The results show that the machine learning-based approach can achieve high accuracy on simple text-based CAPTCHAs, but the accuracy decreases as the CAPTCHA becomes more distorted or noisy.
| CAPTCHA Type | Accuracy | | --- | --- | | Simple text-based CAPTCHA | 90% | | Distorted text-based CAPTCHA | 80% | | Noisy text-based CAPTCHA | 70% | captcha+breaker
Future work includes exploring more advanced machine learning-based approaches, such as deep learning, to improve the accuracy of CAPTCHA breakers. Additionally, we plan to investigate the use of CAPTCHAs in various applications, such as online registration and voting systems, and evaluate their effectiveness in preventing automated programs from accessing these systems. The results show that the machine learning-based approach
[4] J. K. Lal, P. S. Kumar, and S. K. Sahu, "A Survey on CAPTCHA and CAPTCHA Breaking Techniques," Journal of Intelligent Information Systems, vol. 54, no. 2, pp. 267-286, 2020. [3] Y. LeCun
CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) is a widely used challenge-response test designed to determine whether the user is human or a computer. The primary goal of CAPTCHA is to prevent automated programs, also known as bots, from accessing a system or performing certain actions. However, with the advancement of artificial intelligence and machine learning techniques, CAPTCHAs have become increasingly vulnerable to being broken. This paper provides a comprehensive overview of CAPTCHA, its history, types, and vulnerabilities. Additionally, we will discuss various CAPTCHA breaker techniques, including machine learning-based approaches, and analyze their effectiveness.
[3] Y. LeCun, Y. Bengio, and G. Hinton, "Deep Learning," Nature, vol. 521, no. 7553, pp. 436-444, 2015.