• http://midtownseo.com Michael Gold

    When searching for my name, there are some noticeable improvements. The results are nearly all head shots of men named Michael Gold– no women, no animals. On the first page, the only truly inaccurate result is a map from whitepages.com with a description that reads “Michael Gold is most likely to live in Alabama, Arkansas, Arizona.”

  • mvn

    The yellow flower is a Coreopsis.

    Google’s search is really a lot more useful when searching for works of art. Not only you identify the painting, you get a lot of info related to the art.

    Thanks for the nice post.

  • http://campeaux.com/ Brandon

    Your results were way better than mine. When I searched by my avatar, the first image match result was that of Osama Bin Laden. I’m assuming the image matching has more to do with color tone than features. (see: http://twitpic.com/5btwc6 )

    Google Goggles seems to be more accurate entirely.

  • http://emarketingwall.com/ Tharindu Gunawardana

    Is this work only based on similar image recognition or with the help of text. Either file name or description.
    If you are getting more accurate results after you entered the description , what will be the difference compared with Google normal image search?

    I know that this will recognize similar images on what ever technology they used , but how it will enhance the user behavior and help to improve search results ?

  • http://pascal.vanhecke.info Pascal Van Hecke

    Just reading this comparison now.

    I did a few tests with profile images for this blog post:
    http://pascal.vanhecke.info/2011/08/11/reverse-image-search-and-profile-pics-unwanted-linking-of-private-and-public-information-airbnb-as-an-example/

    It turns out that reverse image search does pose some problems for companies that obfuscate users’ full names for privacy reasons, yet allow to have the profile pics indexed. The example I discuss is Airbnb, but also Foursquare comes to mind (where they obfuscate the last name as well).