Moment Calculator and Texture Analysis?

[Andy Weller]

Dear all,

We have been using both the Moment Calculator and Texture Analysis
plugins for sometime now. I have some concerns about using these though.

* The Moment Calculator plugin only works with rectangular ROIs (see
http://rsb.info.nih.gov/ij/plugins/moments.html)

* The Texture Analysis plugin only works for square ROIs (see
http://rsb.info.nih.gov/ij/plugins/texture.html)

Maybe this is of no consequence to the final result?!

If it is, does anyone know how easy it will be to update these plugins
to reflect polygonal ROIs? Does anyone know how to do this? Can anyone
do this?

(I have purposely shied away from plugin development as my programming
skills, I feel, are really not up to it. Maybe with a bit of spare time
in the future I will be able to update these plugins...)

Many thanks, Andy

[Andy Weller]

Dear all,

We have been having an off-group discussion about textural analysis. I
currently take the whole textural measurements from a complete particle
- i.e. not selective parts. However, I would like to know:

Would it be possible to draw a line across, perhaps, the longest
diameter and measure textural (and gray-level) distributions across that
line?

Would this also be possible across concentric circles drawn through our
particles?

Does this make imaging sense?

Many thanks, Andy

[karo03]

Hi Andy,
Perhaps my two cents:
Texture measures like grey-value co-occurrence and/or higher moments  
are strongly dependent on the number of pixels taken to characterize  
a (one) textural homogenous region and of course on the neighborhood  
of these pixels! Hence thin regions (lines) do not or nearly not  
represent textures. Also from logical points of view, is there any  
texture if only a line is visible?
Beside that, have a look into the pleasures of texture features  
(Invariance of textural features in image cytometry under variation  
of size and pixel magnitude. Anal Cell Pathol, 8:117-133, 1995)
Regards
Karsten
Am 22.02.2007 um 10:47 schrieb Andy Weller:

> Dear all,
>
> We have been having an off-group discussion about textural  
> analysis. I currently take the whole textural measurements from a  
> complete particle - i.e. not selective parts. However, I would like  
> to know:
>
> Would it be possible to draw a line across, perhaps, the longest  
> diameter and measure textural (and gray-level) distributions across  
> that line?
>
> Would this also be possible across concentric circles drawn through  
> our particles?
>
> Does this make imaging sense?
>
> Many thanks, Andy