It is our strong belief that the only way to learn image analysis is to try the algorithms on real images with a computer. Therefore, a number of computer exercises are included in this course. They are not compulsory, but it is highly recommended that you do them.

Each exercise will have a number of scheduled class hours in a Computer Room. The exercises (except for the first one) will contain quite a lot of programming when implementing the algorithms and your ideas.

Of course, if you have questions regarding the CEs or the m-files, don't hesitate to send me an email.

Back to Statistical Image Analysis homepage.

This exercise is for you who have none or little experience using Matlab with images.

- Computer Exercise 1: Getting Started [postscript-file | pdf-file] (updated version 050119)

- fumitory-007.jpg (Use the right buttom to 'Save link as...'! )

- Computer Exercise 2: Basic Image Processing [postscript-file | pdf-file]
- Additional info to Computer Exercise [pdf-file]

- The Weed seed images.

**Files:**

- resample.m (If you get stuck when writing your own programme.)
- midpoint.m
- contour.m

Old versions:

Read through the chapter regarding Pattern Recognition
in the Lecture Notes (IASS) **before** you begin with this exercise.

**Document:**

- Computer Exercise 3: Pattern Recognition [postscript-file | pdf-file]

The **Data** used in this exercise can be found
here.

**Files:**

Here is the half-done Matlab file for estimating the error rates:

Here's the last CE, which deals with simulation of noise and Markov Random Fields.

- Computer Exercise 4: Statistical Image Models [postscript-file | pdf-file]

Here is an outline of the main program. You may need the function neighbours.m from CE2. Probably, some adjustments have to be made, though.

Here's the files for the Potts model experiment in Section 3.4 in the CE.

- start64.tif (this image looks totally black when not viewed with Matlab's imshow)
- main_sim_anneal.m with subroutines:

Last modified: by anders.sjogren@math.chalmers.se