ImageJ 2.0.0-beta-7 released

[ctrueden]

Hi everyone,

The ImageJDev team has released the seventh beta of ImageJ2, version 2.0.0-
beta-7. This release achieves two key longstanding goals of the project:
unifying image I/O and better support for very large image data.

ImageJ 2.0.0-beta-7 is a "beta"-quality release, meaning the code is not
finished. The design is more stable than before but still subject to change
until the final 2.0.0 release. It is recommended that people continue to
use ImageJ v1.x for critical work.

The ImageJ2 user interface is modeled after ImageJ v1.x as much as
possible. However, under the hood, ImageJ2 is a complete redesign of
ImageJ. It provides backward compatibility with older versions of ImageJ by
bundling the latest v1.x code and translating between "legacy" and "modern"
image structures.

Development of this release of ImageJ focused on better support for
N-dimensional data, particularly large datasets. This release adds support
for "cell images" which are similar to, but more powerful than, the
"virtual stacks" of ImageJ 1.x. There has also been a further strengthening
and stabilization of the ImageJ2 architecture, with the new plugin system
generalized into a SciJava Common library used by both ImageJ2 and SCIFIO.

For further details on this release, including a discussion of new
features, please see our blog post:
  http://developer.imagej.net/2013/06/12/imagej-v200-beta-7

You can download the release from:
  http://developer.imagej.net/downloads

See the Downloads page for the updated table of planned future releases and
roadmap links for the project. The focus of the next release (2.0.0-beta-8)
is on better multidimensional image analysis and processing routines, such
as computation of measurements and statistics from ROIs.

For more details on the ImageJ developments, see the web site at:
  http://developer.imagej.net/

Regards,
Curtis Rueden
ImageJDev project lead
Laboratory for Optical and Computational Instrumentation
University of Wisconsin-Madison

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

[Frederic V. Hessman]

I have a principle question about ImageJ 2.0 and couldn't find any docs or links (and though I'd ask before plowing through the code): what about metadata?

The 2.0.0 version strips FITS files of their metadata completely.  In ImageJ 1.*, this metadata was available via the image info, admittedly not in any officially organized fashion, but at least it could be accessed (e.g. the FitsJ class in the Göttingen astronomy plugins simply handles the FITS header as a string array and provides methods for key-value queries and handling - primitive but effective).  Why was this functionality removed?  Has it been substituted by something official?  Is the same true for other data formats with metadata (e.g. tiff)?  If the official metadata model works as advertised (simple standard axes like "X" and "Y" provided for basic users but user-defined axes descriptions possible), then it should be trivial to parse FITS headers to create that metadata.

The FITS metadata system is vastly more powerful and complicated than most ImageJ users (or developers) need, but it should be simple to keep those parts of the metadata changed by image manipulations within an ImageJ 2.0 session if they are well-documented internally and be able to feed them back into FITS format for storage.  Of course, a cleaner system would be to be able to modify the metadata dependent upon the metadata model of the original data (e.g. a metadata handling plugin for different formats)….

Rick

------------------------------------------------------------------------------------------------
Dr. Frederic V. Hessman     [hidden email]
Institut für Astrophysik          Tel.  [phone]
Friedrich-Hund-Platz 1         Fax [phone]
37077 Goettingen                 Room F04-133
http://www.Astro.physik.Uni-Goettingen.de/~hessman
-------------------------------------------------------------------------------------------------
MONET: a MOnitoring NEtwork of Telescopes
http://monet.Uni-Goettingen.de
-------------------------------------------------------------------------------------------------

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

[ctrueden]

Hi Rick,

> I have a principle question about ImageJ 2.0 and couldn't find any
> docs or links (and though I'd ask before plowing through the code):
> what about metadata?

Thanks, this is a very opportune question! I am sorry that the beta release
gave the impression that metadata is not a paramount issue. On the
contrary, in scientific image processing, it is at the forefront of our
minds!

As mentioned in the 2.0.0-beta-7 blog post, the I/O subsystem of ImageJ2 is
SCIFIO, which is a library born from the Bio-Formats project, but much more
generalized. SCIFIO is intended for use across several of the SciJava
software projects (http://scijava.org/), not just in the life sciences. As
such, it is a perfect match for general-purpose image processing tools like
ImageJ.

The primary purpose of SCIFIO is to enable interchange of metadata from
open and proprietary file formats across various metadata standards. In
Bio-Formats, the one and only metadata interchange format is OME-XML.
SCIFIO generalizes that to make metadata models pluggable. Of course,
OME-XML will still serve as a flagship data model for life sciences images,
but other models can be used just as well; SCIFIO allows you to define your
own Translators between metadata formats.

Every file format SCIFIO supports defines its own Metadata class, which
houses all the metadata as close to its original structure as possible. We
have not yet exposed this in the ImageJ2 UI, but we have a plan for doing
so, which will be in place before the 2.0.0 final release.

The FITS format implementation in SCIFIO 0.2.0 is a first cut which parses
the metadata as key/value pairs [1] (it was ported from the Bio-Formats
implementation, which does the same [2]). We will be reconciling this code
with the ImageJ1 implementation [3] as needed before the 1.0.0 release of
SCIFIO.

> simple standard axes like "X" and "Y" provided for basic users but
> user-defined axes descriptions possible

I know you also brought up coordinate systems on the ImageJ list when the
ImageJ2 project first started [4]. Thanks to you, we do have a ticket for
more sophisticated coordinate system handling [5], which in theory ImgLib2
is already flexible enough to accommodate. (And note that I come from an
earth sciences background, so I know about coordinate system
transformations, unit conversions, etc., in that similar context [6].) But
for time reasons, to actually provide that sort of feature end to end --
from the data model through to the end-user interfaces -- is outside the
scope of the ImageJ 2.0.0 development; the aforementioned ticket is
tentatively slated for "ImageJ 3.0.0" unless someone wants to champion it
sooner, of course.

As for supporting metadata about multidimensional axes of
Euclidean/Cartesian spaces, I did a lot of work on that at a recent ImgLib2
hackathon. More work still needed, but good progress so far! [7]

So in short: the metadata system for ImageJ2, ImgLib2 and SCIFIO is more
sophisticated than just key/value pairs. :-)

Regards,
Curtis

[1]
https://github.com/scifio/scifio/blob/scifio-0.2.0/components/scifio-devel/src/io/scif/formats/FITSFormat.java#L169
[2]
https://github.com/openmicroscopy/bioformats/blob/v4.4.8/components/scifio/src/loci/formats/in/FitsReader.java#L111
[3]
https://github.com/fiji/ImageJA/blob/v1.47t/src/main/java/ij/plugin/FITS_Reader.java
[4] https://list.nih.gov/cgi-bin/wa.exe?A2=IMAGEJ;e0ecf83a.0912
[5] http://trac.imagej.net/ticket/40
[6] http://visad.github.io/
[7]
https://github.com/imagej/imglib/compare/d2e6e0b6508f18b498961afc5c1822882f65f637...img-metadata


On Thu, Jun 13, 2013 at 5:24 AM, Frederic V. Hessman <
[hidden email]> wrote:

> I have a principle question about ImageJ 2.0 and couldn't find any docs or
> links (and though I'd ask before plowing through the code): what about
> metadata?
>
> The 2.0.0 version strips FITS files of their metadata completely.  In
> ImageJ 1.*, this metadata was available via the image info, admittedly not
> in any officially organized fashion, but at least it could be accessed
> (e.g. the FitsJ class in the Göttingen astronomy plugins simply handles the
> FITS header as a string array and provides methods for key-value queries
> and handling - primitive but effective).  Why was this functionality
> removed?  Has it been substituted by something official?  Is the same true
> for other data formats with metadata (e.g. tiff)?  If the official metadata
> model works as advertised (simple standard axes like "X" and "Y" provided
> for basic users but user-defined axes descriptions possible), then it
> should be trivial to parse FITS headers to create that metadata.
>
> The FITS metadata system is vastly more powerful and complicated than most
> ImageJ users (or developers) need, but it should be simple to keep those
> parts of the metadata changed by image manipulations within an ImageJ 2.0
> session if they are well-documented internally and be able to feed them
> back into FITS format for storage.  Of course, a cleaner system would be to
> be able to modify the metadata dependent upon the metadata model of the
> original data (e.g. a metadata handling plugin for different formats)….
>
> Rick
>
>
> ------------------------------------------------------------------------------------------------
> Dr. Frederic V. Hessman     [hidden email]
> Institut für Astrophysik          Tel.  [phone]
> Friedrich-Hund-Platz 1         Fax [phone]
> 37077 Goettingen                 Room F04-133
> http://www.Astro.physik.Uni-Goettingen.de/~hessman
>
> -------------------------------------------------------------------------------------------------
> MONET: a MOnitoring NEtwork of Telescopes
> http://monet.Uni-Goettingen.de
>
> -------------------------------------------------------------------------------------------------
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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