Ticket #559 (closed defect: fixed)
Opened 2010-09-13T12:12:44-05:00
Last modified 2011-02-02T04:45:00-06:00
LSM reader XT linescan opened as XY
Reported by: | curtis | Owned by: | melissa |
---|---|---|---|
Priority: | minor | Milestone: |
|
Component: | bio-formats | Severity: | serious |
Keywords: | Cc: | white@… | |
Blocked By: | Blocking: |
Description
From Dan White:
The y dim is a single pixel, and the other dim is time, so it should be shown as a 1 pixel high line with the time slider to go through the time points.
The image properties tool and the bioformats metadata in fiji shows that the frame interval as 0 which is wrong, and the y is 30000, also wrong.
File is at zeiss-lsm/dan/2_1mm_pt1.lsm.
Pushing this to 5.0 - it's actually a rather tricky issue.
Originally, we dealt with XZ and XT scans as proposed above - the Y dimension would be set to 1, and the Z or T dimension would be set to whatever is stored in the file. However, this resulted in displayed images that were 1 pixel high, which is typically not what vendor software displays. A lot of people complained that their images looked wrong, and so with #333 we made the executive decision to handle XZ and XT scans based on how they should be displayed.
Obviously, though, that's not quite right. The current roadblock (and it's not a small one) is that the entire stack of OME software - from the OME-XML schema all the way up through Bio-Formats to OMERO - assumes that the first two dimensions in a dataset are XY, and so any image returned by Bio-Formats will be an XY image. This falls under the big umbrella of N-Dim support (see #398, ome-xml:112, NDimensionalData).
The idea is that eventually our modelling of dimensions will be more flexible, so that any two dimensions can be the first two dimensions, and then returned images use the first two dimensions (whatever they may be) as the width and height. We would, then, offer the option of viewing (for instance) an X = 1024 x T = 3000 XT scan as either 3000 1024 x 1 images or 1 1024 x 3000 image.