Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 09:28
    tinevez commented #90
  • 09:26
    imagejan commented #90
  • 09:00

    tinevez on master

    Utility to select text color ba… Select text color in the colorm… (compare)

  • 08:40
    tinevez commented #90
  • Nov 21 20:22
    tinevez closed #133
  • Nov 21 20:22
    tinevez commented #133
  • Nov 21 18:16
    jeffreyEdgarYeung commented #133
  • Nov 21 17:04
    imagejan commented #90
  • Nov 21 17:01
    imagejan commented #90
  • Nov 21 16:24
    tinevez closed #83
  • Nov 21 16:24
    tinevez commented #83
  • Nov 21 16:24
    tinevez opened #110
  • Nov 21 16:23

    tinevez on start-window-2

    An icon for a help item. A Mastodon launcher JFrame that… (compare)

  • Nov 21 15:26
    tinevez commented #90
  • Nov 21 15:24
    tinevez commented #90
  • Nov 21 15:24
    tinevez commented #90
  • Nov 21 08:27
    tinevez commented #90
  • Nov 21 07:31
    imagejan commented #90
  • Nov 20 20:34
    SarenT commented #139
  • Nov 19 22:22
    imagejan synchronize #90
Stephan Saalfeld
@axtimwalde
'The Nujan writer only writes new files. It does not update existing files.'
that sounds like one step back compared to 'does not support parallel writing' (HDF5)
Curtis Rueden
@ctrueden
Dealbreaker for your use case, eh?
@axtimwalde Are you saying there is a different pure-Java HDF5 writer?
Stephan Saalfeld
@axtimwalde
good for compact small files, I guess but it means that you always have to save the complete image or stack or time-series
I don't know of other pure Java HDF5 writers
Curtis Rueden
@ctrueden
I don't understand your comments, then.
What are you comparing Nujan against, if it's the only pure-Java implementation? The JNI-based ones? Of course they are better—they call the reference C implementation, no? But they will be fundamentally limited to the platforms for which they've been compiled.
Kyle I S Harrington
@kephale
N5 :P
Stephan Saalfeld
@axtimwalde
I just tried to understand what it does
and whether it is interesting for me
saving only new HDF5 files is a step back compared to existing JNI based libraries available for Java, and it is not necessary that JNI based implementations are better only because they are written in C
Stephan Saalfeld
@axtimwalde
@kephale, I am pretty happy with N5 currently, but there is no support for it beyond the very limited number of platforms for which a JVM with java.nio exists
Kyle I S Harrington
@kephale
but it should be pretty easy to port, right? its mostly channels + compression libraries
Curtis Rueden
@ctrueden

it is not necessary that JNI based implementations are better only because they are written in C

Agreed. My point is that in this case, the complete (and only) HDF5 reference implementation is written in C, so the Java libraries which simply interface with that are indeed likely to work much better, at least on supported platforms.

very limited number of platforms for which a JVM with java.nio exists

Is java.nio not available on the JVM on some platforms? I would expect it is there for Windows, macOS, Linux x86/x64, Linux ARM/ppc, Linux s390x (what is that?), Solaris, Solaris SPARC, AIX, z/OS, IBM i. That does not strike me as particularly limited?

Curtis Rueden
@ctrueden
I am glad that N5 is working well though. ^_^
Kyle I S Harrington
@kephale
it seems like a good idea, but presumably it will need a proper C/C++ port to get the python audience
Kyle I S Harrington
@kephale
@ctrueden do you have an opinion about default jekyll sites for hosting imagej projects on github pages (sp. i'm figuring out where to redirect imagej.fun)? open to better ideas, like if the suggestion is to redirect to the IJ wiki or something...
Curtis Rueden
@ctrueden
@kephale My opinion is: either put a simple splash/intro page there, or just simply redirect to https://imagej.net/FunImageJ.
Kyle I S Harrington
@kephale
cool, i may swap to a splash in the future, for now i'm just redirecting
Curtis Rueden
@ctrueden
For examples of the former, see http://scif.io/, http://scijava.org/ and http://fiji.sc/. For an example of the latter (sort of), see http://imglib2.net/
I think the redirect is fine, and easiest, and allows everyone to edit the content in a consistent way.
We did a splash for fiji.sc because everyone is so picky about exactly how it gets presented on the top-level page.
Kyle I S Harrington
@kephale
but the splash would be tempting for https://xkcd.com/297/
thanks!
Stephan Preibisch
@StephanPreibisch
Hi @ctrueden, we would like to make a new pom-scijava project for the BigStitcher (migrate it from its current parent pom-fiji), however, its master branch still depends on a branch of SPIM_Registration (accumulative-stitching), which we have on the BigStitcher update site. So both branches are working in parallel. So in order to do make BigStitcher a parent of pom-scijava we would have two Travis versions on SPIM_Registration (master and accumulative-stitching). Would this work? Thanks so much!!
Stephan Preibisch
@StephanPreibisch
I already made a Travis version of the branch, but travis seems to ignore compiling it: https://github.com/fiji/SPIM_Registration/commits/accumulative-stitching-travis Is there a trigger? It tried to compile some versions from the other branch: https://github.com/fiji/SPIM_Registration/commits/accumulative-stitching
Ai P. Vardhanabindu
@Aivard_twitter
Thanks @kephale . @krzysg Thanks very much. I have tried it and it works now
Curtis Rueden
@ctrueden
@StephanPreibisch You can use a different artifactId or a version suffix. Both have pros and cons. Getting Travis to build that branch requires a change to the Travis config.
I would recommend against multiple integration branches if you can avoid it.
Stephan Preibisch
@StephanPreibisch
Thanks a lot @ctrueden .. we could wait transforming the BigStitcher into a pom-scijava child until we finally release it. Could we maybe already set it up as travis job that crashes for now due to the pom-fiji, but so we can quickly change the state? Does that make sense?
Curtis Rueden
@ctrueden
@StephanPreibisch I took a quick look at these repos. I see that the SPIM_Registration branch also depends on net.imglib2.simulation:multiview-simulation:1.0.1 which also does not exist.
Maybe it makes more sense to use your own groupId like de.mdc-berlin.preibischlab for all your stuff?
Then you can have de.mdc-berlin.preibischlab:multiview-simulation and de.mdc-berlin.preibischlab:SPIM_Registration and de.mdc-berlin.preibischlab:BigStitcher?
And you can use master branch for all of those (fork SPIM_Registration from fiji into the PreibischLab org).
I'm happy to file the PRs needed so that Travis can build and deploy the respective master branches of those to the ImageJ Maven repository for you.
This would also be nice because the SPIM_Registration fork, even though the groupId is changed, would still make a JAR with the same name, and thus be overridden when you replace SPIM_Registration on your update site. (For some definition of "nice"—alternately it might make more sense to change the package prefix and artifactId so that both can coexist without conflict.)
Stephan Preibisch
@StephanPreibisch
Hi @ctrueden, thanks a lot for looking into this ... this does sound like a great suggestion. I'll discuss it with David Hoerl, but I am sure this is a nice way to go.
RachelLouise123
@RachelLouise123
Hi, I'm having some issues using the 3D viewer and volume viewer plugins. I'm trying to create a 3D model from micro ct scan data however it keeps telling me that it is running out of memory. I've tried increasing the virtual memory but that hasn't had any effect. I've also tried reducing the number of scan slices being used but that hasn't done anything either. I keep getting the same error message is there any way I can resolve this? Any help will be much appreciated!
Ellen T Arena
@etarena
Hey @RachelLouise123 - it would be best if you can post this on the Forum... that is the best place to ask these types of questions to get you the best help! :)
RachelLouise123
@RachelLouise123
Okay thank you!
Kabouik
@Kabouik
Hi. I would like to analyze colour (or maybe just average brightness, since I'm interested in dark pigmentation) in 8 samples per image, in a series of images. All samples are on a white background that should be relatively easy to extract with a threshold filter. How should I proceed to get Fiji to automatically select the 8 distinct areas in each image, and then what would be the best way to get a color (or value along a lightness-darkness relative scale) measurement?
Niko Ehrenfeuchter
@ehrenfeu
Hi @Kabouik - Hey @RachelLouise123 - it would be best if you can post this on the Forum... that is the best place to ask these types of questions to get you the best help! :)
Kabouik
@Kabouik
Just for illustrating, the images look like this: https://0x0.st/dEJ.png
Niko Ehrenfeuchter
@ehrenfeu
sorry for the wrong / double name, copy-paste bug :blush:
Kabouik
@Kabouik
Oh, I see, sorry about that.
^ @ehrenfeu
Niko Ehrenfeuchter
@ehrenfeu
No problem!
Varun Sreenivasan
@kas_varun_twitter
Hi there, I have a problem with Image J and Fiji on Surface Pro 4, running Windows 10. When I try to adjust brightness/contrast of an image, the window "B&C" does not show up. I can see that the window opens as a taskbar item, but I cannot see the window. Can someone help me fix this?