Quite obviously, the next big thing will be image and video search; in particular, given a sample image of a person, search the internet images or videos of that person, as well as the textual information about that person. The technology already exists for scanning the Super Bowl crowd for terrorists, but is not deployed yet for normal people to use.
It's a nice hard AI problem, because one can not only do straight face recognition, but also use textual context to figure out names of people, locations and events, and follow the thread of those names, to find more images.
And the world will become a creepier place.
No comments :
Post a Comment