Google Picasa 3.5 = 11,000 Faces, 4 People, 1 Large Mosaic

By , September 26, 2009 at 6:50 pm

This began as an experiment using Google Picasa 3.5 and its new people recognition feature. In a few hours I was able to identify over 11,000 portraits of our immediate family. Using a free mosaic program called Andrea Mosaic I was able to make this from our 11,000 faces in a matter of minutes. A brief touch up in Photoshop and export to Zoomify gives you this amazingly detailed picture of our family (for the most part — some pictures of Stacey and family pop up at times as Picasa still has a hard time picking siblings apart it seems). Forthcoming – high resolution panoramas of the pictures I’ve taken! You can see it larger here – click on the image below to zoom in!

4 Responses to “Google Picasa 3.5 = 11,000 Faces, 4 People, 1 Large Mosaic”

  1. Cynthia says:

    This is totally cool! I love it. I loved zooming in and seeing all the photos that were used to make the final picture. The hardest jigsaw puzzle I’ve ever done was a Disney one that was similar to this. All those tiny pictures…wow!

  2. Stacey says:

    Cool! I might have to try that one day. I saw my face pop up quite a few times(but just the same shot, multiple times). Carolyn, I never knew you and I looked so much alike, I guess according to Google, we do!

  3. Jason says:

    Terrific mosaic! I too use Andrea Mosaic…awesome program. One question: how did you get the faces, without the rest of the photos? Picasa will make a mosaic with just the faces, but only with grids…not in the form of another photo. Is there a folder full of the face thumbnails?

  4. Dan says:

    If you download Picasa 3.0+ there is a facial recognition feature that isolates individual faces. You can export these faces as images that you can use with Andrea Mosaic to create this effect. Note that since faces have a relatively narrow color palette, this image above was color-corrected post Andrea Mosaic by overlaying the original photo with a degree of transparency in Photoshop.

Leave a Reply

Copyright 2004-2013 Dan & Carolyn Chan