An Analytical Survey to handle Virtual ATM through Fingerprint and Face Recognition using Deep Learning

The improvement in the technological sector in India has been crucial for the realization of improved convenience and reliability. The banking sector has also flourished that achieved significant enhancements that have been useful to the end user. Facilities such as ATMs or Automated Teller Machines have revolutionized the transaction capabilities and reduced the chances of human error. ATMs allow the dispensing and deposition of cash at all times. The cards issued by the bank can be used for this purpose and allow for a much easier integration. But there has been an increase in fraudulent transactions and card theft that decreases the reliability and security of ATMs. Therefore, to improve the security and reliability of the Automated Teller Machines, there is a need for a methodology that identifies the user while the transaction is being performed, instead of identifying the card which is being done in the current model. The realization of the user identification through a Virtual ATM approach will require the implementation of Biometric authentication mechanism. This approach discusses the use of face recognition along with fingerprint recognition using live streaming, Channel Boosted Convolutional Neural Networks along with One Time Password implementation for achieving a highly robust and secure Virtual ATM. The approach will be elaborated in utmost detail in the upcoming research directives in the future.