They are entirely different data constructs in Wildbook.
Annotation viewpoints exist only on annotations and have a very fixed list for downstream comparison of IDs (left vs left; right vs right, etc.).
Keywords exist on MediaAssets and vary from Wildbook to Wildbook and may indicate scars, unique features, tags, etc. It is an open vocabulary.
We have an ability to reflect viewpoints as Keywords after detection, but this is generally not used. Manual annotation does not kick off this process.
Hi @jason - in next gen, what about changing the Add Annotation viewpoint options to the viewpoints in WBIA? I’m thinking of Up vs Top and Down vs Bottom, in particular. I’m not sure how others feel but I always get asked about up and down as viewpoints and how they should be interpreted and used. They’re a little confusing to many, me included.
Great point! While there may have been historical reasons for “up” and “down” that @parham may remember from the original AI pipeline development, it’s worth considering.
I would recommend posting this as a feature request.