Image windowing for doing emgm

1 view (last 30 days)
bahar h
bahar h on 24 Jul 2012
Edited: Walter Roberson on 26 Jun 2017
Hi I want to do windowing on an image then run EM algorithm on it.I used both blockproc and colfilt but both of them cause error. I face many error for using blockproc but one error:
Error using ==> reshape
To RESHAPE the number of elements must not change.
Error in ==> colfilt at 183
b(i*mb+brows,j*nb+bcols) = ..
for using colfilt. what do you recommend me to do?
Thank you

Answers (1)

Image Analyst
Image Analyst on 24 Jul 2012
I recommend you give enough information for us to be able to solve your problem.
  2 Comments
bahar h
bahar h on 25 Jul 2012
actually I am supposed to simulate a paper in which I should calculate the joint probability density of gray levels of a suitably chosen search window in a mammogram:x=(x1,x2,...xn),assume the joint probability in form of Gaussian components.finally I should have a data set S by pixelwise scaning of the mammogram with the search window :S={x(1),x(2),...}.and then obtain Gaussian mixture model and maximum likelihood by EM algorithm,I think for these parts(GMM &EM)I can use stprtool but I have problem with search window yet.my image is a 1024*1024 uint8 and the size of window used in the paper is 13*13 pixels with trimmed corner?(I don't know trimmed corner and I am not sure it be necessary?)my mammogram belong to a database other than the mentioned paper but I hope they be the same size. Thank you
Image Analyst
Image Analyst on 25 Jul 2012
I'm not sure what the question is, but when you scan an image with a window, yes, you are going to have to know how to handle the edge conditions where the window is "off" the edge of the image. This is a common situation in image processing and there are a variety of ways to handle partial windows, such as assuming image is zero for window pixels not over an image, or just ignoring those pixels, or renormalizing, etc.

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!