counting cells

[thuraya]

Hello everybody,

i need help how to count cells with the dark purple color that exist in the
big circle area
please help me if you can

here is the link for the image,
http://imageshack.us/photo/my-images/834/46609221.jpg/

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

[dscho]

Hi Thuraya,

On Mon, 10 Dec 2012, Thuraya Medhat wrote:

> i need help how to count cells with the dark purple color that exist in the
> big circle area

It looks like you want to use the Colour Deconvolution plugin of Gabriel
Landini:

        http://www.dentistry.bham.ac.uk/landinig/software/software.html

Then you will need the Particle Analyzer:

        http://mirror.imagej.net/docs/guide/146-30.html#toc-Subsection-30.2

> please help me if you can

Sure, that is what this mailing list is for. But you will find that your
studies will require a good deal of knowledge about what you are doing.

> here is the link for the image,
> http://imageshack.us/photo/my-images/834/46609221.jpg/

In favor of my point above, your images are color images and if you do not
calibrate your camera correctly, the deconvolution results will be ever so
slightly off.

So there is no way around studying a lot about scientific image
processing, I am afraid.

Ciao,
Johannes

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

[ctrueden]

Hi Thuraya,

> i need help how to count cells with the dark purple color that exist
> in the big circle area

Did you try Fiji's Color Threshold plugin? I tried it with your sample
image and got pretty good results with the following settings:

Hue: 186 - 216
Saturation: 0 - 255
Brightness: 0 - 180

You can then create a selection & mask and manipulate them in a variety of
ways:

- Click the Select button in the Threshold Color dialog to create a
selection.

- Run the following plugins (press L for command finder):
 -- Create Mask
 -- Fill Holes (on the binary image)
 -- Watershed (chops up blobs into convex shapes/cells)
 -- Analyze Particles... (on the binary image, same options as above)
  --- Size: 10-Infinity
  --- Display results
  --- Clear results
  --- Add to Manager
  --- Exclude on edges

Now you have your cells in the ROI Manager, so you can:

- Close Mask window (Don't save)
- Close Results window
- Finally, in ROI Manager window, check Show All and Labels boxes

Your original image will show the cell outlines. To measure them, run:

- Run "Set Measurements..." (to configure which statistics are calculated)
- Click "Measure" button in ROI Manager

Now your measurements will be based on the original data, using the ROIs
(cells) from the ROI Manager.

Here is a screenshot of the cells I segmented using the above procedure:

    http://code.imagej.net/curtis/fiji-particles.jpg

That was just a quick attempt (e.g. a few of the cells are fused, and there
are some overly-small cells). You can do better by tweaking further with
the Threshold Color settings, and using a bigger minimum Size (perhaps
"25-Infinity"?) in the Analyze Particles dialog.

You could also try CellProfiler (http://cellprofiler.org/); it has many
good techniques for this sort of thing.

Regards,
Curtis


On Mon, Dec 10, 2012 at 5:53 AM, Thuraya Medhat <[hidden email]>wrote:

> Hello everybody,
>
> i need help how to count cells with the dark purple color that exist in the
> big circle area
> please help me if you can
>
> here is the link for the image,
> http://imageshack.us/photo/my-images/834/46609221.jpg/
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

[thuraya]

Dear curties,
you realy did a WONDERFUL job,, just need one thing you told me to do
selection before i do the mask option
but that doesn't work with me could you please do a screen shot about  what
you mean when you said "- Click the Select button in the Threshold Color
dialog to create a selection''

On Mon, Dec 10, 2012 at 8:41 PM, Curtis Rueden <[hidden email]> wrote:

> - Click the Select button in the Threshold Color dialog to create a
> selection
>

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

[ctrueden]

Hi Thuraya,

> could you please do a screen shot about  what you mean when you said
> "- Click the Select button in the Threshold Color dialog to create a
> selection''

Look again at the screenshot I already linked:

    http://code.imagej.net/curtis/fiji-particles.jpg

You see the dialog box entitled "Threshold Color"? There is a button at the
bottom called "Select", which is what I was referring to. This button
creates a selection based on the defined threshold, which you can then
convert to a mask, etc.

-Curtis


On Mon, Dec 10, 2012 at 2:56 PM, Thuraya Medhat <[hidden email]>wrote:

> Dear curties,
> you realy did a WONDERFUL job,, just need one thing you told me to do
> selection before i do the mask option
> but that doesn't work with me could you please do a screen shot about  what
> you mean when you said "- Click the Select button in the Threshold Color
> dialog to create a selection''
>
> On Mon, Dec 10, 2012 at 8:41 PM, Curtis Rueden <[hidden email]> wrote:
>
> > - Click the Select button in the Threshold Color dialog to create a
> > selection
> >
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

... [show rest of quote]

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

[thuraya]

>
> it's me again um realy sorry for bothering but while um doing the order''
> fill holes'' this is the image i got idunnu wt's gone wrong
>


> http://img10.imageshack.us/img10/3083/14615797.png
>

>
> On Mon, Dec 10, 2012 at 11:50 PM, Curtis Rueden <[hidden email]> wrote:
>
>> Hi Thuraya,
>>
>> > could you please do a screen shot about  what you mean when you said
>> > "- Click the Select button in the Threshold Color dialog to create a
>> > selection''
>>
>> Look again at the screenshot I already linked:
>>
>>     http://code.imagej.net/curtis/fiji-particles.jpg
>>
>> You see the dialog box entitled "Threshold Color"? There is a button at
>> the
>> bottom called "Select", which is what I was referring to. This button
>> creates a selection based on the defined threshold, which you can then
>> convert to a mask, etc.
>>
>> -Curtis
>>
>>
>> On Mon, Dec 10, 2012 at 2:56 PM, Thuraya Medhat <[hidden email]
>> >wrote:
>>
>> > Dear curties,
>> > you realy did a WONDERFUL job,, just need one thing you told me to do
>> > selection before i do the mask option
>> > but that doesn't work with me could you please do a screen shot about
>>  what
>> > you mean when you said "- Click the Select button in the Threshold Color
>> > dialog to create a selection''
>> >
>> > On Mon, Dec 10, 2012 at 8:41 PM, Curtis Rueden <[hidden email]>
>> wrote:
>> >
>> > > - Click the Select button in the Threshold Color dialog to create a
>> > > selection
>> > >
>> >
>> > --
>> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>> >
>>
>> --
>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>
>
>

... [show rest of quote]

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

[ctrueden]

Hi Thuraya,

> while um doing the order "fill holes" this is the image i got

I was able to essentially duplicate your (erroneous) result by first
running the "Invert LUT" command before running Fill Holes. So it seems the
Fill Holes command is sensitive to the LUT somehow.

Is your Fiji fully up to date? (Help > Update Fiji)

When I run the "Create Mask" command with a fully up to date Fiji on OS X,
I receive a Mask window with "8-bit (inverting LUT)" in the title area. I
notice in your screenshot that your LUT is not inverted like that. So I
suspect your error has something to do with the LUT. Are you inverting it
manually? Or do you have the "Use inverting lookup table" option in Edit >
Options > Appearance checked? (That option is off for me.)

If you are stumped, you can probably avoid the issue by running "Invert
LUT" followed by "Fill Holes" followed by "Invert LUT" again. Though
ideally it would be best to learn why the problem is happening, of course.

Regards,
Curtis

P.S. Note that it is also possible for the "Fill Holes" command to behave
unexpectedly if you have many dense particles that are all connected in a
large structure. The command interprets it as one giant structure which
needs filling in. So if you used a substantially wider Brightness setting
than 0-180 with your image, you may experience this phenomenon as well. In
short, try to get your particles well-separated when possible.


On Wed, Feb 20, 2013 at 12:06 PM, Thuraya Medhat
<[hidden email]>wrote:

> >
> > it's me again um realy sorry for bothering but while um doing the order''
> > fill holes'' this is the image i got idunnu wt's gone wrong
> >
>
>
> > http://img10.imageshack.us/img10/3083/14615797.png
> >
>
>
> >
> > On Mon, Dec 10, 2012 at 11:50 PM, Curtis Rueden <[hidden email]>
> wrote:
> >
> >> Hi Thuraya,
> >>
> >> > could you please do a screen shot about  what you mean when you said
> >> > "- Click the Select button in the Threshold Color dialog to create a
> >> > selection''
> >>
> >> Look again at the screenshot I already linked:
> >>
> >>     http://code.imagej.net/curtis/fiji-particles.jpg
> >>
> >> You see the dialog box entitled "Threshold Color"? There is a button at
> >> the
> >> bottom called "Select", which is what I was referring to. This button
> >> creates a selection based on the defined threshold, which you can then
> >> convert to a mask, etc.
> >>
> >> -Curtis
> >>
> >>
> >> On Mon, Dec 10, 2012 at 2:56 PM, Thuraya Medhat <
> [hidden email]
> >> >wrote:
> >>
> >> > Dear curties,
> >> > you realy did a WONDERFUL job,, just need one thing you told me to do
> >> > selection before i do the mask option
> >> > but that doesn't work with me could you please do a screen shot about
> >>  what
> >> > you mean when you said "- Click the Select button in the Threshold
> Color
> >> > dialog to create a selection''
> >> >
> >> > On Mon, Dec 10, 2012 at 8:41 PM, Curtis Rueden <[hidden email]>
> >> wrote:
> >> >
> >> > > - Click the Select button in the Threshold Color dialog to create a
> >> > > selection
> >> > >
> >> >
> >> > --
> >> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> >> >
> >>
> >> --
> >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> >>
> >
> >
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

... [show rest of quote]

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

[thuraya]

many many many mannnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnny thanksssssss

On Wed, Feb 20, 2013 at 8:58 PM, Curtis Rueden <[hidden email]> wrote:

> Hi Thuraya,
>
> > while um doing the order "fill holes" this is the image i got
>
> I was able to essentially duplicate your (erroneous) result by first
> running the "Invert LUT" command before running Fill Holes. So it seems the
> Fill Holes command is sensitive to the LUT somehow.
>
> Is your Fiji fully up to date? (Help > Update Fiji)
>
> When I run the "Create Mask" command with a fully up to date Fiji on OS X,
> I receive a Mask window with "8-bit (inverting LUT)" in the title area. I
> notice in your screenshot that your LUT is not inverted like that. So I
> suspect your error has something to do with the LUT. Are you inverting it
> manually? Or do you have the "Use inverting lookup table" option in Edit >
> Options > Appearance checked? (That option is off for me.)
>
> If you are stumped, you can probably avoid the issue by running "Invert
> LUT" followed by "Fill Holes" followed by "Invert LUT" again. Though
> ideally it would be best to learn why the problem is happening, of course.
>
> Regards,
> Curtis
>
> P.S. Note that it is also possible for the "Fill Holes" command to behave
> unexpectedly if you have many dense particles that are all connected in a
> large structure. The command interprets it as one giant structure which
> needs filling in. So if you used a substantially wider Brightness setting
> than 0-180 with your image, you may experience this phenomenon as well. In
> short, try to get your particles well-separated when possible.
>
>
> On Wed, Feb 20, 2013 at 12:06 PM, Thuraya Medhat
> <[hidden email]>wrote:
>
> > >
> > > it's me again um realy sorry for bothering but while um doing the
> order''
> > > fill holes'' this is the image i got idunnu wt's gone wrong
> > >
> >
> >
> > > http://img10.imageshack.us/img10/3083/14615797.png
> > >
> >
> >
> > >
> > > On Mon, Dec 10, 2012 at 11:50 PM, Curtis Rueden <[hidden email]>
> > wrote:
> > >
> > >> Hi Thuraya,
> > >>
> > >> > could you please do a screen shot about  what you mean when you said
> > >> > "- Click the Select button in the Threshold Color dialog to create a
> > >> > selection''
> > >>
> > >> Look again at the screenshot I already linked:
> > >>
> > >>     http://code.imagej.net/curtis/fiji-particles.jpg
> > >>
> > >> You see the dialog box entitled "Threshold Color"? There is a button
> at
> > >> the
> > >> bottom called "Select", which is what I was referring to. This button
> > >> creates a selection based on the defined threshold, which you can then
> > >> convert to a mask, etc.
> > >>
> > >> -Curtis
> > >>
> > >>
> > >> On Mon, Dec 10, 2012 at 2:56 PM, Thuraya Medhat <
> > [hidden email]
> > >> >wrote:
> > >>
> > >> > Dear curties,
> > >> > you realy did a WONDERFUL job,, just need one thing you told me to
> do
> > >> > selection before i do the mask option
> > >> > but that doesn't work with me could you please do a screen shot
> about
> > >>  what
> > >> > you mean when you said "- Click the Select button in the Threshold
> > Color
> > >> > dialog to create a selection''
> > >> >
> > >> > On Mon, Dec 10, 2012 at 8:41 PM, Curtis Rueden <[hidden email]>
> > >> wrote:
> > >> >
> > >> > > - Click the Select button in the Threshold Color dialog to create
> a
> > >> > > selection
> > >> > >
> > >> >
> > >> > --
> > >> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> > >> >
> > >>
> > >> --
> > >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> > >>
> > >
> > >
> >
> > --
> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> >
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

... [show rest of quote]

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