Automated CAPTCHA Detection through Machine Learning Models: A Survey

Automated CAPTCHA Detection through Machine Learning Models: A Survey

Anjali Kushwaha1, Prof. Vipin Kasera2

Research Scholar1, Asst. Professor2

VITM, Indore, India1,2

Abstract— As artificial intelligence is improving its purview in matching human level intelligence, it is necessary to identify human and machine learning intelligence. One of the ways is to use a CAPTCHA, or "Completely Automated Public Turing test to tell Computers and Humans Apart," is a security measure that helps distinguish between human users and automated bots. It is widely used across websites to prevent bots from performing certain actions, such as creating accounts or submitting forms, to safeguard against spam, data breaches, and unauthorized access. With advances in machine learning, particularly in computer vision and natural language processing, CAPTCHA systems have been both challenged and enhanced. Machine learning models trained to recognize CAPTCHAs attempt to decode visual and audio puzzles designed specifically to be hard for computers to solve. This paper presents a comprehensive review on machine learning and deep learning algorithms for CATPCHA recognition with the salient features of contemporary work.

Keywords— CAPTCHA, Machine Learning, Deep Learning, Object Detection, Classificaiton Accuracy.