Design of Face Recognition AttendX for Recording Student Attendance Data Based on Artificial Intelligence Technology

Authors

  • Ankur Singh Bist , Widya Febriani , Chandra Lukita, Sandy Kosasi , Untung Rahardja

Abstract

AttendX is an online attendance platform created by Signy Advanced Technologies from
India which aims to facilitate attendance in the scope of work and college. The way this system works
is very easy by just registering so that the system can verify faces that have been registered for
attendance. This application is designed with modern facial recognition technology that has developed
rapidly, with this sophistication it can make it easier to detect faces on structured parts such as the
forehead, cheekbones, eyes, and even the nose. When doing face detection, it is not only on one side
but there are 3 (three) parts used, so that the results are more optimal, namely the right side, the left
side, and the front side. In fact, to support the government's efforts to comply with the new normal
regulations, the AttendX platform is designed to be able to detect faces if the user is wearing a mask.
This study implements a real review of technological developments, especially making it easier to fill
in attendance online, using face recognition technology, which is the latest development in detecting
and verifying faces, introducing possible uses for aspects of use, by presenting a system called
AttendX-Net facial recognition. The method used is ResNet as well as the Multi-layer feed-forward
network to achieve the designed results. The use of LFW discovered by the University of
Massachusetts researchers did not limit the number of facial recognition by using the identification
data set from this study, including in-depth understanding techniques and explaining how machines
work with the use of anti-spoofing making this a major challenge in facial care. Likewise, the Extended
Mobile Net was used to solve a similar problem, in introducing a pipeline to reach the peak of the
presence system implementation in different cases.

Published

2020-10-16

Issue

Section

Articles