[Insight-developers] TIFF Support of Zeiss Images

Bill Lorensen bill.lorensen at gmail.com
Tue Jul 7 18:14:16 EDT 2009


I have no reason to believe that the image contents will be correct.
I'm surprised that they were read without crashing or exceptions. This
code has been around in TIFFImageIO for over three years and, as far
as I can tell, has never had public testing.

I doubt if the problem is an old tiff library. If we do upgrade, I
think we should fully test/cover what we have first. Otherwise how
will be know if we are fixing problems or creating new ones.

I am surprised at a lack of images for regression testing from the
tiff community. I don't think there exists a set of images that cover
the tiff possibilities. A bit disappointing for such a widely used
library.

Bill


2009/7/7 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>:
>
> It seems that things have been improved since my last check - I thought it
> was during the last OME user meeting... I must have done something wrong.
> I can't see what may have improved the situation with images with several
> channels, but well, it opens the image. Great!
>
> However, I'm not getting the expected content here with Curtis' image, nor
> with http://voxel.jouy.inra.fr/blastocyste.lsm . The first slice is ok, the
> others are black. My ITK build is a bit old ('itk version 3.13.0, itk source
> $Revision: 1.3032 $, $Date: 2009-05-14 02:00:19 $ (GMT)') on my laptop -
> maybe it can explain that problem.
>
> Also, when opening the image with a VectorImage, I get 3 channels instead of
> 2. Do you have the same behavior?
>
> For the test image, if needed, we can produce a small lsm image for ITK with
> no copyright risk.
>
> Gaëtan
>
>
> Le 7 juil. 09 à 21:54, Julien Jomier a écrit :
>
>> Hi Bill,
>>
>> Mathieu Malaterre posted one in the Insight Journal a couple years ago:
>>
>>  http://hdl.handle.net/1926/139
>>
>> maybe that would do the trick. Otherwise let me know and I'll try to find
>> another one.
>>
>> Julien
>>
>> Bill Lorensen wrote:
>>>
>>> Curtis,
>>> Thanks for the data and verification.
>>> If I don't hear from Julien, I'll check in the small file you provided
>>> and add a test. I assume we can distribute the file with ITK?
>>> Bill
>>> On Tue, Jul 7, 2009 at 3:44 PM, Curtis Rueden<ctrueden at wisc.edu> wrote:
>>>>
>>>> Hi Bill,
>>>>
>>>>> Does this look correct?
>>>>
>>>> curtis at monk:~$ showinf 2chZT.lsm -nopix -nometa
>>>> Checking file format [Zeiss Laser-Scanning Microscopy]
>>>> Initializing reader
>>>>       Removing thumbnails
>>>>       Reading LSM metadata for series #0
>>>> Initialization took 0.681s
>>>>
>>>> Reading core metadata
>>>> Filename = 2chZT.lsm
>>>> Series count = 1
>>>> Series #0:
>>>>       Image count = 399
>>>>       RGB = true (2)
>>>>       Interleaved = false
>>>>       Indexed = false (true color)
>>>>       Width = 400
>>>>       Height = 300
>>>>       SizeZ = 21
>>>>       SizeT = 19
>>>>       SizeC = 2 (effectively 1)
>>>>       Thumbnail size = 128 x 96
>>>>       Endianness = motorola (big)
>>>>       Dimension order = XYCZT (uncertain)
>>>>       Pixel type = uint8
>>>>       Metadata complete = false
>>>>       Thumbnail series = false
>>>>
>>>> So, the resolution and image counts match what you're getting. Note that
>>>> the
>>>> data is really 4-dimensional (21 focal planes and 19 time points, with 2
>>>> channels), but the data is likely correct.
>>>>
>>>> -Curtis
>>>>
>>>> On Tue, Jul 7, 2009 at 2:34 PM, Bill Lorensen <bill.lorensen at gmail.com>
>>>> wrote:
>>>>>
>>>>> I just ran the large file through TIFFImageIO and produces a Meta file
>>>>> whose header looks like this:
>>>>>
>>>>> ObjectType = Image
>>>>> NDims = 3
>>>>> BinaryData = True
>>>>> BinaryDataByteOrderMSB = False
>>>>> CompressedData = False
>>>>> TransformMatrix = 1 0 0 0 1 0 0 0 1
>>>>> Offset = 0 0 0
>>>>> CenterOfRotation = 0 0 0
>>>>> AnatomicalOrientation = RAI
>>>>> ElementSpacing = 3.84502e-07 3.84342e-07 7.5e-07
>>>>> DimSize = 400 300 399
>>>>> ElementType = MET_UCHAR
>>>>> ElementDataFile = foo.raw
>>>>>
>>>>> Does this look correct?
>>>>>
>>>>> Bill
>>>>>
>>>>> On Tue, Jul 7, 2009 at 3:24 PM, Curtis Rueden<ctrueden at wisc.edu> wrote:
>>>>>>
>>>>>> Hi Bill,
>>>>>>
>>>>>>> Thanks. The TIFFImageIO read the file. I converted it to png. Does it
>>>>>>> look ok to you?
>>>>>>
>>>>>> Yes, looks good. That is the first frame. There are multiple pages in
>>>>>> the
>>>>>> TIFF, though.
>>>>>>
>>>>>> If you want, you can use the Bio-Formats command line tools to convert
>>>>>> microscopy formats to vanilla TIFFs for verification; see:
>>>>>>
>>>>>>   http://www.loci.wisc.edu/ome/formats-tools.html
>>>>>>
>>>>>> And hopefully soon I can start offering binaries for the Bio-Formats
>>>>>> ITK
>>>>>> linkage.
>>>>>>
>>>>>> -Curtis
>>>>>>
>>>>>> On Tue, Jul 7, 2009 at 2:15 PM, Bill Lorensen
>>>>>> <bill.lorensen at gmail.com>
>>>>>> wrote:
>>>>>>>
>>>>>>> Curtis,
>>>>>>>
>>>>>>> Thanks. The TIFFImageIO read the file. I converted it to png. Does it
>>>>>>> look ok to you?
>>>>>>>
>>>>>>> Bill
>>>>>>>
>>>>>>> On Tue, Jul 7, 2009 at 3:06 PM, Curtis Rueden<ctrueden at wisc.edu>
>>>>>>> wrote:
>>>>>>>>
>>>>>>>> Hi Bill & Gaëtan,
>>>>>>>>
>>>>>>>> Yep, the Zeiss LSM format is quite nasty, with all kinds of edge
>>>>>>>> cases
>>>>>>>> if
>>>>>>>> you want to handle LSMs in the wild. We have been tweaking
>>>>>>>> Bio-Formats
>>>>>>>> to
>>>>>>>> deal with it for years.
>>>>>>>>
>>>>>>>> -Curtis
>>>>>>>>
>>>>>>>> On Tue, Jul 7, 2009 at 2:04 PM, Bill Lorensen
>>>>>>>> <bill.lorensen at gmail.com>
>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>> The TIFFImageIO does have a bunch of code to handle some sort of
>>>>>>>>> tiff
>>>>>>>>> from Zeiss. Julien added it about 3 years ago.
>>>>>>>>>
>>>>>>>>> 2009/7/7 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>:
>>>>>>>>>>
>>>>>>>>>> Le 7 juil. 09 à 20:56, Bill Lorensen a écrit :
>>>>>>>>>>
>>>>>>>>>>> Curtis,
>>>>>>>>>>>
>>>>>>>>>>> Is it a tiff image? Also, it's huge...
>>>>>>>>>>>
>>>>>>>>>> LSM is a kind of tiff. With that name, I guess this is a 2
>>>>>>>>>> channels
>>>>>>>>>> 3d +
>>>>>>>>>> time image.
>>>>>>>>>> ITK's LSM reader is not able to read it, so it can be a good input
>>>>>>>>>> image
>>>>>>>>>> for
>>>>>>>>>> a failing test...
>>>>>>>>>>
>>>>>>>>>> Gaëtan
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> Bill
>>>>>>>>>>>
>>>>>>>>>>> On Tue, Jul 7, 2009 at 2:47 PM, Curtis Rueden<ctrueden at wisc.edu>
>>>>>>>>>>> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> Hi Bill,
>>>>>>>>>>>>
>>>>>>>>>>>> Zeiss has provided one example LSM dataset called 2chZT. You can
>>>>>>>>>>>> download
>>>>>>>>>>>> it
>>>>>>>>>>>> from LOCI's website:
>>>>>>>>>>>>
>>>>>>>>>>>> http://www.loci.wisc.edu/software/visbio/data/2chZT.zip
>>>>>>>>>>>>
>>>>>>>>>>>> HTH,
>>>>>>>>>>>> Curtis
>>>>>>>>>>>>
>>>>>>>>>>>> On Tue, Jul 7, 2009 at 1:40 PM, Bill Lorensen
>>>>>>>>>>>> <bill.lorensen at gmail.com>
>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>> Julien,
>>>>>>>>>>>>>
>>>>>>>>>>>>> itkTIFFImageIO supports reading of Zeiss images. I think you
>>>>>>>>>>>>> added
>>>>>>>>>>>>> this capability a few years ago. We have no example TIFF images
>>>>>>>>>>>>> to
>>>>>>>>>>>>> exercise this class. Do you have a set of small images to test
>>>>>>>>>>>>> this
>>>>>>>>>>>>> code? Right now the code coverage for TIFF is 53% (
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> http://www.cdash.org/CDash/viewCoverageFile.php?buildid=372622&fileid=10863748
>>>>>>>>>>>>> ).
>>>>>>>>>>>>>
>>>>>>>>>>>>> Bill
>>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>>> Powered by www.kitware.com
>>>>>>>>>>>>>
>>>>>>>>>>>>> Visit other Kitware open-source projects at
>>>>>>>>>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>>>>>>>>>
>>>>>>>>>>>>> Please keep messages on-topic and check the ITK FAQ at:
>>>>>>>>>>>>> http://www.itk.org/Wiki/ITK_FAQ
>>>>>>>>>>>>>
>>>>>>>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>>>>>>>> http://www.itk.org/mailman/listinfo/insight-developers
>>>>>>>>>>>>
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Powered by www.kitware.com
>>>>>>>>>>>
>>>>>>>>>>> Visit other Kitware open-source projects at
>>>>>>>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>>>>>>>
>>>>>>>>>>> Please keep messages on-topic and check the ITK FAQ at:
>>>>>>>>>>> http://www.itk.org/Wiki/ITK_FAQ
>>>>>>>>>>>
>>>>>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>>>>>> http://www.itk.org/mailman/listinfo/insight-developers
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> Gaëtan Lehmann
>>>>>>>>>> Biologie du Développement et de la Reproduction
>>>>>>>>>> INRA de Jouy-en-Josas (France)
>>>>>>>>>> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>>>>>>>>>> http://voxel.jouy.inra.fr  http://www.itk.org
>>>>>>>>>> http://www.mandriva.org  http://www.bepo.fr
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>
>>>>>>
>>>>
>
> --
> Gaëtan Lehmann
> Biologie du Développement et de la Reproduction
> INRA de Jouy-en-Josas (France)
> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> http://voxel.jouy.inra.fr  http://www.itk.org
> http://www.mandriva.org  http://www.bepo.fr
>
>


More information about the Insight-developers mailing list