Citation

Yo, Ming Chun and Chong, Siew Chin and Chong, Lee Ying (2024) Optimizing Masked Face Recognition: A Tailored CNN Integrates with Different Classifiers. In: 2024 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET), 26-28 August 2024, Kota Kinabalu, Malaysia.

Text Optimizing Masked Face Recognition_ A Tailored CNN Integrates with Different Classifiers.pdf - Published Version Restricted to Repository staff only Download (2MB)

Abstract

The surge in research on masked face recognition highlights the need for more accurate and efficient systems as wearing masks has become routine after the COVID-19 pandemic. Despite CNN's success in computer vision and image processing tasks, it may still have weaknesses in recognizing masked faces. However, traditional machine learning techniques are commonly used to classify occluded face images which make their integration into the classification stage of CNN highly significant. Therefore, this paper aims to study on optimizing masked face recognition by integrating different classifiers with tailored CNN. Multiple different traditional machine learning techniques including SVM with RBF and linear kernel, k-NN, RF, MLP, GNB and QDA act as classifier and incorporate with the tailored CNN model. The result of the experiment shows that all the survey traditional machine learning techniques fit well with each other. Among of them, the tailored CNN + k-NN method can recognise masked face images of both LFW-SMFRD and RMFRD by providing an outstanding performance that reaches 92.98% and 96.71% accuracy respectively.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	masked face recognition , tailored CNN , deep learning , classifier , machine learning
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	04 Dec 2024 02:50
Last Modified:	04 Dec 2024 02:50
URII:	http://shdl.mmu.edu.my/id/eprint/13212

Downloads

Edit (login required)