gaussian and rayleigh distribution MRI Late Enhancement

Question

Katarzyna b. 2012-5-31

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/39984-gaussian-and-rayleigh-distribution-mri-late-enhancement

评论： A-J KHAN 2016-7-10

Hello,

I am stuck with mri images that is why I ask for help.

I want to make Rayleigh and Gaussian distribution of intensities of the image and then fit it to the histogram but it does not work.

Below you can find my example and sample of code:

clc; clear all; close all;

I=dicomread('image.dcm'); I=uint8(round(double (I)));

I=I(:)'; [muhat, sigmahat]=normfit(I); mean_value=round(muhat); sigma_value=round(sigmahat);

x=0:1:29952; for i=1:size(x')

    gauss(i)=100*exp( - (x(i)-mean_value)^2/(2*sigma_value^2)/(sigma_value*sqrt(2*pi)));
end

x=0:1:29952; for i=1:size(x')

    rayl(i)=500*exp( - (x(i)^2/(2*sigma_value^2)*(x(i)/sigma_value^2)));
end

figure(3) plot(x,gauss, x,rayl) hold on imhist(I)

THANK YOU for your help in advance

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Ryan 2012-5-31

Have you looked into using adapthisteq(Image,'Distribution','rayleigh')? There are additional settings and distributions. It is Contrast-Limited Adaptive Histogram Equalization (CLAHE). You can change the tile size that it filters the image at and the number of bins for color choices it can have in addition to some other things. I'm not sure if that will help you or not.

http://www.mathworks.com/help/toolbox/images/ref/adapthisteq.html

and

http://www.mathworks.com/help/toolbox/images/ref/adapthisteq.html#bqkqi21

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Stephen 2012-5-31

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/39984-gaussian-and-rayleigh-distribution-mri-late-enhancement#answer_49646

在 MATLAB Online 中打开

[c1, c2] = hist(I(:), 50);
f=fit(c2', c1', 'gauss1');

or for the rayleigh, you'd have to specify it in fittype where 'gauss1' is

f contains a structured variable with the amplitude, mean and std for the gausian, but for the rayleigh it should just be sigma and the amplitude.

you do not need a for loop to create the distribution curves. simply put c2 in place of x(i) and replace 100, 500 with the fit amplitude, and the mean/sigmas with their corresponding fit values. i.e.

gauss=f.a*exp( - (c2-f.b).^2/ (2*f.c^2) );

rayl should be similar. then,

hist(I(:),50),hold on, plot(c2,gauss,c2,rayl)