Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Efficient Masked Face Recognition Method during the COVID-19 Pandemic

Walid Hariri

TL;DR

This work tackles masked-face recognition during the COVID-19 era by removing masked regions and leveraging deep features from the eyes and forehead. It fuses three pre-trained CNNs (VGG-16, AlexNet, ResNet-50) with a deep BoF quantization that uses an RBF-based codebook, followed by MLP classification to achieve efficient, accurate recognition. The approach delivers strong results on RMFRD and SMFRD, outperforming transfer-learning, covariance-based, and some deep-feature baselines while also reducing computational load. The method's lightweight, region-focused representation supports potential real-time deployment in surveillance and access-control settings, with future work aimed at ensemble methods and broader applications.

Abstract

The coronavirus disease (COVID-19) is an unparalleled crisis leading to a huge number of casualties and security problems. In order to reduce the spread of coronavirus, people often wear masks to protect themselves. This makes face recognition a very difficult task since certain parts of the face are hidden. A primary focus of researchers during the ongoing coronavirus pandemic is to come up with suggestions to handle this problem through rapid and efficient solutions. In this paper, we propose a reliable method based on occlusion removal and deep learning-based features in order to address the problem of the masked face recognition process. The first step is to remove the masked face region. Next, we apply three pre-trained deep Convolutional Neural Networks (CNN) namely, VGG-16, AlexNet, and ResNet-50, and use them to extract deep features from the obtained regions (mostly eyes and forehead regions). The Bag-of-features paradigm is then applied to the feature maps of the last convolutional layer in order to quantize them and to get a slight representation comparing to the fully connected layer of classical CNN. Finally, Multilayer Perceptron (MLP) is applied for the classification process. Experimental results on Real-World-Masked-Face-Dataset show high recognition performance compared to other state-of-the-art methods.

Efficient Masked Face Recognition Method during the COVID-19 Pandemic

TL;DR

This work tackles masked-face recognition during the COVID-19 era by removing masked regions and leveraging deep features from the eyes and forehead. It fuses three pre-trained CNNs (VGG-16, AlexNet, ResNet-50) with a deep BoF quantization that uses an RBF-based codebook, followed by MLP classification to achieve efficient, accurate recognition. The approach delivers strong results on RMFRD and SMFRD, outperforming transfer-learning, covariance-based, and some deep-feature baselines while also reducing computational load. The method's lightweight, region-focused representation supports potential real-time deployment in surveillance and access-control settings, with future work aimed at ensemble methods and broader applications.

Abstract

The coronavirus disease (COVID-19) is an unparalleled crisis leading to a huge number of casualties and security problems. In order to reduce the spread of coronavirus, people often wear masks to protect themselves. This makes face recognition a very difficult task since certain parts of the face are hidden. A primary focus of researchers during the ongoing coronavirus pandemic is to come up with suggestions to handle this problem through rapid and efficient solutions. In this paper, we propose a reliable method based on occlusion removal and deep learning-based features in order to address the problem of the masked face recognition process. The first step is to remove the masked face region. Next, we apply three pre-trained deep Convolutional Neural Networks (CNN) namely, VGG-16, AlexNet, and ResNet-50, and use them to extract deep features from the obtained regions (mostly eyes and forehead regions). The Bag-of-features paradigm is then applied to the feature maps of the last convolutional layer in order to quantize them and to get a slight representation comparing to the fully connected layer of classical CNN. Finally, Multilayer Perceptron (MLP) is applied for the classification process. Experimental results on Real-World-Masked-Face-Dataset show high recognition performance compared to other state-of-the-art methods.

Paper Structure

This paper contains 21 sections, 2 equations, 8 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related works
  3. Motivation and contribution of the paper
  4. The proposed method
  5. Preprocessing and cropping filter
  6. Feature extraction layer
  7. VGG-16:
  8. AlexNet:
  9. ResNet-50:
  10. Deep bag of features layer
  11. Fully connected layer and classification
  12. Experimental results
  13. Dataset description
  14. Method performance
  15. Performance comparison
  16. ...and 6 more sections

Figures (8)

  • Figure 1: Overview of the proposed method.
  • Figure 2: 2D Face rotation.
  • Figure 3: (1): Masked face. (2): Sampling the masked face image into 100 regions of the same size. (3): Cropping filter.
  • Figure 4: VGG-16 network architecture introduced in simonyan2014very.
  • Figure 5: AlexNet network architecture introduced in krizhevsky2017imagenet.
  • ...and 3 more figures