Created attachment 121
Import with BioFormat 5 and Stitch Tiles option

Please try using the Fiji Stitching plugins instead of Bio-Formats's built-in Stitch Tiles feature. The Bio-Formats stitching is an extremely limited feature which does not make full use of the stage position metadata. The Stitching plugins, conversely, are very configurable and use Bio-Formats internally to read the metadata, which is then used to perform the stitching.

You'll want to use the settings:
* Type = "Positions from File"
* Order = "Defined by image metadata"

If you still have problems with the stitching when using the Fiji Stitching plugins, feel free to reopen this bug.

See also:
* http://loci.wisc.edu/software/fiji-image-stitching
* http://fiji.sc/Stitching

Created attachment 122
LogFile fron Stitching Plugin

Created attachment 123
Screeshot Stitching Plugin Result

Hi Curtis,

this was actually what I tried first. But see the attached results ...

So I decided to try to import the CZI using the stitch Tile option (usually works fine for me ...) to have a quick look, if the data are imported correctly at all. And to me it looks like that there is already a problem during the import since the file size is much smaller than the original CZI --> it is roughly 1/7th of the original CZI file size ...

Sebi

Hi,
are there any news on this problem yet or anything I should test or try out? We use the combination ZEN & Fiji quite extensively, so I rely on this.
Please let me know, if there is anything I can do to find the problem.

Cheers, Sebi

Created attachment 124
Curtis's stitching results (see comment)

I tried stitching the data using the following settings:
* Type = Positions from file
* Order = Defined by image metadata
* Fusion method = Linear blending
* "Add tiles as ROIs" checked
* "Compute overlap" unchecked
* "Increase overlap [%]" set to 0
* No other checkboxes

And the resultant stitched image is substantially different than the one Sebastian got.

It is definitely a problem with how Bio-Formats parses the metadata. You can inspect what Fiji thinks the coordinates are by checking the "Add tiles as ROIs" checkbox. Then all the ROIs appear in the ROI manager, labeled with information that identifies each corresponding tile. You can also see the tile coordinates outputted to the ImageJ Log window. The bulk of the tiles have coordinates (0.0, 0.0) which is almost certainly incorrect.

I have filed a bug upstream in the Bio-Formats bug tracker:
https://trac.openmicroscopy.org.uk/ome/ticket/11750

Sebastian Rhode wrote:
> I saw that this bug has priority minor. Please do not get me wrong, but from my point of view this means I can not use any tile images recorded in ZEN within Fiji. Unfortunately this is a very typical use case since most applications I deal with use a software module Tile & Positions.

Priority minor is the default priority of new bugs. It is up to the bug's owner to assign a "real" priority, milestone, etc., to the ticket, which has not yet been done.

> When do think somebody will find the time to look into this? Or how can we support you guys in solving this problem?

That is a question for the Melissa Linkert, the Bio-Formats maintainer.

Melissa has now updated the OME ticket details. You are CCed, so will receive an email whenever a change happens to the ticket.

OME ticket #11750 [1] is now closed, meaning this issue should be resolved.

Sebastian, please reopen if not. Thanks!

[1] https://trac.openmicroscopy.org.uk/ome/ticket/11750