Facial recognition goes mainstream

Oct. 11, 2006

I found this very interesting piece of news the other day. It pertained to Picasa, which is Google Inc.'s photo management system (kind of like iPhoto from Steve Jobs/Apple). This software basically finds all your photos on your PC in the same way that Google's core web search engine finds pages like this one on the Net.   [Incidentally, I was thinking how this is pretty similar to what Steve Russell at 3VR is doing with its searches for surveillance video, but that might be a bit of a stretch.]

So, the fact that Picasa can organize your photos is nice, but so what? Here's the "What" -- the next version of the Picasa software is reportedly designed to use facial recognition to help you sort the photos. So if you want to see all the photos of Mom, you can do that, without having to add any metatags to let the software know that "this is a photo of mom."

What brings this all the way back around is that to get the technology into Picasa, Google bought Neven Vision, a biometrics/image recognition company that was working on mobile visual recognition problems, including a recognition tool for the LAPD.

Now, as we all know, today's video streams really are nothing more than a series of digital photos, so what Google/Neven Vision (you might as well forget the Neven Vision name now b/c Google likes its own brand) is doing is actually rather significant to our industry as well. If they are able to perfect this technology and get the consumer public accustomed to facial recognition tools by making it available in Picasa, then I wouldn't be surprised if they couldn't relicense this back into commercial applications for IP video surveillance.

In fact, now that Google has YouTube (which replaces its own video management and sharing platform), there's even some people starting to speculate that Google might turn the Neven Vision recognition tools onto video itself. Of course, that's all speculation...

 

Searching is everything. Finding is better, but still, searching is everything.

 

--Geoff