[annodex-dev] spatial tagging of video
wiking at maeth.com
Tue Jul 3 05:03:05 PDT 2007
::: Lately I started to get familiar with CMML and Annodex, and while
reading the CMML draft i was just wondering that if the content is a
video stream, why isn't there a possibility to define an area (any kind
of polygon) within a timeframe (or maybe just a frame) of the content.
So this way not just a time-interval could be anchored by a <clip> tag,
but a clip tag could have a <div> tag (or something similar, like in
html) that would define that that <clip> is for a give area in the
frame(s) for the given time-interval. This one could for instance mark
an object within a time-frame and give an URI to it.... of course this
would lead to many other good possibility to use this 'extension'.
By now i was just wondering how one could
. .. define multiple polygons within the same time-interval (having
multiple <clip> tags for the same timeframe.... although as far as i
understand from the cmml DTD this is possible...)
. .. how to define a polygon: one of the straightforward thing would be
a list of vertexes or just simply use SVG for this.
here's a small example with <div> tag (based on the example of cmml from
More information about the annodex-dev