[Skip navigation]

Overview

Course aim

The aim of the course is to provide the student with tools for handling high-dimensional statistical problems, models, and methods, with practical applications mainly in image analysis and spatial statistics. Of special importance are the Bayesian aspects, since they form the foundation for a large part of the modern image analysis methods. These are, in the course, related to applications in remote sensing and environmental statistics.

Course contents

Bayesian methods for stochastic modelling, classification and reconstruction. Markov fields, Gibbs distributions, deformable templates, such as Snakes. Correlation structures, multivariate techniques, analysis of discrimination. Simulation methods for stochastic inference (MCMC, etc.). Stochastic remote sensing and spatial statistics.

For further details, see the full course description.

Lectures:

10:30am-12:20am, Monday and Friday, in PDL C301, starting March 26
See also the schedule, below.

Computer exercises:

10:00am-12:50am, Wednesday, in CMU B027

Exercises:

Se schedule, below.

Examination

Three home assignments/projects.
w2: 1) Classification, (ps/pdf) handed out 2007-04-02, preferably due 2007-04-18.
w4: 2) Gaussian Markov random fields, (ps/pdf) handed out 2007-04-16, preferably due 2006-05-02.
w7: 3) Deformable templates and other, (ps/pdf) handed out 2007-05-07.
Written presentation of projects 1 and 2.
Written and oral presentation of project 3.
Presentation time:

• Friday: 1/6 10:30am PDL C301

Schedule

Final? Updated 2007-05-11.

Day				Lectures (chapters in the book)	Computer Ex.	Exercises
w1	26/3		L1	Introduction; Statistical modelling (1.1)			ps/pdf
			C1		Images in Matlab		ps/pdf
			L2	Models and Classification I (2.1-2.2, Kroisos)			ps/pdf
w2	2/4		L3	Classification II; The EM-algorithm (2.2-2.3)			ps/pdf
			C2		Classification		ps/pdf
			L4+E1	Markov fields I; Gaussian fields (4.1-4.2.3)		Bayesian Statistics 2.1-2.5	ps/pdf
w3	9/4		L5	Markov fields II; Gaussian fields (4.2.4-4.2.6)			ps/pdf
			C3		Gaussian fields		ps/pdf
			L6	Markov fields III; Discrete Markov fields (4.3)			ps/pdf
w4	16/4		L7	Markov fields IV; MCMC-simulation (4.4-5)			ps/pdf
			C4		MRF-simulation		ps/pdf
			L+E2	GMRF parameter estimation methods		Markov fields 4.4-4.7,(4.8)
Lecture and Comp.exercise break!
w6	30/4		L8	Deformable templates I; Shape description, distance metrics, Procrustes-analysis			ps/pdf
			C5		Shape analysis		ps/pdf
			L9	Deformable templates II; Shape variation and modelling, shape-PCA			ps/pdf
w7	7/5		L10	Deformable templates III; Warping, image-PCA finn-data.mpg: mpeg/gz/ mpeg/zip (0.7MB) finn-impca.mpg: mpeg/gz/ mpeg/zip (1.6MB) finn-rewarp.mpg: mpeg/gz/ mpeg/zip (8.8MB) finn-movie.mpg: mpeg/gz/ mpeg/zip (3.1MB)			ps/pdf
			C6		Warping and image-PCA		ps/pdf
			L11	Deformable templates IV; Shape reconstruction from images, Snakes			ps/pdf
w8	14/5		L12+E3	Weighted PCA and other miscellanea		Statistical shape analysis (the second edition chapter 6 exercises)	ps/pdf
			C7		Snakes		ps/pdf
			L13	Discussion of projects 1 & 2			ps/pdf
w9	21/5	Mon		No more lectures (I'll be in my office during the lecture hours, and in the computer lab during the comp.ex. hours)
w10	1/6	Fri		Project presentations, ca. 10 minutes each

Matlab

Matlab-files and data are available from the Matlab page.
A list of useful Matlab tips and tricks is also available, Matlab-tips (ps)/ (pdf).

Literature

The locally produced book Image Modelling and Estimation - A statistical approach, is available from the literature homepage.

People

Course administrator/lecturer:

Finn Lindgren
room: PDL B213
phone: ?
e-mail: finn@maths.lth.se