Recognize Captcha using Neural Networks

Recognize Captcha using Neural Networks

Matta Venkata Sesh Tej1, Rohitha Ravindra Myla2, Gali Jithendranath Reddy3, Trinaya Kodavai4, Nachiket Timmanagoudar5,Satyam Shahu Kulthe6, Dhruv Gupta7

ABSTRACT

CAPTCHAs (Completely Automated Public Turing test to tell Computers and Humans Apart) involve distorting texts or images in a way that remains recognizable to most humans but poses a challenge for computer recognition. Furthermore, many CAPTCHAs introduce noise or additional elements like blobs or lines into the image to enhance the difficulty for recognition. Their primary purpose is to serve as security measures on websites, preventing bots from gaining access or carrying out transactions on the site.

Conversely, machine learning (ML) algorithms, particularly deep learning (DL) neural networks, have demonstrated considerable success in addressing similar challenges such as handwritten digit recognition. This serves as motivation for the development of an ML-based CAPTCHA breaker designed to map CAPTCHAs to their respective solutions.