More Riya rumors about whether it was going to be snatched up before it even launched a product. I have no idea what Riya's product really is--most people focus on face recognition--but outside of Google (which has quite a few experimental online projects)--I can't imagine someone picking up Riya at this point. Image reco is a hard problem. Wrapping a hard problem with a Web 2.0 interface, doesn't make the problem any more solvable.
But this doesn't mean Riya couldn't define a way around the problem. They may be able to. I don't know. Maybe they're classifying images as landscape, photos with people, and the like--presenting something that maybe could be culled from tomorrow's camera EXIF data. And if their algorithms miss something, who cares?
I've read of some web-based image classification systems in the past--and there have been some notable desktop-oriented face recognition systems too--but no one has been able to commercialize image recognition outside of some very special uses. It's a fascinating problem, but let's get real. $30 million for an image recognition company without a market tested product? Uh, right.
I can't wait to see what the product is like though.
Posted by Loren at December 19, 2005 05:52 AM