Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Centroid estimation of grayscale image

625 views
Skip to first unread message

Florian

unread,
May 26, 2008, 5:28:12 AM5/26/08
to
Hello folks,

I am looking for a function that calculates the centroid (center of mass) of
a grayscale image/matrix with different algorithms.
Does anybody know one?

Thanks in advance, Greetings

Florian


Message has been deleted

Daphne

unread,
May 26, 2008, 3:26:02 PM5/26/08
to

and if you have an earlier version of Matlab, check this
out
http://blogs.mathworks.com/steve/2007/08/31/intensity-
weighted-centroids/

Daphne


ImageAnalyst <imagea...@mailinator.com> wrote in
message <ec2ff76a-cfb0-4b30-a8be-
6677cf...@c58g2000hsc.googlegroups.com>...
> On May 26, 5:28=A0am, "Florian" <florian.m...@dlr.de>
wrote:


> > Hello folks,
> >
> > I am looking for a function that calculates the

centroid (center of mass) =


> of
> > a grayscale image/matrix with different algorithms.
> > Does anybody know one?
> >
> > Thanks in advance, Greetings
> >
> > Florian
>

> ---------------------------------------------------------
---------
> Florian:
> The regionprops() function in the image processing
toolbox does that
> (with R2008a & later only). What do you mean by "with
different
> algorithms"? I only know of the standard definition.
regionprops()
> does also have a non-weighted centroid that just depends
on the
> binarized image and disregards the gray levels in the
image, if you're
> looking for that.
> Regards,
> ImageAnalyst

Florian

unread,
May 27, 2008, 5:47:44 AM5/27/08
to
Thanks for the answers. With different algorithms I mean different ways of
calculating the centroid, especially with subpixel accuracy. E.g. thers is
the center of mass calculation, the use of the Fourier transform of wavelet
transform. I heard that by using the wavelet transform the highest accuracy
is obtained.

Florian


"Daphne " <daphnew_t...@yahoo.com> schrieb im Newsbeitrag
news:g1f2sa$118$1...@fred.mathworks.com...

0 new messages