Advanced Healthcare Systems

16A Comprehensive Analysis on Masked Face Detection Algorithms

Pranjali Singh^1*, Amitesh Garg² and Amritpal Singh¹

¹Dept. of Computer Science & Engineering, Dr B.R. Ambedkar National Institute of Technology, Jalandhar, India

²Sabre Travel Technologies Pvt. Ltd., Bengaluru, Karnataka, India

Abstract

The COVID-19 is an ongoing crisis that has resulted in a large number of fatalities and safety concerns. People also carry masks to cover themselves to effectively prevent the transmission of this virus. In this situation, recognizing a face is very challenging. In certain cases, like facial attendance, face access control, and facial security, this makes traditional facial recognition technology ineffective, for that urgent requirement to improve this recognition performance and use the technology on the masked face. During the current pandemic, the main objective of researchers is to deal with these problems through quick and accurate approaches. Throughout this chapter, we suggest a clear way centered on removing masked area and deep learning–related techniques to resolve the issues of mask detection. Another way of finding the masked face is to go through TensorFlow, YOLOv5, SSDMNV2, SVM, OpenCV, and Keras. The first phase is discarding the area of the masked face. Next, to determine the best aspects from the areas collected for that, we use a pre-instruction of deep CNN (Convolutional Neural Network). We used labeled image data to train the CNN model. With a 98.7% accuracy, ...