nagyistoce / centralslice

Automatically exported from code.google.com/p/centralslice
0 stars 0 forks source link

Reconstructed image appears to be out of range (aliasing?) #2

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1. change the phantom to a square pulse
2. execute "run.m"
3. open "5_reconstruct_xy_real.png"

What is the expected output? What do you see instead?
The expected image should be a white square, but instead it is some large 
circles:

http://code.google.com/p/centralslice/source/browse/branches/5_reconstruct_xy_re
al.png?spec=svn39&r=39

Please use labels and text to provide additional information.

Original issue reported on code.google.com by antony.s...@gmail.com on 26 Feb 2011 at 10:02

GoogleCodeExporter commented 9 years ago
Alising still occurs in reconstructed image in revision 83.

Here's the steps to reproduce the result:
1) download ver8.zip
http://code.google.com/p/centralslice/downloads/detail?name=ver8.zip
2) Make the following settings in file 'start_simulation.m'
{{{
shape = 'Modified Shepp-Logan'
N_image =  231
N_theta =  180
SNR = 0
interp_m = 'nearest'
}}}
3) Open file "start_simulation.m" and run in Matlab
4) Look at the figure titled "Reconstructed Image"

How should I reduce the artefacts in the image?

Original comment by antony.s...@gmail.com on 6 Mar 2011 at 4:56

Attachments:

GoogleCodeExporter commented 9 years ago
Hey, I followed Lewis' reference to reduce artefacts by oversampling. Please 
refer to revision 92 for my modification to the code.

To try this method, please download the attached zip file and extract all files.

Open "start_simulation.m" in Matlab, change the value of "oversampling_ratio" 
and run the script. You will need to wait for a considerably long time (>3 
minutes) and then look for "Reconstructed_Image.png" in the same folder.

This method is very good and it reveals a hidden problem in interpolation!
I will file an separate issue to describe the bug in detail.

Original comment by antony.s...@gmail.com on 6 Mar 2011 at 9:29

Attachments: