Google tests new captcha

If you have signed up for an account on any site on the net, you’ve no doubt used a captcha. Captcha's are designed to stop bots from signing up for spam purposes. Google has developed a new innovative method to stop computers pretending they are humans by orienting images in different positions, and requiring the human to pick out the upright image.

The idea behind this is that people can tell when an image is up or down. Captcha's are usually obscured or distorted text. This new method of captcha still needs a lot of work, including identifying 3D images and distinguishing between cats and dogs.
Rich Gossweiler, Maryam Kamvar, and Shumeet Baluja(Google authors) describes the technique
“The main advantages of our Captcha technique over the traditional text recognition techniques are that it is language-independent, does not require text-entry (e.g. for a mobile device), 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.
We conducted extensive experiments to measure the viability of this technique...Our Captcha technique achieves high success rates for humans and low success rates for bots, does not require text entry, and is more enjoyable for the user than text-based Captcha.”
(source: http://www.richgossweiler.com/projects/rotcaptcha/rotcaptcha.pdf”)

Another problem Google needs to overcome is to find the right balance in the images so that they're not too easy or too confusing. Some images are easier than others to identify because they have obvious clues such as face, blue sky, green grass and text. These cues can also be used by computers to figure out orientation. Other images can be hard to identify their orientation properly due to multiple objects within the image.
Images are drawn from the web and those known to perform well in the tests. When people are consistently unable to discern when an image is correct these images are not included in the library of images used.
Researchers are in favor of this system over the current Google system of distorted or obscured captchas, but this image based captchas aren’t computer proof.

Original article

Follow me on twitter, for free updates! Click Here