Google Tried New Captcha Method

Technology.am (Apr.20 , 2009) — Google, which already devoted a lot of resources to block e-mail and Web spam, has tried a new test to keep the bots at bay.

captcha1Google authors have described the image-orientation technique. This task requires analysis of the often complex contents of an image, a task which humans usually perform well and machines generally do not.

A large repository of images are given from a web search result, we use suitable automated orientation detectors to prune those images that can be automatically set upright easily. We then use a social feedback mechanism to assert that the remaining images have a human-recognizable upright orientation.

The test is the latest variation on a screening technique called a Captcha (completely automated public Turing test to tell computers and humans apart). The idea is that people can often tell which way is up in a photo, but computers have a harder time.

The main advantages of the Captcha technique over the traditional text recognition techniques are that it is language-independent, does not require text-entry, and employs another domain for Captcha generation beyond character obfuscation. This Captcha lends itself to rapid implementation and has an almost limitless supply of images.

6 thoughts on “Google Tried New Captcha Method

  • April 20, 2009 at 3:45 pm
    Permalink

    I hope there’s going to be a NSFW version of this captcha. Can’t wait for people getting it intentionally wrong to see ANOTHER ONE, heh.

  • April 20, 2009 at 5:39 pm
    Permalink

    And, of course, there’s no chance at all that the bots would simply pick an image at random. Why, a website could have 10 images, and the bots would only get through 10% of the time. That sort of success rate would clearly make the spammers pack up and go home.

  • April 25, 2009 at 12:46 pm
    Permalink

    A site with 10 images per captcha would be guessed 1 in 1024 times, not 1 in 10.

    Do a google search for e.g. “dog” and look at how many are upright. Almost all. That means that a computer could generate captcha’s using huge amounts of web data and most spambots would never see the same image twice.

  • May 14, 2009 at 11:57 am
    Permalink

    Can’t belive how much i love this blog! I must have commented on nearly all your articles now?

  • May 15, 2009 at 12:06 am
    Permalink

    Great article, I completely agree with what your saying – as I have posted a lot of comments here now, I was wondering if you have egold as i have some spare cash in my egold account I want to donate to the site!

Leave a Reply

Your email address will not be published. Required fields are marked *