Re: threshholding nuclei

Posted by Joel Sheffield on Jan 16, 2011; 10:50pm
URL: http://imagej.273.s1.nabble.com/threshholding-nuclei-tp3685280p3685284.html

Assuming that your original images are RGB, you might consider separating
the color channels and then carrying out the analysis.  The other issue that
is often a problem is whether you have even illumination across the field.
It might help if you could post a link to some of the images.

Joel


On Sun, Jan 16, 2011 at 2:45 PM, Ralph Common <[hidden email]> wrote:

> It is often best not to use a counterstain when you need to threshold a
> specifically stained feature.  You can have duplicate slides made, one set
> with counterstain, one without.  If this is no longer possible you can try
> to eliminate the affect of the eosin by using a red filter and photographing
> your slides in monochrome mode.
>
> Ralph Common
>
>
> On 1/16/2011 2:13 PM, nwoller wrote:
>
>> Hi,
>>
>> I am a new ImageJ user. I have taken a series of images of a slide with
>> tonsil tissue which has been stained with hematoxylin as well as liquid
>> permanent red (Immunostain). I have been careful to capture all of the
>> images in exactly the same way, same light intensity, same exposure time,
>> same white balance etc. I would like to count the number of
>> hematoxylin-stained nuclei in my images. I threshold the image, apply a
>> watershed and use "analyze particles". Unfortunately, I don't seem to be
>> able to apply the same thresholds to my images. A threshold that works
>> nicely for one image is completely hopeless for another image. It is
>> possible that the intensity of the hematoxylin stain varies quite alot in
>> the same tissue. A colleague therefore suggested that I should threshold
>> by
>> finding the top of the "threshold curve" and using that value for
>> thresholding my images. What do you think? I would really appreciate some
>> help and advice.
>>
>> Nina
>>
>

--


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