[Resources] [Terminology] [Units] [MCQs]


   

Segmentation and Noise Filtering


30 min.
Objectives:
At the end of this practical you should understand the use of rank filtering for;
- simple image segmentation.
- and noise reduction.

Terminology:


See:
Image Segmentation
Median Filtering
Thinning and Thickening
Read the introductory sections in the MATLAB user guide regarding file I/O and displays. Familiarise yourself with the MATLAB "HELP" facility.

Run the matlab "demos" and select the "Region Labelling of Steel Grains" deomo from the "Image Processing" toolbox. Run through the sequence of processing steps provided making sure that you understand the algorithms being used at each stage.

Now run the "noise reduction filtering" demo. Apply the available image noise and then observe the effects of the noise reduction algorithms.

Try to find out;
What is the best filter for "gaussian" random noise?
What is the best filter for "salt and pepper"(ie:"drop-out")?
In the context of our terminology, how would you describe "speckle"?



(c) Imaging Science and Biomedical Engineering 2000 [paul.bromiley@man.ac.uk]

Valid HTML 4.01!