Computer Vision / Datorseende TS1017

The Computer Vision course covers major background for graduate students in the field. PhD students who have not read our Computer Vision undergraduate courses, or some thing equivalent (see TSBB12 Computer Vision Theory  and TSBB13 Computer Vision Systems for details) are supposed to read this course during their first year of PhD studies.

Usually, PhD students can also participate in the undergraduate courses, but we were forced to cancel the respective courses in 2011. As a replacement, we offer this PhD course covering basically the same contents, including a theoretic part and two small projects.

The course is basically the same as a combination of the undergraduate courses TSBB12 Computer Vision Theory  and TSBB13 Computer Vision Systems.

The course gives 12 ECTS credit points, which corresponds to approximately 320 h of work per student.  This time is divided into the following activities

• Lectures:
• Class room lessons:  
• Computer assignments: 
• Written examination:
• Own work (mainly in projects):

The course is given during the VT1 and VT2 periods.

Messages 

• None

People

• Michael Felsberg, lectures, examiner
• Klas Nordberg, lectures
• Per-Erik Forssén, lectures
• Erik Ringaby, lessons, computer exercises
• Kristoffer Öfjäll, lessons, computer exercises

We all have our offices in the B-building, corridor A between entrances 27 and 29, ground floor.

Documents

• LiTH Study Guide for TS1017, 2011
• Course schedule

Literature 

The course is based on the book

• Sonka, Hlavac, Boyle, Image Processing, Analysis, and Machine Vision, 3rd edition, Thomson (2008).

There is a student (paperback) edition of this book (labeled ISE) from Nelson Engineering.

Registration

In order to receive email about the course and have results input to Ladok it is required that you are registered for the course.  If you intend to take the course but is not registered, make sure to register asap, using the Student portal.  If you take the course but are not registered to any program at Linköping University, please contact the course examiner in order to make sure that you receive email about the course.

Examination

The course has a written examination which is offered at three occations during 2010: 

• March 14, 8-12.  Volontary mid-term examination of the first half of the course.
• TBA
• TBA

Grades

How to pass the course and get grades are described here.

Schedule for the lectures VT1+VT2

• SHB = Sonka, Hlavac and Boyle "Image Processing, Analysis, and Machine Vision"
• GW = Gonzalez and Woods, "Digital Image Processing"
• Wikipedia = Search wikipedia for the listed term
• Mathworld = Search http://mathworld.wolfram.com/ for the listed term
• Edupack - orientation is a tutorial found here
• EDUPACK2 = www2.cvl.isy.liu.se/Education/Edupack/orientation2.pdf
• RNDF = J. Weickert: A Review of Nonlinear Diffusion filtering, Scale-Space 1997

Class room lessons VT1 only

Computer assignments VT1+VT2

Projects

The course includes two projects that are made in groups of 2-3 students

• Object tracking (VT1)
• 3D reconstruction (VT2)

Project groups 2011

General resources

We suggest and allow you to use the following software

• OpenCV (Open Source Computer Vision).  Version 1.0.0 is installed on the department's Linux computers.  This is an older version that differs from more recent versions, 2.x, not only in functionality but also in interfaces to functions and there may be problems in getting OpenCV based code developed for 1.0.0 to work on 2.x versions.  Conseuqently, if you want to use OpenCV, and need to use the department's computers, stick to the 1.0.0 version only.  If you have access to your own computers, we recommend to install 2.x and use that version only.  Read the section on OpenCV in the hints and pittfalls page.  There is a cheat sheet for OpenCV.
• David Lowe's SIFT implementation for Matlab.  This includes binary executables that are compiled into mex-code that runs from Matlab.  If you don't know about Matlab mex-files ask your guide.
• VLFeat has a a useful code library, both for Matlab and C/C++.  For example, there is an alternative implementation of SIFT here, and also an implementation of MSER here.  Both are made by Andrea Vedaldi.
• The Visal Geometry Group at Oxford University maintains code for affine invariant region detectors, produced in cooperation with other groups.
• The Computer Vision Laboratory at ETH provides an implementationof SURF.
• The Sparse Bundle Adjustment package (C/C++ code) by M. Lourakis.

Guides

Each project is assigned a project guide.  They are

• Erik Ringaby, ringaby@isy.liu.se
• Kristoffer Öfjäll (VT1), kristoffer.ofjall@liu.se
• Marcus Wallenberg (VT2), wallenberg@isy.liu.se

Subversion

Each project will have access to a part of a Subversion repository, with individual login by each project member.   User names and password are managed according to (use the first that applies)

• If you already are a user of ISY Subversion, use the same name and login as before
• If you are a ISY emplyee, e.g. as a PhD student, use your ISY account name for logging in.  A password will be created and distributed at the beginning of the first project.
• If you have an active LiU student account, use the account name for logging in.  A password will be created and distributed at the beginning of the first project.
• If none of the above applies, a temporary user name will be crated and distributed at the beginning of the first project.  A password will be created and distributed at the beginning of the first project.

In the case that a new password is created and distributed at the beginning of the first project, this password is only temporary and must be changed to new one before the account is actively used.  This is done by following the instructions here

• https://svn.isy.liu.se

The address to the Subversion repositories are

• https://svn.isy.liu.se/student2011/ts1017/grp1 to https://svn.isy.liu.se/student2011/ts1017/grp5

Documentation about Subversion can be found here

• https://svn.isy.liu.se

If you are using a Windows based computer, we recommend that you use the Tortoise client for Subversion.  We also recommend that you look at the section on Subversion in the hints and pittfalls page.

Hints and pitfalls

Based on experiences from previous year's projects, we have accumuated a list of hints and pitfalls for the projects here.  Read them carefully before starting your project work.

Written examination

The written examination of the course TS1017 is the same as for the course TSBB12.  A few older examination theses can be found here.

Calendar

• No item yet