search

ome / bioformats

Bio-Formats is a Java library for reading and writing data in life sciences image file formats. It is developed by the Open Microscopy Environment. Bio-Formats is released under the GNU General Public License (GPL); commercial licenses are available from Glencoe Software.
https://www.openmicroscopy.org/bio-formats
GNU General Public License v2.0
377 stars 242 forks source link

Problem with Reading Image Series Dimension with BioFormats 5.2.1 #2558

Closed sebi06 closed 8 years ago

sebi06 commented 8 years ago

Hi Guys,

i test one of my testfiles yesterday with the BioFormats 5.2.1 and ran into an issue I do not understand yet. The file is a 3x3 tile image with 10% overlap and just one channel, where a frame is 640x640. Wirh former bioformtas this resulted always in 9 image series (one for each tile) with 640x640. Now I wanted to check out the new features, where my expectation is that it results in only one image series with 1792 x 1792 size.

But with BioFormats 5.2.1 and my testscripts now give me 2 image series, which does not really makes sense to me. Is this a bug or am I missing somethig? With 5.1.10 everything works normal and I get only one image series

Testfile: https://dl.dropboxusercontent.com/u/623476/3x3Tile_CH%3D1.czi

Help i really appreciated, Sebi

Here is the output for the testfile using BF 5.1.10 vs. 5.2.1.

3x3Tile_CH=1_BF5.1.10_output.txt 3x3Tile_CH=1_BF5.2.1_output.txt

sbesson commented 8 years ago

Hi @sebi06, what you observe is expected as perhttps://github.com/openmicroscopy/bioformats/pull/2481 whichadded support for the whole slide imaging modality in the Zeiss CZI format. As described in the Pull Request, the implications of these support addition for existing filesets are that:

In order to navigate through the pyramid sub-resolutions, you want to adapt your test script to use the Bio-Formats resolution API - see https://github.com/openmicroscopy/bioformats/blob/v5.2.1/components/formats-gpl/utils/SubResolutionExample.java for an example.

Best, Sebastien

sebi06 commented 8 years ago

Hi Sebastien,

ok, I got it. So then I think the problem is that python_bioformats (at least to my knowledge) offers no parameter getResolutionCount. Probably there are not aware of the sub-resolution API.

I python the following does not work yet:

rdr = bioformats.get_image_reader(None, path=imagefile)

totalseries = np.int(rdr.rdr.getSeriesCount())

for sc in range(0, totalseries): rdr.rdr.setSeries(sc) resolutionCount = rdr.rdr.getResolutionCount()

So would you suggest to contact those people regarding the problem. The python wrapper for BioFormats is an extremly useful and powerful tool, we must sort this out.

Cheers,

Sebi

2016-09-05 10:06 GMT+02:00 Sébastien Besson notifications@github.com:

Hi @sebi06 https://github.com/sebi06, what you observe is expected as perhttps://github.com/openmicroscopy/bioformats/pull/2481 whichadded support for the whole slide imaging modality in the Zeiss CZI format. As described in the Pull Request, the implications of these support addition for existing filesets are that:

  • all 640x640 tiles are now combined into a single full-resolution image of dimensions 1792 x 1792 pixels instead of 9 series/images previously
  • a second series for the pyramid at lowest resolution with dimensions 896 x 896 pixels is also detected

In order to navigate through the pyramid sub-resolutions, you want to adapt your test script to use the Bio-Formats resolution API - see https://github.com/openmicroscopy/bioformats/blob/v5.2.1/components/ formats-gpl/utils/SubResolutionExample.java for an example.

Best, Sebastien

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openmicroscopy/bioformats/issues/2558#issuecomment-244683631, or mute the thread https://github.com/notifications/unsubscribe-auth/ADp9oSwG8u38asIEVwige8K0nwA78G5qks5qm82WgaJpZM4J0ecj .

Dr. Sebastian Rhode Abt-Petto-Straße 27 82041 Oberhaching Tel: +49 89 20970486 Mobil: +49 151 40767993 sebrhode@gmail.com

sebi06 commented 8 years ago

Additionally I am wondering how the user can figure out programmatically which image series contains the full resolution. In ZEN it is easily possible to do things like creating an acquisition with 2 distinct positions, where the 1st position is a 3x3 tile region and the 2nd position is a 9x9 tile region (as an example). This resulted in 8 image series for the test CZI file, which can be found here:

https://dl.dropboxusercontent.com/u/623476/Pos%3D2_3x3_Tile%2B9x9Tile_CH%3D1.czi

In ZEN we call this "2 scenes" and we have a special index for this called SceneIndex. For the test CZI above this would yield in:

Dimension Order CZI : BSCYX0 Shape CZI : 1, 2, 1, 10243, 9567, 1

How is BioFormats 5.2.1 dealing with this issue?

Cheers, Sebi

sbesson commented 8 years ago

@sebi06: so using your example file, the behavior depends on whether subresolutions are flattened as series or not. In the former case, 8 series will be detected (one for each scene and resolution):

$ showinf -nopix -nometa Pos\=2_3x3_Tile+9x9Tile_CH\=1.czi 
Checking file format [Zeiss CZI]
Initializing reader
ZeissCZIReader initializing Pos=2_3x3_Tile+9x9Tile_CH=1.czi
Initialization took 0.377s

Reading core metadata
filename = Pos=2_3x3_Tile+9x9Tile_CH=1.czi
Series count = 8
Series #0 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 1792
    Height = 1792
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #1 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 896
    Height = 896
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #2 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 448
    Height = 448
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #3 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 224
    Height = 224
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #4 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 5248
    Height = 5248
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #5 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 2624
    Height = 2624
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #6 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 1312
    Height = 1312
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #7 :
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 656
    Height = 656
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false

If resolutions are not flattened then 2 series will be detected (one per scene) containing 4 levels of resolution each.

$ showinf -nopix -nometa -noflat Pos\=2_3x3_Tile+9x9Tile_CH\=1.czi 
Checking file format [Zeiss CZI]
Initializing reader
ZeissCZIReader initializing Pos=2_3x3_Tile+9x9Tile_CH=1.czi
Initialization took 0.35s

Reading core metadata
filename = Pos=2_3x3_Tile+9x9Tile_CH=1.czi
Series count = 2
Series #0 :
    Resolutions = 4
        sizeX[0] = 1792
        sizeX[1] = 896
        sizeX[2] = 448
        sizeX[3] = 224
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 1792
    Height = 1792
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false
Series #1 :
    Resolutions = 4
        sizeX[0] = 5248
        sizeX[1] = 2624
        sizeX[2] = 1312
        sizeX[3] = 656
    Image count = 1
    RGB = false (1) 
    Interleaved = false
    Indexed = false (false color)
    Width = 5248
    Height = 5248
    SizeZ = 1
    SizeT = 1
    SizeC = 1
    Thumbnail size = 128 x 128
    Endianness = intel (little)
    Dimension order = XYCZT (certain)
    Pixel type = uint8
    Valid bits per pixel = 8
    Metadata complete = false
    Thumbnail series = false

On the python-bioformats front, looking at the code, it looks like every method needs to be exposed via the Java bridge. Opening an issue on their repository to discuss the API addition might be a starting point as I do not know about theor roadmap.

sbesson commented 8 years ago

See https://github.com/CellProfiler/python-bioformats/issues/61 for the corresponding python-bioformats issue.