count nuclei in low mag image

[katrina]

Hello all,

I am trying to count the number of cells (white dots) and detect the x,y coordinates of each one in this image.  I've tried using Fiji to do this but I can't detect all of the individual cells very well.  This is what I've tried:

1. Process>>Filter>>Gaussian Blur to run a Gaussian Blur Filter (I've also tried without running this filter)
2. Image>>Adjust>>Threshold>>Select Dark Background to separate the cells from the background
3. Process>>Binary>>Watershed to try to separate touching cells, but it doesn't separate larger clumps of cells very well as shown below.
   

Can I do any additional processing so that I can identify individual cells better?  I don't need to preserve the size of the cells, just their position and the total number of cells.  

I would greatly appreciate help with this!!!!  Thank you!

Katrina

[Adam Hughes]

You might want to try an adaptive threshold.  Not 100% how to do it in
imageJ actually, but here's an example:

http://scikit-image.org/docs/dev/auto_examples/plot_threshold_adaptive.html

I'd expect this would clean up the contrast differences which are causing
the larger images to be over-adjusted.


On Thu, Nov 21, 2013 at 9:16 PM, katrina <[hidden email]> wrote:

> <http://imagej.1557.x6.nabble.com/file/n5005647/Cocx_1_2.gif>
> Hello all,
>
> I am trying to count the number of cells (white dots) and detect the x,y
> coordinates of each one in this image.  I've tried using Fiji to do this
> but
> I can't detect all of the individual cells very well.  This is what I've
> tried:
>
> 1. Process>>Filter>>Gaussian Blur to run a Gaussian Blur Filter (I've also
> tried without running this filter)
> 2. Image>>Adjust>>Threshold>>Select Dark Background to separate the cells
> from the background
> 3. Process>>Binary>>Watershed to try to separate touching cells, but it
> doesn't separate larger clumps of cells very well as shown below.
> <http://imagej.1557.x6.nabble.com/file/n5005647/Cocx_1_2_segmented.gif>
>
> Can I do any additional processing so that I can identify individual cells
> better?  I don't need to preserve the size of the cells, just their
> position
> and the total number of cells.
>
> I would greatly appreciate help with this!!!!  Thank you!
>
> Katrina
>
>
>
>
> --
> View this message in context:
> http://imagej.1557.x6.nabble.com/count-nuclei-in-low-mag-image-tp5005647.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

[Jacqueline Ross]

Hi Katrina,

If you only need to know how many cells there are and where they are (but not size, intensity, etc.), then I would do this by using Process - Find Maxima. You can get the coordinates for each maxima plus total number of cells/nuclei. You can play around with the Noise tolerance if needed to make sure that you are getting the right selection and Preview point selection. You can add the selections to the ROI Manager and save them if you want to do that. You have choices for the Output type as well.

I assume that your original image is larger than this but even this one worked well for me.

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 1142, NEW ZEALAND

Tel: [phone]
Fax: [phone]

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


-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of katrina
Sent: Friday, 22 November 2013 3:17 p.m.
To: [hidden email]
Subject: count nuclei in low mag image

<http://imagej.1557.x6.nabble.com/file/n5005647/Cocx_1_2.gif>
Hello all,

I am trying to count the number of cells (white dots) and detect the x,y coordinates of each one in this image.  I've tried using Fiji to do this but I can't detect all of the individual cells very well.  This is what I've
tried:

1. Process>>Filter>>Gaussian Blur to run a Gaussian Blur Filter (I've also tried without running this filter) 2. Image>>Adjust>>Threshold>>Select Dark Background to separate the cells from the background 3. Process>>Binary>>Watershed to try to separate touching cells, but it doesn't separate larger clumps of cells very well as shown below.
<http://imagej.1557.x6.nabble.com/file/n5005647/Cocx_1_2_segmented.gif>    

Can I do any additional processing so that I can identify individual cells better?  I don't need to preserve the size of the cells, just their position and the total number of cells.  

I would greatly appreciate help with this!!!!  Thank you!

Katrina




--
View this message in context: http://imagej.1557.x6.nabble.com/count-nuclei-in-low-mag-image-tp5005647.html
Sent from the ImageJ mailing list archive at Nabble.com.

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

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

[chin]

Hi Katrina,

You may use the  ITCN (Image-based Tool for Counting Nuclei) plugin.

http://www.bioimage.ucsb.edu/automatic-nuclei-counter-plug-in-for-imagej

Try the "width" from 6-8 pixels,  "threshold" from 0.3-0.5, you could get various images with
detected cells.

chin