Getting all points which are inside of the convexhull

I have an image in 3D. I'd like to get all points which are inside its convexhull, but don't belong to the image. So I'd like to do something like this:

[K, V] = convhull(image);
result= zeros(size(image));
result(K == 1 & granule == 0) = 2;

It doesn't work, because K is of size x (the number of rectangles which create the convex hull) by 3.

How can I do it properly?

Answer 1

[Image Analyst]

Don't use image as the name of your variable because it's the name of a built-in function.

Don't use convhull(). Use bwconvhull() instead. Then do an XOR with the binary image.

[user6567]

Thanks! This is exactly what I am looking for!

[user6567]

Unfortunately, this function also returns the convex hull of the size x (the number of rectangles which create the convex hull) by 3.

[Image Analyst]

I don't know what you mean that the convex hull is three times the size of x. And I don't know what you mean that the convex hull is created by rectangles. Rectangles don't create it, though it will have a bounding box. Post your image and illustrate what you mean and point out how something is three times as big as it should be.

[user6567]

I didn't say my convex hull is three times the size of x. I said that the size of the convex hull is x by 3, what is the result of the following code:

[K, V] = convhull(image);
disp(num2str(size(K)));

For instance, the above conde prints out "1242 3". So, nothing is three times as big as it should be.

Sorry, I didn't mean rectangles, I meant polygons. And what I was trying to say is that it consists of polygons. But I guess it has nothing to do with the question. Sorry for the confusion.

In principle, when my image is of the size, let's say 100 by 100 by 100, I expect my convex hull to be of size 100 by 100 by 100. And it is of size x (i.e. 1242 ) by 3.

I can't post my data, because it doesn't belong to me and I shouldn't publish it. It also is very huge (1024^3 points). But I'll try to generate an example to illustrate the problem.

[Image Analyst]

You need to use convhulln(), not convhull().

[user6567]

Why? convhull() should also work in 3D, shouldn't it? And it still doesn't solve my problem.