Finding area, average and moments of parts of a segmented image

I have an image divided into sections. How can I find the area, mean, second moment and third moment of those sections? I used the "bwconncomp" function to detect the partitions.

Answer 1

[Image Analyst]

See my Image Segmentation Tutorial in my File Exchange:

Image Analyst's File Exchange

It's a generic, general purpose demo of how to threshold an image to find blobs, and then measure things about the blobs, and extract certain blobs based on their areas or diameters.

Also check out the attached demo on image moments.

[Image Analyst]

You have a labeled image -- where each blob of aa binary image has been labeled with an identifying number. OK, fine. Then you say (ambiguously) "How can I find the area, mean, second moment and third moment of those sections?"

Finding the area of each blob (what you call "section") is trivial

props = regionprops(cc, 'Area'); % cc came from bwconcomp()
allAreas = [props.Area]

OK, now for mean. Calling it average instead of mean does not help. I figured that already. But what do you mean by the "mean of those sections"? Do you want a mean for each blob? What would the mean of one blob mean to you??? Or the mean area over all blobs, like this:

meanArea = mean(allAreas)

I have no idea what you mean. Please clarify.

OK, now for moments. Do you mean like variance is the second moment, skewness is the third moment, and kurtosis is the fourth moment? Do you want those of the areas? If so, you get one overall moment, not a moment for each blob.

If you want the moment for each blob separately, then you must be talking about spatial moments, like intertial moments. My attached demo gives the "spatial central moments" of each blob, like the moment of inertia etc. The first central spatial moment is just the centroid (location) of the blob. And from your physics/mechanics class, you know what moment of intertia is.

[besaret koçak]

I want to find the area, mean, second moment, third moment and variance for each blob separately. I wrote a code like this for this. İs it true

for i=1:1:CC.NumObjects
    moment1(i,1) = moment(image(CC.PixelIdxList{1,i}),1)
    moment2(i,1) = moment(image(CC.PixelIdxList{1,i}),2)
    moment3(i,1) = moment(image(CC.PixelIdxList{1,i}),3)
    area(i,1) = bwarea(CC.PixelIdxList{1,i})
    means(i,1) = mean(GPR(CC.PixelIdxList{1,i}))
    var_result(i,1)= var(double(image(CC.PixelIdxList{1,i})))
end

[Image Analyst]

image is a function so you would not want to use that for your variable name. PixelIdxList just gives indexes. I think you'd want PixelList which gives (x,y) coordinates. I don't know what your GPR function or array is, but your means and var_result do not make sense to me. Maybe by mean you mean the mean x and y, in other words the centroid which you should get directly from regionprops.