Color Deconvolution and the subsequent image merge

[Przemko]

Hi!
The title says it all. I have DAB (brown) stained tissue sections,
counterstained with hematoxylin (a result of immunohistochemistry).
Since my brown staining is sometimes weak, I would like to enhance it. I
used Color Deconvolution plugin to separate the colors (GREAT PLUG-IN!)
and enhance the brown color. The problem however starts here. I would
like to put my images back, but to no avail. Stacks don't work, RGB
MERGE changes colors, Image calculator changes colors, Color plugin
changes colors. So, How could I separate the two color spaces, enhance
one of the colors (that far I can do it) and then merge the images back
to get the same blue (from hematoxylin) and enhanced brown (from DAB). I
looked in various tutorials etc but cannot find anything.
TIA
Przemko

--
Przemko Tylzanowski Ph.D.
LSD & Joints
O & N1, box 813- dark green
University of Leuven
Herestraat 49
3000 Leuven, Belgium

phone: [phone]
fax  : [phone]



Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

[jchanson]

Przemko -
        Since the white has a value of "255", these images need to be
inverted (to make the background "0")  before adding them together.  Try -

- (running Color Deconvolution)
- (enhancing the colors as you see fit)
- convert images to RGB (Image | Type | RGB Color)
- invert both images (Ctrl-Shift-i),
- add them together (Process | Image Calculator)
- and then invert the final image

Jeff
Senior Imaging Analyst
[hidden email]



Przemko <[hidden email]>
Sent by: ImageJ Interest Group <[hidden email]>
06/24/2008 06:06 AM
Please respond to
ImageJ Interest Group <[hidden email]>


To
[hidden email]
cc

Subject
Color Deconvolution and the subsequent image merge






Hi!
The title says it all. I have DAB (brown) stained tissue sections,
counterstained with hematoxylin (a result of immunohistochemistry).
Since my brown staining is sometimes weak, I would like to enhance it. I
used Color Deconvolution plugin to separate the colors (GREAT PLUG-IN!)
and enhance the brown color. The problem however starts here. I would
like to put my images back, but to no avail. Stacks don't work, RGB
MERGE changes colors, Image calculator changes colors, Color plugin
changes colors. So, How could I separate the two color spaces, enhance
one of the colors (that far I can do it) and then merge the images back
to get the same blue (from hematoxylin) and enhanced brown (from DAB). I
looked in various tutorials etc but cannot find anything.
TIA
Przemko

--
Przemko Tylzanowski Ph.D.
LSD & Joints
O & N1, box 813- dark green
University of Leuven
Herestraat 49
3000 Leuven, Belgium

phone: [phone]
fax  : [phone]



Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

[Gabriel Landini]

On Tuesday 24 June 2008 19:28:13 Jeffrey C Hanson wrote:
> - (running Color Deconvolution)
> - (enhancing the colors as you see fit)

Since I wrote that plugin let me just add some further comment, for the
record.

Before anybody thinks that this is a good idea, enhancing weak immunostains
selectively (i.e. separating dyes, adjust one of them and re-merge them to
then recreate an enhanced image), is likely to make unaware observers believe
that antibody avidity is higher than it really is, and/or that the antigen
presence is more prominent that it is.

Of course I am not saying that the original post was suggesting to do any of
this, but I think that, peer reviewed journals are likely see this "practice"
as misrepresentation of the data.

There are a few articles about this type of problem which are an interesting
read: http://www.jcb.org/cgi/content/full/166/1/11

I would strongly advice against doping that. I would just leave weak staining
to remain weak. Otherwise, there is no point in following staining protocols
with controlled times, dilutions and temperatures.
There are enough problems already with the arbitrary meaning of immunostain
intensity.

Regards.

Gabriel