Facial Recognition in Picasa is Amazing
No Comments »
October 3rd, 2009
I’ve always been a big fan of Picasa – it is a great, free application for organizing (and making small edits to) your photos. I especially like the ability to easily straighten, rotate, and crop pictures. And now it just got a whole lot better with facial recognition built-in to the latest version (finally!). This has been a web-only feature for Picasa Web Albums for awhile now, and I have been waiting for Google to integrate it into the desktop application too. Fortunately, the wait is over and it works very well.
Facial recognition is like magic. After it scans your entire photo library (which can take awhile), Picasa will ask you to identify the faces it has found. This is linked to your Gmail Contacts to make it easy to find people’s names. After this, it will start suggesting other pictures that it thinks are the same person, and I’d estimate that over 90% of the time it is correct. Even if the person is wearing a hat, making a funny face, turned to the side, or wearing a fake mustache, Picasa is usually able to identify the person correctly. Like I said, magic.
And the best part is that after you have gone through the minimal effort to label everyone, you instantly have a very well organized photo collection. You can easily find any picture with a particular person in it, or even a particular combination of people. Awesome.
Tricking Picasa with the “Tim Face”
So I went through a phase where I made the same face in just about every photo taken of me. I’m sure you’ve seen it before, it’s not particularly original – I just have my mouth wide open like I am really, really excited about something. (It was college. I didn’t know any better). Anyway, some of my friends started referring to this expression as the “Tim face” and often copied (i.e. mocked) me in their own photos. Apparently they must all do a really good impression of me because it was good enough to fool Picasa. That’s right, Picasa’s facial recognition assumed that anyone making the Tim face was me! See example below:
