Abstract:
ATMs or Automatic Teller Machines (ATMs) are one of the most common devices used by people today. The number of people carrying out cash withdrawal transactions with ATMs...Show MoreMetadata
Abstract:
ATMs or Automatic Teller Machines (ATMs) are one of the most common devices used by people today. The number of people carrying out cash withdrawal transactions with ATMs is growing every day. Automatic teller machines (ATMs) are one of the most important devices in the world. Conventional ATMs are vulnerable to theft due to the speed of technological advancement. More than 270,000 debit card fraud reports were filed in 2021, making it the most common type of identity theft. You need an ATM that's safe and efficient to make your transaction easier, more convenient, and more enjoyable. In this day and age, the field of Machine Perception is growing at a rapid rate. Modern advances in biometric verification technologies, such as finger print, retina scan, and facial recognition, have significantly contributed to the remediation of the security risk at the ATM. In particular, this project aims to provide a machine perception approach to address the security hazard inherent in ATM machine access. In the event of widespread adoption of this technology, facial data as well as accounts would be safeguarded. We'll create a face verification Clickbait Link and send it to the proprietor of account to confirm the personality of the illegitimate person using specialized Artificial Intelligence (AI) bots for remote certification with 92.10% accuracy, 96.34% precision, and 92.70% recall with a FPR of 0.05. While it is clear that human biometrics cannot be tampered with, this proposal will significantly address the issue of account security, allowing only the real owner of the account to access it. This eliminates the risk of fraudulent activity caused by the theft and reproduction of an ATM card. Experimental data on real-time datasets shows that the proposed approach outperforms Latest high level investigation techniques with respect to learning speed and matching precision.
Date of Conference: 14-15 December 2023
Date Added to IEEE Xplore: 19 March 2024
ISBN Information: