Annotate an image fragment inside a html page

Hi, I introduce myself.
I'm currently working to an Annotation client/server project (
http://thepund.it/).
Our architecture is based on a client that allows  the user to select a
text fragment inside a html page and insert additional informations about
it.
On the server side we have a Rest API service that on annotation creation
does some business logic operations and then saves annotation in a RDF
dataset repository (we are currently testing with success blazegraph).

We want to be as much as possible Web Annotation compliant.
I mean each annotation is written with metadata, body and target triples
following your model and stored into the repository.

In the case mentioned above we have successfully used "fragment selectors"
as described here:
http://www.w3.org/TR/2014/WD-annotation-model-20141211/#fragment-selector
Otherwise if we want to annotate a fragment image in a html page we did not
found
an applicable model.
You introduce the "svg selectors" in order to select a fragment inside an
image but in this case the source is the image url (http://.../image.jpg).
We should use both fragment (in order to select the image inside the page)
and svg (in order to select the fragment inside the image) attached to the
"sptarget" element but the source is the html page not the image url.

Could you suggest me some model to use in order to fit this use case?
Many thanks