Fintech and Cryptocurrency

13Deep Neural Network in Security: A Novel Robust CAPTCHA Model

Manasi Chhibber, Rashmi Gandhi^*, Aakanshi Gupta and Ashok Kumar Yadav

Department of Computer Science and Engineering, Amity School of Engineering & Technology, Amity University Noida, Uttar Pradesh, India

Abstract

The CAPTCHA verifies that the user is a human by using a password to strengthen personal identification security on the web services. Web services can be made vulnerable by automated attacks on websites. CAPTCHA is a well-known security method for protecting websites from being attacked by automated attack tools. When a CAPTCHA is given a high level of distortion to make it resistant to automated attacks, it becomes difficult for humans to recognise it. The problem can be addressed in a variety of ways using a neural network. Deep learning models Dense net, Mobile Net, and VGG were used to test the security in the presented research work. To assess the model’s fitness, the models performed batch normalisation using the required additional layer. With loss and accuracy, the best model is visualised. The experimental findings confirmed the neural network’s superiority over the current state-of-the-art technique for captcha recognition.

Keywords: Security, deep learning models, captcha, dense net, mobile net, VGG, neural network

13.1 Introduction

CAPTCHA involves generating random alpha-numeric images that are used to block out non-human access to a website [12]. CAPTCHA helps in protecting you from password ...