Enhancing security of Urdu language websites through Urdu CAPTCHA

摘要

Nowadays almost every daily activity such as in education, entertainment, communication, e-commerce (to name a few) are done through Internet. For secure login to such activities, user have to sign-up to access those resource over the Internet. However, a bot (web robot) can sign-up automatically by entering fake information to access those resources. To prevent from automated-bots attack, CAPTCHAs are used to secure websites. CAPTCHA stands for Completely Automated Public Turing tests to tell Computers and Humans Apart and it is a Turing test to confirm that the end user is human or a robot (bot). CAPTCHA was introduced by gang of four (Luis Von Ahn, Manuel Blum, Nicholas J. Hooper) at CMU in 2000. There are many types of CAPTCHAs including text based, basic questions, mathematical questions and selecting images, whereas scrambled text-based CAPTCHAs are mostly used. However, nowadays bots are very intelligent to break alpha-numeric scrambled text-based CAPTCHAs. Most of CAPTCHAs are in English, European and in east Asian languages but unfortunately Urdu language has less focus. According to the literature, CAPTCHAs should be designed in local and regional languages including Urdu and Sindhi to improve the security, accessibility and understandability of our regional websites. This research aims to propose an Urdu language-based CAPTCHA for regional URDU websites. This research highlights the limitations of exiting available CAPTCHAs and compares the efficiency of our proposed CAPTCHA with exiting work. The results show that our proposed CAPTCHA provides robustness and efficiency in terms of complexity.