Discussion:
4-colour merge?
Shane Harding
2007-11-10 21:25:55 UTC
Permalink
Hi All,
I'm a relatively new ImageJ user and would like to merge 4-colours
into one composite image (CY3, CY5, Alexa488 and DAPI). Is this
possible in ImageJ? How would I do this?

Thank you,
Shane
Dan White
2007-11-11 10:47:14 UTC
Permalink
Dear Shane.

Your eyes have 3 color sensors, red green and blue,
your computer screen had red green and blue pixels
(and in printing its cyan, yellow, magenta and black)

So, now if you want to merge 3 colours,
thats fine, make each colour channel one of red green and blue,
It will look very pretty (hopefully).

Now with 4 colours you are faced with a big problem.
Your eyes are not capable of picking out 4 different colours
out of an RGB image.

As far as your eyes are concerned,
all colours are made of some mixture of red, green and blue.
So if you already have a red, green and blue channel,
then try to add 1 more, what colour would you pick?
Magenta... no, because thats a mix of red and blue.
Similar for cyan and yellow, and other colours.

So, if you do mix red green blue and magenta,
you will end up with a pretty image,
but it will be impossible to tell which parts
of the red part of the image come from
the red channel or from the magenta channel.

Some people use
red, green blue and white,
but you are faced with that same problem .
There cold be a place with almost equal red green and blue intensity
(= white)
which you could not tell from a white pixel

If you are just trying to make a pretty picture,
and not say anything quantitative or too specifically qualitative,
you might get away with RGB+white.
If i reviewed it or saw it in talk,
I would raise the points made above.

If you are trying to look for colocalisation
(or more specifically spatial intensity correlation)
you can do this in pairwise way,
using the imageJ colocalisation plugins
(read the docs at
http://www.macbiophotonics.ca/imagej/
)


Looking at a 2D intensity histogram / scatterplot
of once channel vs a second gives a much better way
of visualising the spatial intensity correlation between then
than a colour merge image.

In my opinion, colour merge images are hard to interpret,
can be misleading, are a poor way of visualising "colocalisation"
and this researcher would send back any paper to an author
who has used colour merge images to try to show
between channel spatial intensity correlations (colocalisation)

In fact I would go as far as to say,
wrong interpretation of colour merge images could
ruin you career.
2D hisotgrams / scatter plots are a much better way to show it.

cheers

Dan
Date: Sat, 10 Nov 2007 16:25:55 -0500
Subject: 4-colour merge?
Hi All,
I'm a relatively new ImageJ user and would like to merge 4-colours
into one composite image (CY3, CY5, Alexa488 and DAPI). Is this
possible in ImageJ? How would I do this?
Thank you,
Shane
Dr. Daniel James White BSc. (Hons.) PhD
Senior Microscopist / Image Processing and Analysis
Light Microscopy Facility
Max Planck Institute of Molecular Cell Biology and Genetics
Pfotenhauerstrasse 108
01307 DRESDEN
Germany

+49 (0)176 61262128 (German Mobile)
+49 (0)351 210 2627 (Work phone at MPI-CBG)
+49 (0)351 3265048 (German Festnetz "landline" - cheaper)
+49 (0)351 210 1078 (Fax MPI-CBG LMF)
+358 (0) 468102840 (Finnish mobile, only when I'm in Finland)
http://www.bioimagexd.net
http://www.chalkie.org.uk
dan-***@public.gmane.org
( white-***@public.gmane.org )
Joachim Wesner
2007-11-11 15:53:30 UTC
Permalink
Hi there,

working with my copy of ImageJ 1.38x i just discover that the macro
language does not accept statements like

a = 5e-3 ;

or
a = 5.e-3;
or
a = 5.0e-3;

(a = 0.005 works, however)

Am I totally confused here? Is that a "feature" that only recently crept
in?


Sincerely


Joachim





______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
______________________________________________________________________
Ben Tupper
2007-11-11 17:41:33 UTC
Permalink
Post by Joachim Wesner
Hi there,
working with my copy of ImageJ 1.38x i just discover that the macro
language does not accept statements like
a = 5e-3 ;
or
a = 5.e-3;
or
a = 5.0e-3;
(a = 0.005 works, however)
Am I totally confused here? Is that a "feature" that only recently crept
in?
Hi,

You could use the pow() function. For example...

print(5 * pow(10,-3));

results in 0.0050

Cheers,
Ben

Ben Tupper
Bigelow Laboratory for Ocean Science
180 McKown Point Road
POB 475
West Boothbay Harbor, ME 04575
Johannes Schindelin
2007-11-11 20:28:02 UTC
Permalink
Hi,
Post by Joachim Wesner
working with my copy of ImageJ 1.38x i just discover that the macro
language does not accept statements like
a = 5e-3 ;
or
a = 5.e-3;
or
a = 5.0e-3;
(a = 0.005 works, however)
Am I totally confused here? Is that a "feature" that only recently crept
in?
No, this is a long standing shortcoming. I just fixed it in ImageJA, and
will send Wayne a patch. If you are interested in the patch, see
http://repo.or.cz/w/imageja.git?a=commitdiff;h=2ff92a2c562bacce89f51a20171006914ba9d615.

Ciao,
Dscho
Joachim Wesner
2007-11-12 10:12:49 UTC
Permalink
Hi,

thanx a lot. It´s not that I could not live without it, it´s just a strange
behaviour that I noted when working on macro.
I think/thought that it should accept exponential format, as the macro
language in all other respect is already very similar to basic Java or C.

Wayne has already supplied me with a new IJ snapshot that has the patch
built in, great.

I wish other software would have such a quick bugfix cycle, excellent!

Sincerely


Joachim





Johannes
Schindelin
<Johannes.Schinde An
lin-***@public.gmane.org> IMAGEJ-9srhZJH3/***@public.gmane.org
Gesendet von: Kopie
ImageJ Interest
Group Thema
<IMAGEJ-9srhZJH3/r/hvxM+***@public.gmane.org Re: Macro language exponential
GOV> number format?


11.11.2007 21:28


Bitte antworten
an
ImageJ Interest
Group
<IMAGEJ-9srhZJH3/r/hvxM+***@public.gmane.org
GOV>






Hi,
Post by Joachim Wesner
working with my copy of ImageJ 1.38x i just discover that the macro
language does not accept statements like
a = 5e-3 ;
or
a = 5.e-3;
or
a = 5.0e-3;
(a = 0.005 works, however)
Am I totally confused here? Is that a "feature" that only recently crept
in?
No, this is a long standing shortcoming. I just fixed it in ImageJA, and
will send Wayne a patch. If you are interested in the patch, see
http://repo.or.cz/w/imageja.git?a=commitdiff;h=2ff92a2c562bacce89f51a20171006914ba9d615
.

Ciao,
Dscho



______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
______________________________________________________________________
Justin McGrath
2007-11-12 15:16:43 UTC
Permalink
Joachim Walter wrote a nice plugin called Image5D. It's found at
http://rsb.info.nih.gov/ij/plugins/image5d.html . You can load your
four channels and pick a color for each. I don't know much about
colocalization or quantifying fluorescence, but I would only use this
to make demonstration images.

Justin
Post by Shane Harding
Hi All,
I'm a relatively new ImageJ user and would like to merge 4-colours
into one composite image (CY3, CY5, Alexa488 and DAPI). Is this
possible in ImageJ? How would I do this?
Thank you,
Shane
Michael Weber
2007-11-12 15:40:49 UTC
Permalink
Shane,

there is a plugin called "Colour Merge" which can merge every colour,
also violet plus yellow and so on.

As already mentioned, such an extensive overlay doesn't make sense for
quantitation. In my opinion one can use it if the structures are labeled
with one colour at once only, i.e. nuclei, membrane, endosomes etc.
separately - to get an overview of the situation in the cell/tissue.

cheers,
Michael
Post by Justin McGrath
Joachim Walter wrote a nice plugin called Image5D. It's found at
http://rsb.info.nih.gov/ij/plugins/image5d.html . You can load your
four channels and pick a color for each. I don't know much about
colocalization or quantifying fluorescence, but I would only use this
to make demonstration images.
Justin
Post by Shane Harding
Hi All,
I'm a relatively new ImageJ user and would like to merge 4-colours
into one composite image (CY3, CY5, Alexa488 and DAPI). Is this
possible in ImageJ? How would I do this?
Thank you,
Shane
Shane Harding
2007-11-12 15:56:08 UTC
Permalink
Thank you all for your help.

I think the solution to my question is to use the DAPI image to create
a boundary in which I can use the RGB labels. Is there an easy way to
create an image with the DAPI regions outlined so that I may overlay
my other 3 RGB coloured on this?

I agree that pairwise will also work, I'm just searching for
different, but acceptable options.

Thanks Again,
Shane
Post by Michael Weber
Shane,
there is a plugin called "Colour Merge" which can merge every colour,
also violet plus yellow and so on.
As already mentioned, such an extensive overlay doesn't make sense for
quantitation. In my opinion one can use it if the structures are
labeled with one colour at once only, i.e. nuclei, membrane, endosomes
etc. separately - to get an overview of the situation in the
cell/tissue.
cheers,
Michael
Post by Justin McGrath
Joachim Walter wrote a nice plugin called Image5D. It's found at
http://rsb.info.nih.gov/ij/plugins/image5d.html . You can load your
four channels and pick a color for each. I don't know much about
colocalization or quantifying fluorescence, but I would only use this
to make demonstration images.
Justin
Post by Shane Harding
Hi All,
I'm a relatively new ImageJ user and would like to merge 4-colours
into one composite image (CY3, CY5, Alexa488 and DAPI). Is this
possible in ImageJ? How would I do this?
Thank you,
Shane
Senger Fabrice
2007-11-12 16:02:52 UTC
Permalink
Post by Justin McGrath
Joachim Walter wrote a nice plugin called Image5D. It's found at
http://rsb.info.nih.gov/ij/plugins/image5d.html . You can load your
four channels and pick a color for each. I don't know much about
colocalization or quantifying fluorescence, but I would only use this
to make demonstration images.
Justin
Post by Shane Harding
Hi All,
I'm a relatively new ImageJ user and would like to merge 4-colours
into one composite image (CY3, CY5, Alexa488 and DAPI). Is this
possible in ImageJ? How would I do this?
Thank you,
Shane
Hi there,

to only make a demo image you could use aswell Photoshop or the freely
available "the gimp" and work around with overlays.

Sheers,

Fabrice.
--
Senger Fabrice
John Alexander
2007-11-12 17:37:15 UTC
Permalink
I have found that a quick way to merge images of non-RGB colors is to
use the Past Control and set it to Add ...



John
Post by Joachim Wesner
Post by Justin McGrath
Joachim Walter wrote a nice plugin called Image5D. It's found at
http://rsb.info.nih.gov/ij/plugins/image5d.html . You can load your
four channels and pick a color for each. I don't know much about
colocalization or quantifying fluorescence, but I would only use this
to make demonstration images.
Justin
Post by Shane Harding
Hi All,
I'm a relatively new ImageJ user and would like to merge 4-colours
into one composite image (CY3, CY5, Alexa488 and DAPI). Is this
possible in ImageJ? How would I do this?
Thank you,
Shane
Hi there,
to only make a demo image you could use aswell Photoshop or the freely
available "the gimp" and work around with overlays.
Sheers,
Fabrice.
--
John K. Alexander, Ph.D.
Post-Doctoral Fellow
William Green Laboratory
University of Chicago
Dept. Neurobiology, Pharmacology, and Physiology
947 East 58th Street
Abott Hall 402
Chicago, IL 60637
(off) 773-702-9386
(fax) 773-702-3774
jalexand-t4+EzPmVLfD2fBVCVOL8/***@public.gmane.org
Eric BADEL
2007-11-12 15:47:14 UTC
Permalink
Hi,
I would like to import a sequence of raw images (indicating the
parameters like the x and y sizes, the Offset of the header to skip....).
I am sure it is a common job...

and then save as an another raw format (after removing the header).
Thank you.
Eric
Christophe Meessen
2007-11-13 08:45:35 UTC
Permalink
Hello,

I had the same problem so I created a new file format I name IJG for
ImageJ Graphics.
It is very simple. I don't know if there would be much interest in it. I
have a program to convert any raw data file into ijg file type. It may
drop any number of header bytes in the raw data file. It basically
simply add the 16 header bytes in front of the raw data.

I also heard of a small program that could add a DICOM header which is
probably more interesting because you will be able to read the file with
all programs supporting DICOM. I would be curious to compare complexity
of the code and file changes. IJG requires only a 16byte header addition
that specify the file type, the pixel type and the width, height and
depth size.

It would be nice if ImageJ could add support for a specific
FileReaderPlugins. This plugin class would be derived from the Plugin
class and simply add one method to tell if the reader can open the given
file. The arguments would be the file name and a buffer holding the 1000
first bytes as is currently done in the HandleFileReader.java program.
ImageJ could try all the plugins in InputOutput that are derived from
the FileReaderPlugin class. Even those without an underscore in their
names. The file open attempt would be performed in the lexicographic
plugin name order. Adding a number in front of the name would then allow
to control the order of plugin usage.

The change required to support this would ensure backward compatibility
since the HandleFileReader.java would try use FileReaderPlugins before
the current code which would work as a fallback.
Christophe Meessen
2007-11-13 10:01:21 UTC
Permalink
Here is the link to the programs to convert raw images into dicom. It's
for Linux

http://www.dclunie.com/dicom3tools.html
--
Bien cordialement,

Ch. Meessen
Wayne Rasband
2007-11-13 15:20:11 UTC
Permalink
You can import a sequence of raw images as a stack by using
File>Import>Raw and checking "Open All Files in Folder" in the dialog.
You can save a stack as a raw image sequence by using File>Save
As>Image Sequence and and choosing "Raw" as the format.

-wayne
Post by Eric BADEL
Hi,
I would like to import a sequence of raw images (indicating the
parameters like the x and y sizes, the Offset of the header to
skip....).
I am sure it is a common job...
and then save as an another raw format (after removing the header).
Thank you.
Eric
Eric BADEL
2007-11-13 23:16:30 UTC
Permalink
Very good. One more detail. I openned all the files as they are little
endian order. But when I save them, they are are not saved with Little
Endian Order (I need to uncheck the option to read them again). It is
bizarre. How could I force the little endian order using the Save
As>Image Sequence and and choosing "Raw" as the format ?
Thank you.
Eric
Post by Wayne Rasband
You can import a sequence of raw images as a stack by using
File>Import>Raw and checking "Open All Files in Folder" in the dialog.
You can save a stack as a raw image sequence by using File>Save
As>Image Sequence and and choosing "Raw" as the format.
-wayne
Post by Eric BADEL
Hi,
I would like to import a sequence of raw images (indicating the
parameters like the x and y sizes, the Offset of the header to skip....).
I am sure it is a common job...
and then save as an another raw format (after removing the header).
Thank you.
Eric
Rasband Wayne
2007-11-14 03:40:29 UTC
Permalink
Check "Export Raw in Intel Byte Order" in the Edit>Options>Input/
Output dialog box.

-wayne
Post by Eric BADEL
Very good. One more detail. I openned all the files as they are little
endian order. But when I save them, they are are not saved with Little
Endian Order (I need to uncheck the option to read them again). It is
bizarre. How could I force the little endian order using the Save
As>Image Sequence and and choosing "Raw" as the format ?
Thank you.
Eric
Post by Wayne Rasband
You can import a sequence of raw images as a stack by using
File>Import>Raw and checking "Open All Files in Folder" in the
dialog.
Post by Wayne Rasband
You can save a stack as a raw image sequence by using File>Save
As>Image Sequence and and choosing "Raw" as the format.
-wayne
Post by Eric BADEL
Hi,
I would like to import a sequence of raw images (indicating the
parameters like the x and y sizes, the Offset of the header to
skip....).
I am sure it is a common job...
and then save as an another raw format (after removing the header).
Thank you.
Eric
Loading...