Captcha Recognition and Robustness Measurement using Image Processing Techniques

Abstract

The advances in web-based technology have revolutionized the way people communicate and share information, necessitating firm security measures. Network security prevents and monitors unauthorized access, misuse, modification, or denial of a computer network and network-accessible resources. A CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) is a standard security mechanism for addressing undesirable or malicious Internet bot programs. CAPTCHA generates and grades tests that are human solvable, but beyond the capabilities of current computer programs. Captcha prevents quality degradation of a system by automated software, protect systems that are vulnerable to e- mail spam and minimizes automated posting to blogs, forums and wikis. This paper carries out a systematic study of various Text-based Captchas and proposes the application of Forepart based prediction and Character-Adaptive Masking to break these captchas to evaluate their robustness. Captcha segmentation and recognition is based on Forepart prediction, necessity sufficiency matching and Character-adaptive masking. Different classes of captchas like simple captchas, Botdetect captchas & Google captchas are taken into consideration.