2013 Volume 15 Issue 4 Pages 337-352
In the case of user authentication on the network, most of the existing methods which tell humans and computers apart automatically make use of the difference in perceptual recognition between humans and computers. Researchers point out that it is difficult for visually and/or hearing impaired people to leverage it. In this paper, we propose a method which generates linguistic questions to avoid relying on the specific perceptual ability. Based on the difference of a cognitive ability between humans and computers, we generate linguistic questions such as contextual interpretation. Moreover, we utilize contents on the network to generate almost an infinite number of questions.