Image J

[Sebastian COSTA]

Hi,
I've got difficulties selecting the brown color from this image using color threshold.
Do you know if there is a special filter for brown (DAB) or do you have a tip ??

thanks
Sebastian COSTA        
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Michael Schmid]

Hi Sebastian,

your image has bad compression artifacts (JPG compression). Assuming these are absent in the original image, I guess that 'Threshold Color' in HSB Color space should work reasonably well.
Select low brightness (e.g. 0-110) and low Hue values (e.g. up top 120).

Michael
________________________________________________________________
On Sep 16, 2013, at 11:24, Sebastian COSTA wrote:

> Hi,
> I've got difficulties selecting the brown color from this image using color threshold.
> Do you know if there is a special filter for brown (DAB) or do you have a tip ??
>
> thanks
> Sebastian COSTA  

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Andrei Stefan]

Hi,

I think that you should try *Colour Deconvolution* plug-in (link
here<http://www.dentistry.bham.ac.uk/landinig/software/cdeconv/cdeconv.html>)
or the *Trainable Weka Segmentation* in FIJI (link
here<http://fiji.sc/Trainable_Weka_Segmentation>
).


Enjoy!

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[carlo bianco]

hello
also color segmentation with manual selection works well.
http://bigwww.epfl.ch/sage/soft/colorsegmentation/
best wishes
carlo-bologna




________________________________
 Da: Sebastian COSTA <[hidden email]>
A: [hidden email]
Inviato: Lunedì 16 Settembre 2013 11:24
Oggetto: Image J
 

Hi,
I've got difficulties selecting the brown color from this image using color threshold.
Do you know if there is a special filter for brown (DAB) or do you have a tip ??

thanks
Sebastian COSTA                             
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[John Dunsmuir]

Sebastian,
  A trick that might help is to split the image into three color channels, then divide the green channel by the red one with 32bit result.  The brown regions should be easier to segment.

Regards,
John D.

 

Sebastian COSTA wrote

Hi,
I've got difficulties selecting the brown color from this image using color threshold.
Do you know if there is a special filter for brown (DAB) or do you have a tip ??

thanks
Sebastian COSTA        
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html


10B.jpg (715K) <http://imagej.1557.x6.nabble.com/attachment/5004795/0/10B.jpg>

[Gabriel Landini]

Sebastian COSTA wrote
> I've got difficulties selecting the brown color from this image using
> color threshold.
> Do you know if there is a special filter for brown (DAB) or do you have a
> tip ??

The majority of cases of histological stains for optical (brightfield)
microscopy, the dyes combine "subtractively", so classifying pixels (eg Colour
Threshold plugin) that contain mixtures of 2 or more dyes without considering
the contribution of each dye separately carries a considerable and
uncontrollable error.

The ideal solution to this problem is colour deconvolution (read the paper by
A. Ruifrok), where you separate the colour image into the original
contribution of each dye.
There is a plugin for this, but note that not all dyes behave as expected
(where optical density is a function of concentration, DAB is one of them) and
most important immunostains are no stoichiometric reactions, so you can't
really associate with any reliability the intensity of stain to amount of
reactant.

Note that you won't get any good results with lossy JPEG images. You also need
background-corrected images for the colour deconvolution to work.

Hope this helps.
Gabriel

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html