Thread Subject:

CT image 2

Ok, I have a problem with CT image segmentation. I created a code for image segmentation of the lungs parenchyma, but I have one problem that I just cannot figure out how to solve it.
Hee is the image.
http://i1192.photobucket.com/albums/aa325/privatemeet/table.jpg
I want to remove the table under the lungs. Is there any way to do it automaticaly, without any interaction of the user? Also, this table can be moved to the left, right, down, up. The height is also not always the same.
I tried bunch of methods, but I'm stuck...

Any help?

Here is the another example of the image. In this case I dodnt converted to binary image.
http://i1192.photobucket.com/albums/aa325/privatemeet/1.jpg
I need to remove the table, and also you can see that the table is moved to the left...

Find the centroid of each distinct object in segemented image using bwlabel+regionprops. The table is the object which has the centroid with smallest (or largest) y-coordinate.

"privatemeet none" <privatemeet@yahoo.com> wrote in message <ibirbu$31d$1@fred.mathworks.com>...
> Ok, I have a problem with CT image segmentation. I created a code for image segmentation of the lungs parenchyma, but I have one problem that I just cannot figure out how to solve it.
> Hee is the image.
> http://i1192.photobucket.com/albums/aa325/privatemeet/table.jpg
> I want to remove the table under the lungs. Is there any way to do it automaticaly, without any interaction of the user? Also, this table can be moved to the left, right, down, up. The height is also not always the same.
> I tried bunch of methods, but I'm stuck...
>
> Any help?

Smart thinking!
Do yo have any code regarding this?

Thanks

"privatemeet none" <privatemeet@yahoo.com> wrote in message <ibis65$nff$1@fred.mathworks.com>...
> Here is the another example of the image. In this case I dodnt converted to binary image.
> http://i1192.photobucket.com/albums/aa325/privatemeet/1.jpg
> I need to remove the table, and also you can see that the table is moved to the left...

You could do this all in one shot with my shrinkWrap function:
%%
I = rgb2gray(imread('table.jpg')); %Read

M = shrinkWrap(I,'objthresh',3200,'biggest'); %create map

Ibw = I.*uint8(M); %apply map

imtool(Ibw) %show
%%

shrinkWrap is available here:
http://www.mathworks.com/matlabcentral/fileexchange/29175-shrinkwrap

Thanks!
It is working, I need only to see what will happen when I apply this code on a series of 200 images...

"privatemeet none" <privatemeet@yahoo.com> wrote in message <ibjdsb$ian$1@fred.mathworks.com>...
> Thanks!
> It is working, I need only to see what will happen when I apply this code on a series of 200 images...

You can apply it to an entire image volume at once if you wish. The syntax is the same.