This tutorial explains how to use the Intensity Rescaler program in command line.

Requirements

   You can download the Intensity Rescaler program in our download section

   You can download sample data here (zip format - 2.13Mo)

   IntensityRescaler can process images with the following format: GIPL (.gipl), MetaImage (.mha), Analyze (.hdr)

Overview

   What you need:

- 1 target image (.gipl, .hdr or .mha): This image is the reference image
- 1 target segmentation image: segmentation (Gray matter, White matter and CSF) of the target image.
- N source image(s) (.gipl, .hdr or .mha): Image(s) you want to adjust based on the target image
- N source segmentation image(s): Segmentation(s) of the source image(s)

   Intensity adjustment in 3 steps

IntensityRescaler adjust images using 3 different steps

Original images: - Target and Source max values are not the same - White matter, Gray matter and CSF distributions are not the same between the 2 images
Step 1: Target intensity windowing: - Target maximum is computed using: newmax = Mean(WhiteMatter)+(sigma*Variance(WhiteMatter)) - Apply intensity windowing to stretch histogram with new maximum.
Step 2: Source intensity windowing: - Source maximum is computed using: newsourcemax = Mean(WhiteMatter)+(sigma*Variance(WhiteMatter)) - Apply intensity windowing to stretch histogram to make: newsourcemax = targetmax
Step 3: Class matching: - This is an iterative process which try to adjust mean values for each labels. For example: MeanTarget(White) = MeanSource(White) MeanTarget(Gray) = MeanSource(Gray) MeanTarget(CSF) = MeanSource(CSF) - This is very prowerfull if the source and the target images have the same maximum.

Example

   The first thing is to create a configuration file. To do that type:
> IntensityRescaler -create irescalscript.txt

This will create a irescalscript.txt file which is a demonstration script
Important: You have to edit this file before running IntensityRescaler with it.

   Edit the script file using a text editor
You can see the script generated in the sample data directory.

   Execute IntensityRescaler:
> IntensityRescaler -input irescalscript.txt

if you want to rum it in verbose mode:
> IntensityRescaler -input irescalscript.txt -v

Result

These images are generated using the example script

Original images with segmentions (Target image: Left and Source images: 2 image right)

Original images before Intensity Rescaling

Original images after Intensity Rescaling