It looks like there is a new company called “Riya”:http://www.riya.com that is creating a new photo management application that uses computer vision to help catalog your digital photo collection. The app should be going into very limited alpha in the next day or so.
I’ve been keeping an eye on Computer Vision for a few years now, and I’ve been carrying around an idea for a web app that would make use of it for almost as long. It’s a bit of a bummer seeing it reaching the consumer level, since I’ve been sitting on my ass for so long. On the other hand their service and my idea bare only the most superficial resemblance.
It sounds like Riya’s big feature is facial recognition. The application processes the photos to try to identify all the different people in the photos. You then have to train the application by validating or invalidating some of its guesses about faces from photo to photo. Then you can use the app to tag existing and new photos by the people present in the picture.
What’s both interesting and distrubing is that they are going to share training data across users. Each time a new user joins, they’ll benefit from faces already in the system, something that will be even more valuable as people learn about the service from their freinds, reinforced, no doubt, with weighting due to proximity in their social network.
TechCrunch » First Screen Shots of Riya
Riya leverages potent facial and text recognition technology with an intelligent interface to help people make sense of the thousands of untitled and untagged photos that are building up on their hard drives (and on the web).