Lowercase Letters in Text-Based CAPTCHA: A Visual Perception Analysis

CAPTCHA needs a string of letters to verify if a user is human. Each letter has distinctive visual features, making it perceivably different. No empirical work has yet examined which letters included in text-based CAPTCHA would have high correct response rates from users. The current study hopes to fill this gap. A user test on 6,389 iterations of the lowercase letters used in 1,844 CAPTCHA tests confirmed (1) users' correct response rates to the letters with the short feature had higher correct response rates than those without this feature, and (2) neither the ascending nor the descending feature had impact on the correct response rate. In addition to the theoretical insight into the use of letter perception literature in text-based CAPTCHA design, the findings do have implications for security practitioners.