An Automated Reverse Turing Test Using Facial Expressions

This paper proposes a new image based CAPTCHA test, an automated reverse turing test using facial expressions based on facial emotion recognition. In this test the user is presented with a set of images of human faces having common emotion and a set of labeled smiley images with different emotions. Of the smileys, one will be having the emotion as that of the human faces. The user has to identify the common emotion associated with the human faces and choose it from a given list of expressions provided to pass the test. User studies indicates that this CAPTCHA can be solved with 95.82% success rate while that of ESP PIX is 87.2% and SQUIGL PIX is 67.8%.