Sholl analysis composite and point selection issue

[mia2005]

Dear all,
I am trying to use the Sholl analysis plugin to batch analyze a group of images.

My images are originally in 16 bit color and they are z-stacked.

I start bye thresholding the image and convert it to binary format (0 or 255).

Then I select a point with the "point selection tool".

Then I start the Sholl analysis however, I continue having the same problem:

"Composite images are not supported" even though i set the threshold and I only have 0 or 255 pixels...

If I make a z-projection into a single plane I am able to run the Sholl analysis, however, I am not able to SAVE the point selection point into the file, even though I save it with Tif.

Does anyone has any suggestions for these problems?

1) how to run Sholl on z-projection

2) how to save a point done with the point selection tool

Thanks

[Tiago Ferreira-2]

Dear Cátia,

> My images are originally in 16 bit color and they are z-stacked.

When you say "16-bit color" I will assume you have a 16-bit grayscale multi-channel Z-stack. An example
of this type of images would be "File>Sample Images> Organ of Corti". This is effectively a "composite
image" (In other words: If your image window has a "C" slider, then you have composite image).

Composite images are slightly special and for several reasons (mainly historical) some ImageJ commands
cannot analyze them directly. You must have noticed such limitations already. E.g., when you applied
your threshold: Unless you change the display mode in the "Channels" widget (Image>Color>Channels Tool)
to either "Color" or "Grayscale" ImageJ will throw a warning that "Composite" mode images cannot be
thresholded.

To make it short: Even if all your channels are binary data, if that "C" slider is present, the image
remains a multi-channel "composite". And the Sholl Plugin cannot really parse those images, as it does
not know which channel to parse.
The solution:
You will have to extract individual channels from the "composite" and run the plugin on the channel that
contains your arbor. You can do that in multiple ways. The easiest would be to use "Image>Color>Split
Channels" (The command is also available in the "Channels" widget drop-down menu), or "Image>Duplicate"
by specifying the channel to be duplicated.


> I am not able to SAVE the point selection point into the file, even though I save it with Tif.

I will guess that you are using Fiji. My upfront guess is that these issues are a consequence of having
SCIFIO enabled. Could you confirm and uncheck the "Use SCIFIO" checkbox in Edit>Options>ImageJ2?

AFAIK, the active ROI should always be saved in the Tiff header. If this is not the case upon disabling
SCIFIO, I'd need to ask you for a sample image.

The following links should provide you more details on SCIFIO:
http://fiji.sc/SCIFIO
http://fiji.sc/Bug_reporting_best_practices#Disable_SCIFIO

Do let us know if this solves your problems.

-tiago

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

[mia2005]

Dear Tiago,

(to have the actual plugin developer answering my question, I couldn't have asked for better)

Originally I did not have a C-slider but indeed I managed to go around that error by using "split channels" or as you said (Image>Color>Channels Tool) and selecting gray scale.

When I unchecked SCIFIO option, this did fix my error of not saving the "point selection" but it also changed the way FIJI was reading my image, and after I unchecked this, I no longer need to change my image to gray scale; as soon as I apply my threshold, the image is indeed binary and the Sholl plugin reads it as such. So this actually fixed both issues (perfect!).

I will now threshold all my images and try to follow the rest of the detailed information found here:

http://fiji.sc/Sholl_Analysis#Batch_Processing

in order to batch analyze my samples.

Once again, thank you so much for your fast and detailed information.
Obrigada,
Cátia