Preliminaries
- This file is "Examples..." under the "Help" pull-down menu.
- Without a valid licence file, use is very restricted. The licence
file is a one-line text file called: cea_lic. It can be set up using
the "Install licence" option under the "File" pull-down menu. Or it
can be copied as a file to the bin subdirectory.
- The self-extracting binaries for MR1, MR2, and MR3, produce the
main program, and bin and images (data, in the case of MR1)
subdirectories. The bin subdirectory contains all of the multiresolution
analysis programs.
- If you would like to use the MR1, MR2, or MR3 packages, we would
recommend that you put the cea_lic file in your c: or some other
convenient directory; and then copy it into the bin directory corresponding
to any one of the MR1, MR2, and MR3 packages.
- "Open..." under the "File" pull-down menu supports various 2D image
formats: JPEG, PGM and FITS. FITS is recommended for most processing
because it supports 32-bit floating pixel values.
- Possible warning: extremely long path names will cause a problem
since a fixed total length is assumed for the path and file name.
- Sample images available in the images subdirectory are:
- Opening the galaxy image gives the following. Scroll bars allow
for a very large image. Both halving and doubling the display size are
possible (under the "View" pull-down menu).
Segmentation
- A segmentation with 5 classes (Gaussian mixture modelling of the
marginal distribution) gives the following. Numeric information
is presented in a pop-up window which can be saved as a text file.
Contrast Stretching
- Let us load galaxy.jpg.
Instead of a histogram equalization we stretch
the wavelet coefficients using under the "Tools" pull-down menu the "Contrast"
program. The menu itself is as follows.
The wavelet-based contrast stretching gives the following result.
Wavelet Transform
- The wavelet transform allows for display of just one resolution
scale, or an aggregate of a number of resolution scales. So let us
look for the more smooth aspects of the galaxy image. We will use
"Multiresolution Transform" under the "Filtering" pull-down menu.
This then gives as output:
Now we look for the sharper (more high frequency) information. We
use the following resolution scales.
This gives the following.
Edge Detection
- Using the stones image, we apply "Edge detect" from the "Tools"
pull-down menu.
This gives:
Applying the same wavelet coefficient contrast stretching as before gives
the following.
History of Processing
- A history of processing is kept in text file mr2d_v2.log