Nuclei detection - how can I expand my ROI for the cell membrane?

[Robert Niestroj-Pahl]

Hi folks,
I want to analyse fluorescence probes inside of cells. I have images in
DAPI, FITC and CY5.
I had success with nuclei detection by thresholding and the particle
analysis. Then I save my particle analysis result as an ROI to use it with
the other channels and to count the fluorescence particles. But I want to
count the fluorescence particles not only in the nuclei but also in the
cytoplasm. Is it possible to expand the ROI around each nuclei for lets say
5 or 10 pixels to also measure particles in the cytoplasm of the cells?
Any ideas?

Many greetings

Robert



--
Sent from: http://imagej.1557.x6.nabble.com/

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

[Krs5]

Dear Robert,

Yes, see Edit > Selection > Enlarge

Best wishes

Kees


Dr Ir K.R. Straatman
Senior Experimental Officer
Advanced Imaging Facility
Centre for Core Biotechnology Services
University of Leicester
www.le.ac.uk/advanced-imaging-facility



-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Robert Niestroj-Pahl
Sent: 16 April 2018 13:11
To: [hidden email]
Subject: Nuclei detection - how can I expand my ROI for the cell membrane?

Hi folks,
I want to analyse fluorescence probes inside of cells. I have images in DAPI, FITC and CY5.
I had success with nuclei detection by thresholding and the particle analysis. Then I save my particle analysis result as an ROI to use it with the other channels and to count the fluorescence particles. But I want to count the fluorescence particles not only in the nuclei but also in the cytoplasm. Is it possible to expand the ROI around each nuclei for lets say
5 or 10 pixels to also measure particles in the cytoplasm of the cells?
Any ideas?

Many greetings

Robert



--
Sent from: http://imagej.1557.x6.nabble.com/

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

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

[Herbie]

Good day Robert,

you may have a look at

"Edit >> Selection >> Enlarge..."

and it is always a good idea to study the ImageJ user guide:
<https://imagej.nih.gov/ij/docs/guide/index.html>

Regards

Herbie

::::::::::::::::::::::::::::::::::::::::::::::::::
Am 16.04.18 um 14:11 schrieb Robert Niestroj-Pahl:

> Hi folks,
> I want to analyse fluorescence probes inside of cells. I have images in
> DAPI, FITC and CY5.
> I had success with nuclei detection by thresholding and the particle
> analysis. Then I save my particle analysis result as an ROI to use it with
> the other channels and to count the fluorescence particles. But I want to
> count the fluorescence particles not only in the nuclei but also in the
> cytoplasm. Is it possible to expand the ROI around each nuclei for lets say
> 5 or 10 pixels to also measure particles in the cytoplasm of the cells?
> Any ideas?
>
> Many greetings
>
> Robert
>
>
>
> --
> Sent from: http://imagej.1557.x6.nabble.com/
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

[Robert Niestroj-Pahl]

Thanks Herbie,

I found this command. My selection of the nuclei is now enlarged, but the
watershed segmentation is gone! I can't count the single cells any more. Do
you know a way how to keep the nuclei segmentation?
Best regards

Robert



--
Sent from: http://imagej.1557.x6.nabble.com/

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

[G. Esteban Fernandez]

When you do Analyze Particles, select the option to "Add to Manager"
to add each individual ROI to the list in the ROI Manager; that will
keep them separate. Then you can loop the enlargement and further
analysis steps through each ROI in the list, e.g.:

numROIs = roiManager("count");
for(i = 0; i < numROIs; i++){
     roiManager("Select", i);
     run("Enlarge...", "enlarge=5"); //5 pixel enlargement
     roiManager("Add");
}

//now loop through enlarged ROIs only to do analysis


-Esteban



On Mon, Apr 16, 2018, 5:33 AM Robert Niestroj-Pahl
<[hidden email]> wrote:

>
> Thanks Herbie,
>
> I found this command. My selection of the nuclei is now enlarged, but the
> watershed segmentation is gone! I can't count the single cells any more. Do
> you know a way how to keep the nuclei segmentation?
> Best regards
>
> Robert
>
>
>
> --
> Sent from: http://imagej.1557.x6.nabble.com/
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

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

[Michael Schmid]

Hi Robert,

if you create the selection to a mask, you could try Gabriel Landini's
BinaryDilateNoMerge8 or on the mask, then convert it back to a selection.

http://www.mecourse.com/landinig/software/software.html

Unfortunately, it does dilation with a square structure element, not a
circular one.


An alternative, starting with a binary image, with circular dilation:

//assumes you have white background, black particles:
run("Options...", "iterations=1 count=1 pad do=Nothing");
//the 'watershed' segmentation you mentioned
run("Watershed");
//and now the operations for a no-merge dilation
run("Invert");
run("Distance Map");
run("Find Maxima...", "noise=0 output=[Segmented Particles]");
run("Create Selection");
close();
run("Restore Selection");
run("Make Inverse");
run("Set...", "value=255");
setThreshold(0, 5); // 5 pixels dilation (max. 254)
run("Convert to Mask");



Michael
________________________________________________________________
On 16/04/2018 14:32, Robert Niestroj-Pahl wrote:
> Thanks Herbie,
>
> I found this command. My selection of the nuclei is now enlarged, but the
> watershed segmentation is gone! I can't count the single cells any more. Do
> you know a way how to keep the nuclei segmentation?
> Best regards
>
> Robert

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

[Robert Niestroj-Pahl]

Hi Esteban,

very elegant solution thank you! But now I have overlapping ROI. There might
be the chance of counting particles twice. Do you also know a workaround for
that?

Best wishes

Robert



--
Sent from: http://imagej.1557.x6.nabble.com/

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

[Robert Niestroj-Pahl]

Hi Michael,
thanks for the code, for me very appealing because it produces no
overlapping ROI (like Estebans proposal). But the macro only works from time
to time... sometimes I get only a white Image and sometimes I get
skeletonized structures, only in one of three or four runs it works. Do you
know what can cause the problems? I tried to clear my ROI manager first, but
that seems not to be the problem.

Best wishes

Robert



--
Sent from: http://imagej.1557.x6.nabble.com/

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

[Michael Schmid]

Hi Robert,

could you post a test image that does not work for you? (or upload it to
some site)
Then I can try to find out what is wrong.

By the way, I don't use the ROI Manager in my macro.


Michael
________________________________________________________________
On 17/04/2018 10:31, Robert Niestroj-Pahl wrote:

> Hi Michael,
> thanks for the code, for me very appealing because it produces no
> overlapping ROI (like Estebans proposal). But the macro only works from time
> to time... sometimes I get only a white Image and sometimes I get
> skeletonized structures, only in one of three or four runs it works. Do you
> know what can cause the problems? I tried to clear my ROI manager first, but
> that seems not to be the problem.
>
> Best wishes
>
> Robert

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

[Robert Niestroj-Pahl]

Hi Micheal,

thanks for help. See my screenshot below
<http://imagej.1557.x6.nabble.com/file/t381965/Screen_Shot_2018-04-17_at_10.png>

That is the Image I tried you macro on:
<http://imagej.1557.x6.nabble.com/file/t381965/S5104_MP_IMage_1_Dapi_500_ms_1g_20x.jpg>

Robert



--
Sent from: http://imagej.1557.x6.nabble.com/

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

[Michael Schmid]

Hi Robert,

ok, I see, the 'Restore selection' sometimes fails.
I could solve the problem by adding a "Select none", which makes sure
the selection is saved as "previousRoi".
I hope that it works reliably now...

//assumes you have white background, black particles:
run("Options...", "iterations=1 count=1 pad do=Nothing");
//the 'watershed' segmentation you mentioned
run("Watershed");
//and now the operations for a no-merge dilation
run("Invert");
run("Distance Map");
run("Find Maxima...", "noise=0 output=[Segmented Particles]");
run("Create Selection");
run("Select None");
close();
run("Restore Selection");
run("Make Inverse");
run("Set...", "value=255");
setThreshold(0, 5); // 5 pixels dilation (max. 254)
run("Convert to Mask");


Michael
________________________________________________________________


On 17/04/2018 10:50, Robert Niestroj-Pahl wrote:
> Hi Micheal,
>
> thanks for help. See my screenshot below
> <http://imagej.1557.x6.nabble.com/file/t381965/Screen_Shot_2018-04-17_at_10.png>
>
> That is the Image I tried you macro on:
> <http://imagej.1557.x6.nabble.com/file/t381965/S5104_MP_IMage_1_Dapi_500_ms_1g_20x.jpg>
>
> Robert

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

[Robert Niestroj-Pahl]

Hi Michael,
now it works! Thank you!

Robert



--
Sent from: http://imagej.1557.x6.nabble.com/

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

[Stoyan Pavlov]

Hi Robert,
There is an alternative simpler (and faster) yet not as accurate (in
finding nuclear borders) solution as Landini's "Dilate no merge".
1. Duplicate the segmented nuclei image and do a Voronoi transformation on
the second image (Process->Binary->Voronoi). Then threshold the resulting
image with foreground between 1 and 255. This creates binary image of the
Neighborhood (aka Voronoi's/Dirichlet's cells) of each nucleus: points in
spaces closer to the center to each nucleus are white and correspondingly
the borders between these neighborhoods are black .
2. Second step is to do the desired dilation (i.e. 5 px) on the first image
of segmented nuclei. In the resulting image the nuclei are dilated but
close spaced nuclei fuse.
3.In third step separate the fused dilated nuclei by doing a boolean
conjunction between the two images.: Using Process-> Image calculator do a
boolean "AND" (or you can simply multiply the two images) between the
dilated image and the thresholded voronoi-image . Be sure that the
background (i.e. background between nuclei and the lines between the cells
in the Voronoi-image) in both images is black.
Best wishes,
Stoyan

---
Assoc.Prof. Stoyan P. Pavlov, MD, PhD
Departament of Anatomy, Histology and Embryology
Medical University "Prof. Dr. Paraskev Stoyanov", Varna
Prof. Marin Drinov Str.55
9002 Varna
 Bulgaria
Tel: +359 (0) 52 - 677 - 052
e-mail: [hidden email]
           [hidden email]

2018-04-17 13:42 GMT+03:00 Robert Niestroj-Pahl <[hidden email]>:

> Hi Michael,
> now it works! Thank you!
>
> Robert
>
>
>
> --
> Sent from: http://imagej.1557.x6.nabble.com/
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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