[annotator-dev] RFC: new ideas for specifying image regions
Robert Sanderson
azaroth42 at gmail.com
Wed Feb 5 16:47:39 UTC 2014
Thanks Robert!
Do you mind if I forward the message to the open annotation list? It's a
valuable perspective that needs to be taken into account :)
Rob
On Wed, Feb 5, 2014 at 9:43 AM, Robert Casties
<casties at mpiwg-berlin.mpg.de>wrote:
> Hi Rob,
>
> On 05.02.14 17:01, Robert Sanderson wrote:
> > Is there an issue with what Open Annotation recommends, being the xywh
> > media fragment for rectangular bounding boxes, and SVG for everything
> else?
>
> The media fragment stuff is not enough as its only rectangular and only
> pixels or (integer) percent and SVG is too much as it allows arbitrary
> coordinate systems and transforms and its XML.
>
> > http://www.openannotation.org/spec/core/core.html#FragmentURIs
> > http://www.openannotation.org/spec/core/specific.html#SvgSelector
>
> You have to specify a subset of SVG like the spec says and you have to
> somehow deal with the XML as the spec says.
>
> I am currently only looking for a solution that is easily parseable as
> JSON to be used by an Annotator.js plugin.
>
> I would like the format to be close to what is used by OAC but I am not
> satisfied with the current choice of only media fragment or SVG. I would
> like to talk with other people implementing image annotations to see if
> we could maybe come up with a better selector for this use case.
>
> Cheers
> Robert
>
> > On Wed, Feb 5, 2014 at 6:29 AM, Robert Casties
> > <casties at mpiwg-berlin.mpg.de>wrote:
> >
> >> I am starting to work again on our image annotation tool for our image
> >> server digilib. I want to implement polygon-shaped image regions too and
> >> I have been rethinking the JSON-format for specifying image regions.
> >>
> >> After looking at GeoJSON (http://geojson.org/geojson-spec.html) I have
> >> reworked our annotation format slightly:
> >>
> >> <
> >>
> https://it-dev.mpiwg-berlin.mpg.de/tracs/Annotations/wiki/json-annotation-format
> >>>
> >>
> >> To be close to GeoJSON the "type" of the shape is now in the "geometry"
> >> object.
> >>
> >> Also from GeoJSON I took the idea to specify the shapes only by points
> >> and not lengths. So the Rectangle is specified by two corner points
> >> instead of one point, height and width. This also makes coordinate
> >> transformations easier.
> >>
> >> GeoJSON does not have a Rectangle type, only Polygon (which makes sense
> >> when working with different coordinate systems). Should we also rather
> >> specify the rectangles as polygons?
> >>
> >> I also thought about renaming the member for the image region in the
> >> annotation from "shapes" to "regions" which would make it easier to deal
> >> with the older format. Or should I stick with "shapes", what do you
> think?
> >>
> >> Thanks
> >> Robert
> >> _______________________________________________
> >> annotator-dev mailing list
> >> annotator-dev at lists.okfn.org
> >> https://lists.okfn.org/mailman/listinfo/annotator-dev
> >> Unsubscribe: https://lists.okfn.org/mailman/options/annotator-dev
> >>
> >
>
>
> --
> Dr. Robert Casties -- Information Technology Group
> Max Planck Institute for the History of Science
> Boltzmannstr. 22, D-14195 Berlin
> Tel: +49/30/22667-342 Fax: -299
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/annotator-dev/attachments/20140205/42e09422/attachment-0004.html>
More information about the annotator-dev
mailing list