Develop an Artificial Intelligence Model Solution to Refine CAPTCHA
G. Geetha, Kesavan M, Manoj Kumar M, Udhaya Kiran M R, Ram Srinivasan · International Journal for Research in Applied Science and Engineering Technology · 2025
Abstract: Completely Automated Public Turing tests to narrate Computers and Humans Apart (CAPTCHAs) are established for protection purposes, but their growing complicatedness frequently hampers consumer occurrence. This project presents a Machine Learning model to refine CAPTCHA by reinforcing protection while asserting approachability. The model influences deep education methods, particularly Convolutional Neural Networks and Recurrent Neural Networks to analyse CAPTCHA patterns, discover proneness, and improve their design. A fruitful approach utilizing Generative Adversarial Networks guarantees CAPTCHAs remain opposing to computerized solvers while being handy. Additionally, Optical Character Recognition models are used to judge CAPTCHA strength and upgrade human readability. The projected resolution aims to balance protection and utility by underrating dishonest contradiction while guaranteeing elasticity against advanced bots. The model is prepared on a various dataset of CAPTCHAs to boost changeability. This approach improves confirmation systems, providing a secure still approachable proof design across mathematical podiums