LOCI plugin problem with .zvi 3 channel z-stack - minor bug?

[Robert Baer]

I have some 3-channel, fluorescent z-stacks saved by Zeiss Axiovision 4.8 as a
.zvi files.  When I read these files with ImageJ (FIJI), it uses the LOCI
plugin to open them.  It opens the files as a 16-bit hyperstack with
pseudo-colored
channels.  When I view the channels, there is visual information for
channels 2 and 3, but the frames for channel 1 appear completely black.
When I convert the hyperstack into a stack and then convert the stack back
into a hyperstack, channel 1 information is now  visible and pseudo-colored
red.  Further, if I do analyze histogram on the original hyperstack
with the blank
channel 1, it appears from the histogram ribbon that the white pixels are
originally being "pseudo-colored" black.  This black pseudo-coloring
choice seems the likely cause of the
apparently blank channel 1 frames.  In fact, I can also repair the
problem using the Image | hyperstack | Channels
tool on channel 1 of the image.  Of course, I may be making egregious
assumptions about what’s going wrong in the background.

This leads to a "feature request" unless it already exists, and I missed it.
The 3 channels are by default mapped to the pseudo colors R, G, and B
in that order.  The dye
(pseudo-color) order on all my .zvi files is blue, green, red for C1, C2, and
C3 respectively.  Would there be some way to specify this (or some
other set) as a default
pseudo-coloring order to be used by the LOCI plug-in?  One might argue
that the RGB order is sacred.
An alternative, I guess, would be to be to have a mechanism to specify
which of my channels mapped to each of
those three colors, a priori.   Of course, I can always write a macro to use the
Image | hyperstack | channels tool to fix the hyperstack coloring
after the fact.  Although this is somewhat more flexible,
it is also a bit less convenient when your entire image collection
uses a diffferent standard order.


As a side note, I will also note that there are some other places
ImageJ and its plugins that channels are
"assumed" to be pseudo-colored in the RGB order.  Dialogs therefore
refer to channels
by color names.rather than more generic chan1, chan2, chan3 names.  I
would like to suggest NOT
assuming "colors" in as many places as possible when we specify
interfaces for N-dimentional
images..

Thanks for looking into this.

Rob

------------------------------------------
Robert W. Baer, Ph.D.
Professor of Physiology
Kirksville College of Osteopathic Medicine
A. T. Still University of Health Sciences
800 W. Jefferson St.
Kirksville, MO 63501
660-626-2322
FAX [phone]

--
____________________
Robert W. Baer, Ph.D.
Professor of Physiology
Kirksville College of Osteopathic Medicine
A. T. Still University of Health Sciences
800 W. Jefferson St.
Kirksville, MO 63501 USA

[Melissa Linkert-2]

Hi Rob,

> I have some 3-channel, fluorescent z-stacks saved by Zeiss Axiovision 4.8 as a
> .zvi files.  When I read these files with ImageJ (FIJI), it uses the LOCI
> plugin to open them.  It opens the files as a 16-bit hyperstack with
> pseudo-colored
> channels.  When I view the channels, there is visual information for
> channels 2 and 3, but the frames for channel 1 appear completely black.
> When I convert the hyperstack into a stack and then convert the stack back
> into a hyperstack, channel 1 information is now  visible and pseudo-colored
> red.  Further, if I do analyze histogram on the original hyperstack
> with the blank
> channel 1, it appears from the histogram ribbon that the white pixels are
> originally being "pseudo-colored" black.  This black pseudo-coloring
> choice seems the likely cause of the
> apparently blank channel 1 frames.  In fact, I can also repair the
> problem using the Image | hyperstack | Channels
> tool on channel 1 of the image.  Of course, I may be making egregious
> assumptions about what’s going wrong in the background.
>
> This leads to a "feature request" unless it already exists, and I missed it.
> The 3 channels are by default mapped to the pseudo colors R, G, and B
> in that order.  The dye
> (pseudo-color) order on all my .zvi files is blue, green, red for C1, C2, and
> C3 respectively.  Would there be some way to specify this (or some
> other set) as a default
> pseudo-coloring order to be used by the LOCI plug-in?

If I understand correctly, setting the "Color mode" option to "Custom"
and then manually setting the colors to blue, green, and red would solve
the problem (if you then use the Bio-Formats windowless importer, these
options will be used by default).

However, the color assignment is supposed to just work if you have the
"Color mode" set to "Default".  Does the Fiji updater report that you have
the most recent version of loci_tools.jar?  If not, could you please try
updating and see if that fixes things?

Regards,
-Melissa

On Thu, May 24, 2012 at 11:20:59AM -0500, Robert Baer wrote:

> I have some 3-channel, fluorescent z-stacks saved by Zeiss Axiovision 4.8 as a
> .zvi files.  When I read these files with ImageJ (FIJI), it uses the LOCI
> plugin to open them.  It opens the files as a 16-bit hyperstack with
> pseudo-colored
> channels.  When I view the channels, there is visual information for
> channels 2 and 3, but the frames for channel 1 appear completely black.
> When I convert the hyperstack into a stack and then convert the stack back
> into a hyperstack, channel 1 information is now  visible and pseudo-colored
> red.  Further, if I do analyze histogram on the original hyperstack
> with the blank
> channel 1, it appears from the histogram ribbon that the white pixels are
> originally being "pseudo-colored" black.  This black pseudo-coloring
> choice seems the likely cause of the
> apparently blank channel 1 frames.  In fact, I can also repair the
> problem using the Image | hyperstack | Channels
> tool on channel 1 of the image.  Of course, I may be making egregious
> assumptions about what’s going wrong in the background.
>
> This leads to a "feature request" unless it already exists, and I missed it.
> The 3 channels are by default mapped to the pseudo colors R, G, and B
> in that order.  The dye
> (pseudo-color) order on all my .zvi files is blue, green, red for C1, C2, and
> C3 respectively.  Would there be some way to specify this (or some
> other set) as a default
> pseudo-coloring order to be used by the LOCI plug-in?  One might argue
> that the RGB order is sacred.
> An alternative, I guess, would be to be to have a mechanism to specify
> which of my channels mapped to each of
> those three colors, a priori.   Of course, I can always write a macro to use the
> Image | hyperstack | channels tool to fix the hyperstack coloring
> after the fact.  Although this is somewhat more flexible,
> it is also a bit less convenient when your entire image collection
> uses a diffferent standard order.
>
>
> As a side note, I will also note that there are some other places
> ImageJ and its plugins that channels are
> "assumed" to be pseudo-colored in the RGB order.  Dialogs therefore
> refer to channels
> by color names.rather than more generic chan1, chan2, chan3 names.  I
> would like to suggest NOT
> assuming "colors" in as many places as possible when we specify
> interfaces for N-dimentional
> images..
>
> Thanks for looking into this.
>
> Rob
>
> ------------------------------------------
> Robert W. Baer, Ph.D.
> Professor of Physiology
> Kirksville College of Osteopathic Medicine
> A. T. Still University of Health Sciences
> 800 W. Jefferson St.
> Kirksville, MO 63501
> 660-626-2322
> FAX [phone]
>
> --
> ____________________
> Robert W. Baer, Ph.D.
> Professor of Physiology
> Kirksville College of Osteopathic Medicine
> A. T. Still University of Health Sciences
> 800 W. Jefferson St.
> Kirksville, MO 63501 USA

[Martin Höhne]

Hi Rob,

I also know the phenomenon of black channels from zvi-images. Instead of converting to a stack and reconverting into a hyperstack you can also try assigning a LUT to the channel that appears to be black. At least in my case this renders the information in that channel visible.

Martin