Google today released an upgraded version of Picasa Web Albums which will allow you to tag people’s pictures with their names.  The software will identify all the pictures that appear to be of the same person, group them together, and present them to you for tagging.

Stephen Shankland of CNET News posted a good review of the service, including a couple of amusing Picasa “gotchas”: it thought an image containing a mask and a picture of his bike wheel spokes were both human faces.

That’s predictable for a new service.  But it got me thinking.  Each time Shankland corrected the tags, he taught Picasa’s software a little more about how to recognize faces.  The more people who do that, the more Picasa learns.  And Google is the beneficiary of our free labor.  (Hey, if I want to make a little money on the side, I can get $2 for every 1,000 CAPTCHAs I type.)

“Oh come now,” you say, “stop being so paranoid.”  All I can do is point to this interview with Google’s Marissa Mayer:

“You may have heard about our [directory assistance] 1-800-GOOG-411 service.  Whether or not free 411 is a profitable business unto itself is yet to be seen.  I myself am somewhat skeptical.  The reason we really did it is because we need to build a great speech-to-text model … that we can use for all kinds of different things, including video search.  The speech recognition experts that we have say: If you want us to build a really robust speech model, we need a lot of phonemes, which is a syllable as spoken by a particular voice with a particular intonation.  So we need a lot of people talking, saying things so that we can ultimately train off of that.  So 1-800-GOOG-411 is about that: Getting a bunch of different speech samples so that when you call up or we’re trying to get the voice out of video, we can do it with high accuracy.“

Submitted by: Dan Giancaterino, Internet Librarian
on September 02, 2008 - 1:10 pm

| More Google, Multimedia, Pictures, Web 2.0 Posts