Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: topological filter
Posted:
Jan 20, 2010 4:03 AM


ImageAnalyst <imageanalyst@mailinator.com> wrote in message <c2d7c1c31b1a435abe0f22ae62032d09@u41g2000yqe.googlegroups.com>... > Joseph : > Yes. You can look at the Euler Number of the object. This is > returned from regionprops() in the image processing toolbox. > Basically the process is this (for almost all image processing > applications: > 1. threshold image to binarize it > 2. connected components labeling, using bwlabel() or bwconncomp() > 3. regionprops(), to make the measurements > > Here is the definition of Euler Number: > 'EulerNumber' — Scalar that specifies the number of objects in the > region minus the number of holes in those objects. This property is > supported only for 2D input label matrices. regionprops uses 8 > connectivity to compute the EulerNumber measurement. > > My demo at http://www.mathworks.com/matlabcentral/fileexchange/25157 > will help you to do the measurement.
Hi, continue on this project... My study objects are the diameters of bubbles taken by CCD. Using 1)binarization +2)bwlabel + 3)regionprops I can manage to get all ring shapes bubbles. However, Other bubbles are in U shapes after binarization because of their initial blur boundaries, I wonder if I can use the scale invariant moments to differentiate the U shape and loaf shape objects both of their eular number =1..?? Could you give me some suggestions? Thanks in advance...



