Search results

This paper aims to describe the development of a new experimental CAPTCHA (completely automated public turing test to tell computers and humans apart), which is supposed to provide better protection against spam bots compared to the existing captcha solutions.

In the new CAPTCHA, the authors are using animated images of hand gestures to represent letters and numbers. A lot of people use sign language as their primary means of communication, but an automatic algorithm able to reliably recognize such gestures from videos without any additional data has yet to be developed.

The experiment showed that at first, it takes time for people who do not use sign languages in their everyday lives to adapt to the new CAPTCHA, but after several successful recognitions, they have no more trouble with it than with a typical captcha implementation.

The paper shows that people with little to no knowledge of sign languages can still recognize gestures on video relatively fast. Therefore, a gesture-based implementation can be used not only on websites aimed at sign language speakers but also as a general-purpose captcha service.