Spatial Statistics with Image Analysis HT12

News

First lecture on September 4:th 8.15-10.00 in E:C

Introductory lecture F0 given at the Image analysis (FMA170) course on 2011-10-04.

Examination

Three home assignments/projects. The assignments will be handed out during the 2:nd, 4:th and 6:th course week.
v2: 1) Classic Kriging
Data: temp1997_asgn1.csv
temp1997_asgn1_valid.csv
temp1997_covar.csv
Handed out 2012-09-11, due 2012-09-28.
v4: 2) Gaussian Markov fields and parameter uncertainty
Data: asgn2_ndvi_2006_05.mat
asgn2_ndvi_2006_06.mat
asgn2_ndvi_2006_07.mat
asgn2_ndvi_2006_08.mat
asgn2_ndvi_2006_09.mat
asgn2_ndvi_2006_10.mat
gmrf_negloglike_skeleton.m
Handed out 2012-09-25, due 2012-10-12.
v6: 3) MRFs, non-Gaussian data, etc.
Data: HA3_forest.mat
titan.jpg
lab5simple.dat
asgn1_lund.mat
Functions: gmrf_negloglike_Po_skeleton.m
gmrf_taylor_Po_skeleton.m
Estep_corrupt_skeleton.m
Mstep_corrupt_skeleton.m
normmix_posterior.m
mrf_sim.m
Handed out 2012-10-09. Written and oral presentation of project 3.
Presentation times:
Sign up for presentations using the Doodle, at most four groups per time (presentations should hopefully be shorter than the full aloted time slots).

• Friday: 26/10, 11:00-12:00, TBD
• Wednesday: 31/10, 10:15-12:00, MH:309C
• Wednesday: 31/10, 13:15-15:00, MH:330
• Thursday: 1/11, 13:30-15:00, MH:330
• Thursday: 1/11, 15:15-16:30, MH:330

Schedule

Last updated 2012-10-18.

Recommended reading in parentheses. FL:## denotes chapters in F. Lindgren, Image Modelling and Estimation - A statistical approach, HSS:## gives chapters in the Handbook of Spatial Statistics.

Day				Lectures	Exercises
v1/36	Tue	4/9	L1	Introduction; Statistical modelling (FL: 1.1; HSS: 1, 7.1-7.2)
	Thu	6/9	C1		Image data and dependence in Matlab
	Fri	7/9	L2	Spatial models I; Random fields and covariance models (FL: 3; HSS: 2)
v2/37	Tue	11/9	L3	Spatial models II; Classic Kriging and covariance estimation (FL: 3; HSS: 3)
	Thu	13/9	C2		Covariance estimation
	Fri	14/9	L4	Spatial models III; Covariance estimation (ML) Discussion of Lab 2 and Home Assignment 1 (HSS: 4.1-4.6)
v3/38	Tue	18/9	L5	Markov fields I: Gaussian fields (FL: 4.1-4.2.3; HSS: 12.1.1-12.1.4, 13.2)
	Thu	20/9	C3		Gaussian Markov random fields
	Fri	21/9	L6	Markov fields II: Gaussian fields (FL: 4.2.4-4.2.6; HSS: )
v4/39	Tue	25/9	L7	Hierarchical models II: Bayesian analysis (HSS: 7, 12.1.7)
	Thu	27/9	C4		Work on HA1 or HA2
	Fri	28/9	L8	Classification, EM-algorithm (FL: 2.1-2.3)
v5/40	Tue	2/10	L9	EM-algorithm (cont.) (FL: 2.1-2.3)		See L8
	Thu	4/10	C5		Classification
	Fri	5/10	L10	Markov fields III: Discrete fields (FL: 4.3; HSS: 12.1.8)
v6/41	Tue	9/10	L11	Markov fields IV: Parameter estimation for discrete fields (FL: 4.4,5.2; HSS: 12.1.9)
	Thu	11/10	C6		Discrete field simulation and estimation
	Fri	12/10	L12	Discrete field reconstruction, the EM algorithm
v7/42	Tue	16/10	L13	Non-Gaussian data (HSS: 14)
	Thu	18/10	C7		Work on HA3
	Fri	19/10	L14	Discussion of projects 1 & 2

General Information

The course is given once a year. It is worth 7.5 ECTS credits and consists of 14 2-hour lectures, 7 3-hour computer exercises, and three home assignments.

Introductory lecture F0 and course plan english (svenska)

Literature

F. Lindgren, Image Modelling and Estimation - A statistical approach. The Compendium is available from the department or from the literature homepage.

Most of the material is also coverd in: A. Gelfand P. Diggle P. Guttorp M. Fuentes (Eds), Handbook of Spatial Statistics, which is available as an E-book from Lund University Libraries.

Prerequisits

A basic course in mathematical statistics, and at least one of: Markov processes (FMSF15/MASC03) Stationary Stochastic Processes (FMSF10/MASC04) Image analysis (FMA170) Familiarity with Matlab is recommended.

Recommended related courses (not prerequisits)

Monte Carlo methods for stochastic inference (FMS091/MASM11) Linear and Logistic Regression (FMSN30/MASM22) Computer vision (FMA270)