Implementation of an Embedded Masked Face Recognition System using Huskylens System-On-Chip Module

Globally, Facial recognition systems have been increasingly adopted, by governments, as a viable means of identification and verification in public spaces such as the airport, train stations, and stadiums. However, in the wake of the COVID-19 outbreak, the World Health Organization (WHO) declared th...

Full description

Saved in:
Bibliographic Details
Main Authors: Kennedy, Okokpujie, Chiamaka, Anicho-Okoro, Princess, Okokpujie Imhade, Julius-Olatunji, Okesola
Format: Conference Proceeding
Language:English
Subjects:
COVID-19
Face recognition
facial recognition
Government
HuskyLens
Kendryte
masked faces
Pandemics
Real-time systems
Safety
System performance
System-on-chip
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Globally, Facial recognition systems have been increasingly adopted, by governments, as a viable means of identification and verification in public spaces such as the airport, train stations, and stadiums. However, in the wake of the COVID-19 outbreak, the World Health Organization (WHO) declared that wearing face masks is an essential safety precaution. As a result, current facial recognition systems have difficulties recognizing faces accurately, which motivated this study. This research aims to implement an embedded masked face recognition system using the HuskyLens SoC module to identify people, even while wearing a face mask. The developed method was actualized using the Kendryte K210 chip embedded in the HuskyLens module. This system-on-chip design was integrated with other peripherals using an Arduino Pro-mini board. The results of testing and evaluating the system's performance show that the system's facial recognition accuracy with masked and without masks faces was 90% and 95%, respectively. Implementing this solution in our environment would enable accurate real-time recognition of masked and unmasked faces
ISSN:2377-2697
DOI:10.1109/NIGERCON54645.2022.9803092