Segmentation of oligodendrocyte (cell) staining - DAB

[Jacqueline Ross]

Dear All,

 

One of the graduate students here needs to count the number of
oligodendrocytes stained in his sections. The staining is
immunohistochemistry with DAB substrate.

 

This labelling not only stains the area around the cell but also the
cell processes. However, he just needs to know how many cells are
stained not the area occupied by fibres, etc.

 

He has a large number of images to analyse so we have been trying to see
if we can segment out the cell body staining which looks like a ring
around the cell in order to then batch-process.

 

I have tried a few things with varied success and would really value
some assistance on this. At the moment, my conclusion is that some
processing to enhance the image with subsequent manual counting may be
the only solution.

 

I have put 3 examples on our website here:
http://www.fmhs.auckland.ac.nz/sms/biru/links/imagej.aspx which vary
from easy to hard!

 

There are 3 cropped images with the full sized image also available to
download below. The images do look blue (probably lamp temperature
issue) but the staining is reasonably clear.

 

I have put arrowheads on one of the images to indicate what the student
needs to measure.

 

All suggestions welcome!

 

Kind regards,

 

Jacqui

 

Jacqueline Ross

Biomedical Imaging Microscopist
Biomedical Imaging Research Unit
School of Medical Sciences
Faculty of Medical & Health Sciences
The University of Auckland
Private Bag 92019
Auckland, NEW ZEALAND

Tel: [phone] Ext 87438
Fax: [phone]

http://www.fmhs.auckland.ac.nz/sms/biru/
<http://www.fmhs.auckland.ac.nz/sms/biru/>

[dpoburko]

Hi Jacqui,

  Have a look the Houch Circles plugin (
http://rsb.info.nih.gov/ij/plugins/hough-circles.html ). With some
patience trying to the appropriate parameters, this approach should
probably be able to identify the rings around the oligodendrocytes. Your
student will have to threshold their images first. They might be able to
reduce some of the non-ring staining with Bandpass filters of various
sizes, or by using the Subtract Background function built into IJ.

Hope that get you somewhere.

Cheers,
Damon



--

Damon Poburko, PhD
Postdoctoral Research Fellow
Stanford University School of Medicine
Dept. of Molecular & Cellular Physiology
279 Campus Dr., Beckman B103, Stanford, CA 94305
Ph: 650 725 7564, fax: [phone]




Jacqui Ross wrote:

> Dear All,
>
>  
>
> One of the graduate students here needs to count the number of
> oligodendrocytes stained in his sections. The staining is
> immunohistochemistry with DAB substrate.
>
>  
>
> This labelling not only stains the area around the cell but also the
> cell processes. However, he just needs to know how many cells are
> stained not the area occupied by fibres, etc.
>
>  
>
> He has a large number of images to analyse so we have been trying to see
> if we can segment out the cell body staining which looks like a ring
> around the cell in order to then batch-process.
>
>  
>
> I have tried a few things with varied success and would really value
> some assistance on this. At the moment, my conclusion is that some
> processing to enhance the image with subsequent manual counting may be
> the only solution.
>
>  
>
> I have put 3 examples on our website here:
> http://www.fmhs.auckland.ac.nz/sms/biru/links/imagej.aspx which vary
> from easy to hard!
>
>  
>
> There are 3 cropped images with the full sized image also available to
> download below. The images do look blue (probably lamp temperature
> issue) but the staining is reasonably clear.
>
>  
>
> I have put arrowheads on one of the images to indicate what the student
> needs to measure.
>
>  
>
> All suggestions welcome!
>
>  
>
> Kind regards,
>
>  
>
> Jacqui
>
>  
>
> Jacqueline Ross
>
> Biomedical Imaging Microscopist
> Biomedical Imaging Research Unit
> School of Medical Sciences
> Faculty of Medical & Health Sciences
> The University of Auckland
> Private Bag 92019
> Auckland, NEW ZEALAND
>
> Tel: [phone] Ext 87438
> Fax: [phone]
>
> http://www.fmhs.auckland.ac.nz/sms/biru/
> <http://www.fmhs.auckland.ac.nz/sms/biru/>
>  

[vbindokas]

Dear Jacqui,
     Try using the Template Matching plugin,
http://rsb.info.nih.gov/ij/plugins/template-matching.html .  You can
draw a ROI around a good example, then use the 'crop template from
image' or just use the 'create circular template' routine. Define
thresholds and sizes to match the range you want, then press the
"perform statstica correlation' button. It will output a correlation
image that encodes how all features matched your template. Threshold
this to something like 0.6-1.0 (60-100% match range], and count those  
[analyze particles or use binary_find maxima for similar tolerance
range]. This plugin converts data to 32bit [b/w], so you may want to
tweak color settings or do it yourself vs automatic mode.
     It seemed to do a reasonable job.
regards,

On 2/10/2010 11:43 PM, Jacqui Ross wrote:

> Dear All,
>
>
>
> One of the graduate students here needs to count the number of
> oligodendrocytes stained in his sections. The staining is
> immunohistochemistry with DAB substrate.
>
>
>
> This labelling not only stains the area around the cell but also the
> cell processes. However, he just needs to know how many cells are
> stained not the area occupied by fibres, etc.
>
>
>
> He has a large number of images to analyse so we have been trying to see
> if we can segment out the cell body staining which looks like a ring
> around the cell in order to then batch-process.
>
>
>
> I have tried a few things with varied success and would really value
> some assistance on this. At the moment, my conclusion is that some
> processing to enhance the image with subsequent manual counting may be
> the only solution.
>
>
>
> I have put 3 examples on our website here:
> http://www.fmhs.auckland.ac.nz/sms/biru/links/imagej.aspx which vary
> from easy to hard!
>
>
>
> There are 3 cropped images with the full sized image also available to
> download below. The images do look blue (probably lamp temperature
> issue) but the staining is reasonably clear.
>
>
>
> I have put arrowheads on one of the images to indicate what the student
> needs to measure.
>
>
>
> All suggestions welcome!
>
>
>
> Kind regards,
>
>
>
> Jacqui
>
>
>
> Jacqueline Ross
>
> Biomedical Imaging Microscopist
> Biomedical Imaging Research Unit
> School of Medical Sciences
> Faculty of Medical&  Health Sciences
> The University of Auckland
> Private Bag 92019
> Auckland, NEW ZEALAND
>
> Tel: [phone] Ext 87438
> Fax: [phone]
>
> http://www.fmhs.auckland.ac.nz/sms/biru/
> <http://www.fmhs.auckland.ac.nz/sms/biru/>
>    

--
__

Vytas Bindokas, Ph.D.
Research Assoc. / Assoc. Prof.,
Director, BSD Light Microscopy Core Facility
Dept Neurobiol Pharmacol Physiol MC0926
947 E 58th Street
The University of Chicago
Chicago IL 60637
Room Abbott 129
773-702-4875

email [hidden email]
web site for LMCF:
http://digital.uchicago.edu/index.html

[Joel Sheffield]

---------- Forwarded message ----------
From: JOEL B. SHEFFIELD <[hidden email]>
Date: Thu, Feb 11, 2010 at 10:06 AM
Subject: Re: Segmentation of oligodendrocyte (cell) staining - DAB
To: Jacqui Ross <[hidden email]>


Hi Jacqui

I suppose that it depends on how much error you can tolerate, but here is a
relatively simple approach.  The newer versions of IJ include an
experimental threshold routine, in the image>adjust menu, that allows an
enormous range of selections.  I chose to use the saturation parameter in
HSB imaging.  When I did this, I was able to crudely threshold the cell
bodies, with some scattered bits of other material.  I then converted the
RGB image, with the threshold to 8-bit, so that I could use the analyze
particles routine to count the cells.  I restricted the routine to just
count cells larger than a given value.  I through the I might be able to use
circularity as a second parameter, but the images were actually too
irregular.  I ticked "include holes", as well.


Joel



On Thu, Feb 11, 2010 at 12:43 AM, Jacqui Ross <[hidden email]>wrote:

> Dear All,
>
>
>
> One of the graduate students here needs to count the number of
> oligodendrocytes stained in his sections. The staining is
> immunohistochemistry with DAB substrate.
>
>
>
> This labelling not only stains the area around the cell but also the
> cell processes. However, he just needs to know how many cells are
> stained not the area occupied by fibres, etc.
>
>
>
> He has a large number of images to analyse so we have been trying to see
> if we can segment out the cell body staining which looks like a ring
> around the cell in order to then batch-process.
>
>
>
> I have tried a few things with varied success and would really value
> some assistance on this. At the moment, my conclusion is that some
> processing to enhance the image with subsequent manual counting may be
> the only solution.
>
>
>
> I have put 3 examples on our website here:
> http://www.fmhs.auckland.ac.nz/sms/biru/links/imagej.aspx which vary
> from easy to hard!
>
>
>
> There are 3 cropped images with the full sized image also available to
> download below. The images do look blue (probably lamp temperature
> issue) but the staining is reasonably clear.
>
>
>
> I have put arrowheads on one of the images to indicate what the student
> needs to measure.
>
>
>
> All suggestions welcome!
>
>
>
> Kind regards,
>
>
>
> Jacqui
>
>
>
> Jacqueline Ross
>
> Biomedical Imaging Microscopist
> Biomedical Imaging Research Unit
> School of Medical Sciences
> Faculty of Medical & Health Sciences
> The University of Auckland
> Private Bag 92019
> Auckland, NEW ZEALAND
>
> Tel: [phone] Ext 87438
> Fax: [phone]
>
> http://www.fmhs.auckland.ac.nz/sms/biru/
> <http://www.fmhs.auckland.ac.nz/sms/biru/>
>

--


Joel B. Sheffield, Ph.D
Department of Biology
Temple University
Philadelphia, PA 19122
Voice: 215 204 8839
e-mail: [hidden email]
URL:  http://astro.temple.edu/~jbs <http://astro.temple.edu/%7Ejbs>




--


Joel B. Sheffield, Ph.D
Department of Biology
Temple University
Philadelphia, PA 19122
Voice: 215 204 8839
e-mail: [hidden email]
URL:  http://astro.temple.edu/~jbs