Authors: Walid Hariri

Published on: May 07, 2021

Impact Score: 8.0

Arxiv code: Arxiv:2105.03026

Summary

• What is new: A novel method that combines occlusion removal with deep learning for masked face recognition, utilizing the Bag-of-features paradigm.
• Why this is important: The COVID-19 pandemic has increased mask-wearing, complicating face recognition technologies due to occluded facial features.
• What the research proposes: A method involving occlusion removal and deep learning-based feature extraction from visible facial regions (eyes and forehead) followed by classification through Multilayer Perceptron.
• Results: High recognition performance on the Real-World-Masked-Face-Dataset, surpassing other contemporary methods.

Technical Details

Technological frameworks used: Deep Convolutional Neural Networks

Models used: VGG-16, AlexNet, ResNet-50, Multilayer Perceptron

Data used: Real-World-Masked-Face-Dataset

Potential Impact

Security and surveillance, smartphone manufacturers, access control systems in various sectors

Want to implement this idea in a business?

We have generated a startup concept here: MaskID.