Development of a new CAPTCHA System for Visually Impaired Arabic Speakers/ Users

Abstract

The CAPTCHA verification codes are a test with questions created and correctable by the computer, this technique is used to distinguish between human and machine. This work has developed a new CAPTCHA that depends on vocal interaction of the internet applications users and uses speech recognition , that is because the issuance, production and pronunciation of words is usually a human task while the machines lack the fluency and accuracy in processing words like humans. The new CAPTCHA depends on complexities in the techniques of transforming voice to text after being pronounced by the user, but the strength of the new technique lays in the fact that those user vocal instructions has to be preceded by the user's listening to them being randomly generated from the systems' database, subsequent to that the same user pronounce those words giving the verification process a doubled strength. This work concentrates on the development of CAPTCHA depending on the intact cognitive styles. The study in this paper depends on the participation of a number people with different types of visual impairment, who use various types of screen reader technologies for performance and usability tests of the recent in-use verification codes systems and the new verification codes system. The tests results showed a 70% rise in the levels of passing the verification questions, and a 95% reduction in the time consumed in gaining access to the verification code protected content after the use of the proposed new technique for websites verification.