Large scale captcha survey

Abstract

In this research, we scanned the top 30,000 Alexa web pages to nd out how many web pages are using captcha systems. Our other goal was to classify the captcha types and evaluate the known captchas to determine if they have any kind of weak- nesses or vulnerabilities. We designed a web crawler that utilized the Beautiful Soup library to parse the top 30,000 web pages and nd evidence of captchas in the URL of the web pages by looking for keywords such as login, cart, subscribe, password, sign, register, join, auth, upload, account and registration. After scanning the top 30,000 web pages we discovered that only 10,017 of the web pages are using captcha systems. The captchas that we discovered were audio-based, image-based, text-based, captcha, reCaptcha, FunCaptcha, slider, math, custom and text/image-based captchas.