Author: 
J. Anthony Parker, MD PhD ( J.A.Parker@IEEE.org ) 
Dates: 
2002/05/22: First version 
Source: 
The source file, TwoD_HistogramTP.java
, is also contained in Align Stacks. 
Installation: 
Download TwoD_HistogramTP.class
to the plugins or "plugins/Align Stacks" folder and restart ImageJ. 
Description: 
This ImageJ PlugIn shows a histogram of the intensities in
two images. The xaxis represents the intensity in the first
image. The yaxis represents the intensity in the second image.
Several options for the displayed are defined:
1. Pdf = the value at point x,y in this joint histogram is the
probability of pixel pairs in the source images where the value of
Image1(x',y') = x and the value of Image2(x',y') = y.
2. Joint Entropy = Pdf * log(Pdf).
3. Product = the product of one dimensional marginal histograms.
4. Mutual Information = Pdf * log(Pdf/Product).
5. Conditional Entropy(12) = Pdf * log(Pdf/marginal Pdf for image 1)
6. Conditional Entropy(21) = Pdf * log(Pdf/marginal Pdf for image 2)
Pdf = probability density function. The sum of values in a Pdf is
1.
For display, the values have been multiplied by the number of
pixels
in the histogram, e.g. the average value in a pdf is 1.
This plugin may be interesting for visualizing similarity measures used
in registration.
Zhu YM: Volume image registration by crossentropy optimization. IEEE
Trans Med Imaging 21:174180, 2002.
