JSimaging wrote:I'm working on software to read OME sets, and I have some questions:
1) For a given set of images from a single acquisition, it seems like the OME format allows for a separate OME-XML file that would be an overview of an entire set and would have a .OME extension, in the same folder as the OME-TIFF files that contain the pixel data for the set. It also seems possible that you might not bother with the separate "header" file since each OME-TIFF file will have the XML in it, and that this might still be valid under the OME spec. So should my software expect to see a .OME header file, or should it treat any directory containing a set of OME-TIFF files as a single ND set?
This isn't something that's currently supported thought it is something we're looking to add as soon as possible. As of the released version of the schema it's expected that
ALL OME-TIFF files in a given set contain
ALL XML metadata for the entire set. You can watch this progress here:
http://trac.openmicroscopy.org.uk/ome/ticket/3538JSimaging wrote:2) I've seen some mention of a "ScreenPlateWell" high-content screening format in the Wiki, but could not find mention of simple multi-site dimensionality (acquiring channels and/or Z stacks at multiple user-defined positions). The basic location for an image plane within the ND set only seems to include channel, Z position and time point.
a) Is this kind of multi-site dimensionality supported?
Yes it is, though we as a project don't consider multi-site to be a
first class dimension, those are just
X, Y, Z, C and T. The way you would store multi-site data now is to use the
<StagePosition> for each image. We're in the process of re-organizing some of the documentation so that this is easier to see and that there are examples explaining how to do this. We'll let you know as soon as we've made progress.
JSimaging wrote: b) The format documents the X/Y position of each image plane. So should I infer a multi-site experiment if I find multiple OME-TIFF files with the same C/Z/T dimensionality but different X/Y positions?
Not necessarily. There's no way to currently
infer such a thing exclusively across all implementations as we don't have constructs within the OME data model to aggregate multi-site acquisition sets together. It's actually not something that's really come up until now. I assume based on you asking the questions so specifically that it's a pretty important construct within your software?
JSimaging wrote:3) Are samples available of all the variations in dimensionality that are possible? These will be quite helpful if you have real or at least different data at each location in the image set, and all the possible meta-data.
Thanks,
John
We certainly don't have examples with all the variations but there are a great deal available here:
http://git.openmicroscopy.org/?p=ome.gi ... hb=developWe'll also be working on a few things to make the examples easier to access and better explained. Right now they're a bit difficult to get to unless you know exactly what you're looking for.
