Image subtraction, distance maps and skeletons to measure width

[eburell]

Dear members,

I am trying to measure the thickness of a sponge spicule. I was told to dilate a binary image (from a HSB brightness slide) of the spicule and subtract the original to create a perimeter. In the next step, a distance map was to be created and I assume this can be done with the EDM plugin (I'm currently having problems with this plugin and I am unsure of what a proper output from using this would be). Finally, a skeleton of the spicule was to be created.

I am however not sure on how to calculate the distance from the skeleton to the perimeter using the distance map. Could anyone be so kind as to offer me some tips on this step?

One of the images I am working on can be found here: https://skitch.com/aiden/r9e7y/capture-3

Thanks in advance,
Eli

[Gabriel Landini]

On Friday 13 May 2011,  eburell wrote:

> I am trying to measure the thickness of a sponge spicule. I was told to
> dilate a binary image (from a HSB brightness slide) of the spicule and
> subtract the original to create a perimeter. In the next step, a distance
> map was to be created and I assume this can be done with the EDM plugin
> (I'm currently having problems with this plugin and I am unsure of what a
> proper output from using this would be). Finally, a skeleton of the
> spicule was to be created.
>
> I am however not sure on how to calculate the distance from the skeleton to
> the perimeter using the distance map. Could anyone be so kind as to offer
> me some tips on this step?

Better than that, binarise the spicules, duplicate the image and compute their
distance map in the duplicate.
The maximum grey value in each spicule (in the distance map image) corresponds  
to the thickness in pixel units (or rather the distance from that foreground
pixel to the closest boundary pixel in the object).
You can get the thickness of each spicule using the particle analyser on the
binary image and "redirecting" the measurements to the distance map image to
extract the "Maximum" grey value.

Cheers

Gabriel

[karo03]

In addition to Gabriel you could make from another duplicate of the
binarized image the skeleton and measure the skeleton with redirection as
recommended using the MEAN. This will be some sort of mean thickness, in
fact half of it.
Karsten


2011/5/14 Gabriel Landini <[hidden email]>

> On Friday 13 May 2011,  eburell wrote:
> > I am trying to measure the thickness of a sponge spicule. I was told to
> > dilate a binary image (from a HSB brightness slide) of the spicule and
> > subtract the original to create a perimeter. In the next step, a distance
> > map was to be created and I assume this can be done with the EDM plugin
> > (I'm currently having problems with this plugin and I am unsure of what a
> > proper output from using this would be). Finally, a skeleton of the
> > spicule was to be created.
> >
> > I am however not sure on how to calculate the distance from the skeleton
> to
> > the perimeter using the distance map. Could anyone be so kind as to offer
> > me some tips on this step?
>
> Better than that, binarise the spicules, duplicate the image and compute
> their
> distance map in the duplicate.
> The maximum grey value in each spicule (in the distance map image)
> corresponds
> to the thickness in pixel units (or rather the distance from that
> foreground
> pixel to the closest boundary pixel in the object).
> You can get the thickness of each spicule using the particle analyser on
> the
> binary image and "redirecting" the measurements to the distance map image
> to
> extract the "Maximum" grey value.
>
> Cheers
>
> Gabriel
>

[Gabriel Landini]

I wrote:
> The maximum grey value in each spicule (in the distance map image)
> corresponds to the thickness in pixel units ...

Actually it is "half the maximum thickness". Thanks Karsten for mentioning
this in the previous email.

Cheers
Gabriel

[eburell]

Thank you Gabriel, Karsten.

The EDM-methods you mentioned worked great!

Regards,
Eli