Decoding Captcha using Machine Learning for Identifying Criminal Attempts

Abstract

Completely Automated Public Turing Tests to Tell Computers and Humans Apart (CAPTCHAs) is a computer program which is meant to set humans and machines apart. In simple terms it is a system which identifies whether the user is a human or a machine. We use Captcha very often in our daily basis when we try to access internet. Captcha is meant for security. Decoding or recognizing the patterns of Captcha can help us know about any morally offensive activities taking place prior. Also we can stalk the systems of those who tend to misuse their powers and cause harm to the country. They are designed in such a way that they are solved by humans and unsolvable by machines.We can use convolutional neural networks and recurrent neural networks instead of the conventional methods of CAPTCHA breaking based on segmenting and recognizing a CAPTCHA. This paper mainly focuses on training a machine to decode Captcha which can be used to identify the criminal activities stalking the criminal activities in advance as well.