HTK analysis of text-based CAPTCHA using phonemic restoration effect and similar pronunciation with an Asian accent

Abstract

In Recent years, bot (robot) programs have been a threat on the web. Some kinds of the bots acquire accounts of web services in order to use the accounts for SPAM mails, phishing, etc. CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) is one of the countermeasures for preventing bots from acquiring the accounts. Especially, text-based CAPTCHA is applied to almost all famous web services since it can be implemented easily. However, CAPTCHA faces a problem that evolution of algorithms for analysis of printed characters disarms text-based CAPTCHA. Stronger distortion of characters is an easiest countermeasure of the problem. However, it makes recognition of characters difficult not only for bots but also for human beings. Therefore, we proposed a new CAPTCHA with resistance to the analysis by computers. We focus on the human abilities of phonemic restoration and recognition of similar pronunciation. Furthermore, we also pay attention to the fact that words with an Asian accent can be heard by English speakers. The proposed CAPTCHA makes machinery presumption difficult for bots, while providing easy recognition for human beings. In this paper, analysis of our CAPTCHA by computers with HMM (Hidden Markov Model) is evaluated.