Facial Recognition Authentication

Research output: Contribution to journalArticlepeer-review

Abstract

Keycard-based entry has been widely known as the default option for an entry system even
though they are considered to be unsafe in today’s standards. One can also forget or lose their
card. Since entry sometimes takes more time than usual with the user pulling up their card and
the system recognizing it, it might cause long lines in front of the keycard reader. This report
explores an alternative entry system that is based on biometric authentication, to be more
specific, an entry system based on facial recognition technology (FRT).
The vision of our team was to create a new entry system that can replace the AUK’s
old system. A system that works efficiently, only needing one to look at the camera for it to
confirm their identity and grant access. As many may know, facial recognition technology
works on artificial intelligence. In order to create this system, our team had to study and gain
an understanding of Artificial Intelligence, construct the source code for the system and build
a small dataset for testing reasons. However, even though the system has been successfully
implemented on a smaller scale, problems were faced when the user was wearing glasses or a
hat. Furthermore, the AI would have low accuracy where there is a difference of lighting
between the pictures of the user in the dataset and the same user waiting to be authenticated in
front of the camera. Despite the difficulties initially faced, the project was deemed successful.
This report will go into detail of the planning, code, implementation and the results
Original languageAmerican English
Article number1
Pages (from-to)1-6
Number of pages6
JournalGlobal Journal of Computer and Engineering Technology (GJCET)
Volume3
Issue number1
StatePublished - 15 Jan 2023

Keywords

  • Artificial Intelligence
  • Facial Recognition
  • Biometric Based Entry
  • Biometric Authentication

Fingerprint

Dive into the research topics of 'Facial Recognition Authentication'. Together they form a unique fingerprint.

Cite this