background noise and nearst point

hi experts , i have two questions and hope you can help me.

1- i was able to form a 3d image of a brain from a 2d slices. but i want to set all the gray level values out side the brain to zero (i.e eliminate the noise at the background)

2- A(x,y,z) is a point within the 3 D volume , it could be within the brain or outside the brain. i wish to find the nearest point to A(x,y,z) on the surface of the brain.

Answer 1

[Image Analyst]

% Make a mask:
mask3D = image3D > thresholdValue;
% Fill it so we don't change low gray level voxels inside the head.
mask3D = imfill(mask3D, 'holes');
% Mask volume outside the mask
image3D(~mask3D) = 0;
% Then call bwdist() to get the closest distance from each point to the outer surface.

[zawaiter]

thanks image analyst for your help.i didn't yet try your code, a little busy but i think i know the idea now. from what i understand, can the thresholdvalue be computed with a formula so i can use the code for different images without need to look at the gray levels each time? also can bwdist() get the coordinate of the nearest point not just the distance.

[Image Analyst]

Look at the second return argument from bwdist(). I've never used it though. Be aware that there can be several voxels that are the closest one to a particular reference voxel.

Answer 2

[Amith Kamath]

For either case, I would imagine you will need a mask for the brain regions. This can be done either manually using http://www.mathworks.com/help/images/ref/imfreehand.html (which would be painful indeed) or you could also use tools like http://www.mathworks.com/matlabcentral/fileexchange/19567-active-contour-segmentation for a 3D dataset like yours.

for 1. http://www.mathworks.com/help/images/ref/bwareaopen.html is the best way to eliminate small noisy regions.

for 2. http://www.mathworks.com/help/images/ref/bwdist.html may provide some insight.

[zawaiter]

thank AMITH for help i need some time to try your help.