# clustering an image using matlab

34 views (last 30 days)
izyan hanum on 24 Mar 2015
Commented: izyan hanum on 24 Mar 2015
i have error in coding below. can someone help me.? when i run they only calculate the area of cell but error when i want to cluster them based on area ..please help me

Image Analyst on 24 Mar 2015
I showed you how to segment and measure the regions in your prior question. Please explain what "cluster them based on area" means. You can make a can chart of the blob areas. Are you saying that you, for some reason, expect the areas to be clustered into two size ranges? So that the histogram would be bimodal (with two humps)? If so, what do you based that one?

#### 1 Comment

izyan hanum on 24 Mar 2015
cluster them based on area mean : after i run the coding they will show me the area or the cell. i need to cluster them as simple as i can. mean
area < 2000 is pus cell figure(4), imshow(puscell); area > 2000 is epithelial cell figure(5), imshow(cepithelial cell);
i already make and run the coding but have error. here is my coding. it make me really stress because it so difficult . are you understand now sir.?

izyan hanum on 24 Mar 2015
sir image analysis are you understand what i mean sir? i need cluster them as simple as i can like
area < 2000 is pus cell figure(4), imshow(puscell);
area > 2000 is epithelial cell figure(5), imshow(cepithelial cell);

Image Analyst on 24 Mar 2015
OK, I see from your duplicate question that you actually mean "select" or "classify" or "do size filtering" rather than "cluster" - the wrong terminology is what threw me off.
I suggest you look over my Image Segmentation Tutorial where I show how to exactly that. I get images of the two sizes of coins. I know you got your answer from the other person, but this tutorial has lots of other good stuff in it besides size filtering.
izyan hanum on 24 Mar 2015
yes sir that i mean. sorry make you confuse. i will look your tutorial and study more because i need to enhance the image. thank you for your tutorial and helping me. thanks again sir .