[ome-users] Problem decoding Slidebook files in bioformats 4.4

[Melissa Linkert]

Hi Dmitry,

Thank you for the feedback.

> 1) The problem happening now is that in many images each individual
> plane of each multi-dimensional series is decoded as a separate series
> and so in a file with for example 2 time series with 30 frames I get
> instead 60 2D series. So the number of series decoded now is exactly
> the cumulative number of individual planes available in all series.

That is not unexpected.  The idea is that now you can at least see all
of the images.

> 2) Another problem happening is the number of channel per plane. For
> example in an image with multiple 2D and 3D series with 1-2 channels
> each, in the "Bio-Formats Series Options" I see per each series the
> following: "Series 1: Capture 5: 512x512; 3 planes" and each of these
> series is decoded with 3 channels, whereas they only have either 1 or
> 2 channels.

OK, that should be fixed for 4.4 then.  I only see this problem with
3-25-11.sld; does that sound right?

> 3) third problem is the pixel size of each frame, in some images (e.g.
> cxs-11-2.sld) image width is half of what is should be. In
> cxs-11-2.sld is shows 1024x2048 instead of 2048x2048

Again, something that we'll fix for 4.4 - not sure how that
slipped by in testing.  As far as I can tell, it's just cxs-11-2.sld
that exhibits that problem, right, and not any of the other files that
you have sent?

Regards,
-Melissa

On Tue, Jul 03, 2012 at 05:29:49PM -0700, Dmitry Fedorov wrote:
> He Melissa,
> 
> Thank you so much for giving me access to the 4.4 pre-release for
> testing. I have tested the new version on all images I have here and
> unfortunately the decoding is still not perfectly correct. Actually
> for files that the decoder used to work correctly it is now worse.
> 
> 1) The problem happening now is that in many images each individual
> plane of each multi-dimensional series is decoded as a separate series
> and so in a file with for example 2 time series with 30 frames I get
> instead 60 2D series. So the number of series decoded now is exactly
> the cumulative number of individual planes available in all series.
> 
> 2) Another problem happening is the number of channel per plane. For
> example in an image with multiple 2D and 3D series with 1-2 channels
> each, in the "Bio-Formats Series Options" I see per each series the
> following: "Series 1: Capture 5: 512x512; 3 planes" and each of these
> series is decoded with 3 channels, whereas they only have either 1 or
> 2 channels.
> 
> 3) third problem is the pixel size of each frame, in some images (e.g.
> cxs-11-2.sld) image width is half of what is should be. In
> cxs-11-2.sld is shows 1024x2048 instead of 2048x2048
> 
> 
> I shared again some of the files I sent you for testing some time back
> along with the text file describing geometry and contents of each
> file:
> http://vision.ece.ucsb.edu/~fedorov/slidebook/
> 
> For example:
> 
> 1) The file "cxs-11-2.sld" contains 12 series with 1 or 2 channels and
> 2048x2048 pixels and 20 time points for 3D cases. First 2 series true
> geometry is:
> Capture 1       2D  1ch 2048x2048
> Capture 2       3D  2ch 2048x2048x20
> 
> The new bioformats decodes: 12 series with first two series geometry:
>  1024x2048 2 planes
>  1024x2048 80 planes
> 
> As usual thank you for your attention and let me know if I can help somehow!
> Dmitry