portfolio

More instructions (if needed) will be announced later in the semester.

Lab classes will start in week 3 and finish in week 12 of the semester.

• Lab Sheet 1 Introduction to grey scale manipulation.
• Lab Sheet 2 Binary images and object moments.
• Lab Sheet 3 Morphology.
• Lab Sheet 4 The Fourier Transform and the convolution theorem.
• Lab Sheet 5 Constructing filters in the frequency domain.
• Lab Sheet 6 Edge detection and the Hough Transform.
  Note that information in the lab sheet on finding local maxima in the Hough transform has been updated.
• Lab Sheet 7 Camera calibration
  
• Lab Sheet 8 Stereo reconstruction - The last lab sheet!

How to construct your portfolio. This document provides an introduction to HTML and also explains how to save and resize images and plots from MATLAB for incorporation in your portfolio.

Help

• We want you to help each other (within reason!) Post questions, look for answers and respond to other people's questions at help4240 (Discussion Forum)   If you have a problem please look here!
  Please only email the lecturers regarding issues that are personal to yourself.
• How To... solve some common problems.
• Frequently Asked Questions.
• An introduction to Linux
• MATLAB lecture notes for Computing for Engineers and Scientists, our 1st year MATLAB unit. Lectures 4 to 13 will cover all you need to know.

Some useful MATLAB references

• Some MATLAB functions which you may find useful in doing your labs.
• An Introduction to Matlab (version 2.3) by David F. Griffiths from University of Dundee.
• Elements of Matlab by Lloyd D. Fosdick, Elizabeth R. Jessup, and Carolyn J. C. Schauble.

Images

• Class of 2007 A collection of images will be built up in /cslinux/examples/CITS4240/ClassOf2007 - get your image taken at lab sessions. If you don't like your image we can take more until you are happy!. If you want your image removed from this collection email me specifying the image name.
• Class of 2006
• Class of 2005
• Class of 2004
• Class of 2003
• The Images Directory contains some images that will be used in labs.

Past Exams

• The 2003 exam paper (pdf)
• The 2002 exam paper (pdf)
• The 2001 exam paper (pdf)
• The 2000 exam paper (pdf)
• The 1999 exam paper (pdf)
• The 1998 exam paper (pdf)

Texts

• Image Processing, Analysis, and Machine Vision by Sonka, Hlavac, and Boyle (ITP, 1999) MPSL 006.37 1999 IMG.
• Introductory Techniques for 3-D Computer Vision by Emanuele Trucco and Alessandro Verri (Prentice Hall 1998) MPSL 006.37 1998 INT
• Machine Vision: Theory, Algorithms, Practicalities by E. R. Davies (Morgan Kaufmann Publications 2005)
• Machine Vision by Ramesh Jain, Rangachar Kasturi and Brian Schunk (McGraw Hill 1995). MPSL 006.42 1995 MAC
• Digital Image Processing by Rafael C Gonzalez and Richard E Woods (Addison Wesley, 1992) MPSL 621.367 1992 DIG.
• Digital Image Processing using MATLAB by Rafael C Gonzalez, Richard E Woods, and Steven L. Eddins (Pearson Prentice Hall, c2004) MPSL 621.367 2004 DIG.
• Multiple view geometry in computer vision by Richard Hartley and Andrew Zisserman (Cambridge University Press, 2004, 2nd Ed)
• A Guided Tour of Computer Vision by Vishvjit S. Nalwa (Addison Wesley, 1993) MPSL 006.42 1993 GUI.
• Robot Vision by Berthold Klaus Paul Horn (MIT Press, 1986) MPSL 629.892 1986 ROB.
• The Reconfigured Eye by William J. Mitchell (MIT Press, 1992) EDFAA 621.367 1992 REC.

Some useful links

Links marked with a * are outside PARNET. You will need to use the proxy server to access these sites.

• MATLAB links
  • MATLAB Functions for Computer Vision and Image Analysis which you may find useful.
  • * MathWorks' links to sites containing MATLAB vision functions
  • Course notes for Computing for Engineers and Scientists, our 1st year MATLAB unit.
  • MATLAB Programming Style Guidelines by Richard Johnson
  • *CMU's MATLAB page
  • *Octave GNU's MATLAB equivalent.
  • *Scilab is another MATLAB equivalant. It seems to be under more active development than Octave.
• Vision reference material
  • A guide to the vision literature in the libraries at UWA.
  • *The Carnegie Mellon Computer Vision Homepage
  • *CVonline An on-line compendium of computer vision maintained at the University of Edinburgh.
  • *CVED the Computer Vision Education Digital Library Collection hosted at Swarthmore College, Pennsylvania, USA.
  • *The University of Karlsruhe Bibliographies on Computer Graphics and Vision
  • * Information on image formats. See also http://www.sharefile.com/content/image-formats.aspx
  • *A Stereogram Page
  • *The Lena story
  • *Webvision A superb web book describing all aspects of the human eye.
  • *Michael Bach's Optical Illusions and Visual Phenomena. One of the most comprehensive, and best referenced, sites of this kind I have found.
• Mathematical reference material
  • *Eric Weisstein's World of Mathematics - an astonishing on-line mathematical encyclopedia.
  • *Numerical Recipes On-Line
• Software and commercial sites
  • *Adept Electronic Solutions is a local machine vision company. They also award the student prize for Computer Vision CITS4240.
  • *DDD Dynamic Digital Depth is a local company that specialises in converting 2D movies into 3D, and developing various tools for facilitating this process.
  • *The Official XV Home Page
  • *Intel's OpenCV library, a very comprehensive open source vision library of C functions.
  • *Vision 1 a commercial site with some useful info