Integrate Watson API for image recognition

[helen]

Image recognition seems particularly interesting, as does speech-to-text.

[jeffpaul]

Note that I split out the speech-to-text API in #47

[jeffpaul]

From @ryanwelcher via #28:

This could be applied to Featured Images or media uploads. It would make sense for us to use the Watson tool for our first offering.

[jeffpaul]

Will want to ensure this includes UI for filtering in the Media Library (currently in backbone.js).