Design of an Accented Character-Based Text Captcha System

For the purpose of securing online transactions, the Completely Automated Public Turing Test to Tell Computers and Humans Apart (CAPTCHA) uses text, image, audio, and video as authentication techniques. The first and most popular version of CAPTCHA is text-based, but breaches are possible since CAPTCHA only accepts Latin characters. The use of other characters in text-based CAPTCHA design is being explored because the benefits of text-based CAPCTHA outweigh other types of CAPTCHAs. The goal of the study was to design a CAPTCHA system that would strengthen the security of online transactions by using Latin and accented characters from two Nigerian languages, Yoruba and Igbo called NaijaCAPTCHA. Latin and accented characters were used to create the CAPTCHA code types using a modified Gimpy algorithm. The design generated sixteen CAPTCHA code types using a mix of obfuscation methods. NaijaCAPTCHA was designed to generate five letters including two accented characters.