how to count the number of pixels of each region in an image ( searching for big regions) ?

I want to count the number of pixels in each region of an image. The image contains more than one region and i want to ignore the big regions, so i want to search for big regions. Does any one know how to do this

[Matt Kindig]

How are the "regions" defined in your image? Can these regions be identified from using bwlabel() or similar? Linking to your image would also help.

[Shahd Ewawi]

The regions are defined as black & white.

Answer 1

[Image Analyst]

Use regionprops(). Assuming you already have a binary image of your regions, get all the areas like this:

% Measure all blob areas.
blobMeasurements = regionprops(binaryImage, 'Area');
allAreas = [blobMeasurements.Area]
% Sort in ascending order
[sortedAreas, sortIndexes] = sort(allAreas, 'ascend');
% Take the first 10 smallest:
smallAreas = sortedAreas(1:10);

[Shahd Ewawi]

 i cant use regionprops() . because currently i work on matlab as simulation and i want to convert the code to php so i cant use this function

[Image Analyst]

Then you'll have to write your own connected components labeling routine. Why are you even using MATLAB at all then? Why not just do everything in php right from the start, since you can't use any built-in MATLAB functions?