CODEWITHSUNDEEP
matlabimage-processingmatrix
mugetsu
mugetsu

Reputation: 4398

Matlab matrix cropping of clusters

Given a matrix:

0 0 0 0 0 0 0 0
0 0 0 1 1 1 0 0
0 0 1 1 0 1 0 1
0 0 1 1 0 1 0 0
0 0 0 0 1 0 0 0

notice how the 1's are mainly clustered except for the 1 in the 3rd row, last column

what I want to do is get rid of excess 0's and keep only the cluster of 1's. This is what I'm trying to do with an image, so the cropped result should be rectangular:

0 1 1 1 
1 1 0 1
1 1 0 1
0 0 1 0

the above crop is the smallest rectangular crop I can do to cover the most amount of 1's.

The problem I have right now is how do to this in matlab? I can't simply strip all 0's and I have to watch out for outlier 1's. What is the most efficient method to tackle this problem?

I can only think of manually going row by row, column by column to look for clusters.

Upvotes: 1

Views: 930

Answers (1)

dantswain
dantswain

Reputation: 5467

What you're trying to do is equivalent to labeling 2D connected components in a binary image. If you have access to the image processing toolbox, the function is bwlabel: http://www.mathworks.com/help/toolbox/images/ref/bwlabel.html

If you don't have acccess to bwlabel, there are several algorithms out there. The fastest one I know of is the one that OpenCV uses, which finds closed contours in the image and is based on

F. Chang, C.-J. Chen, and C.-J. Lu, “A linear-time component-labeling algorithm using contour tracing technique,” Computer Vision and Image Understanding, vol. 93, no. 2, pp. 206–220, 2004

Upvotes: 3

Related Questions