A Novel Hybrid Approach to Masked Face Recognition using Robust PCA and GOA Optimizer

Abstract

The use of face masks has become ubiquitous across a wide range of industries and professions, including healthcare, food service, construction, manufacturing, retail, hospitality, transportation, education, and public safety. To overcome this challenge, masked face recognition has emerged as a vital technology in recent years. By utilizing advanced algorithms and deep learning techniques, masked face recognition enables accurate identification and authentication of individuals even in scenarios where masks are worn. This paper presents a novel method for recognizing faces with masks. The proposed method integrates deep learning-based mask detection, landmark and oval face detection, and robust principal component analysis (RPCA) to accurately identify and authenticate individuals wearing masks. A pretrained ssd-MobileNetV2 model is utilized to detect the presence and location of masks on a face, while landmark and oval face detection are used to identify and extract important facial features. RPCA is applied to separate the occluded and non-occluded components of an image, making the method more reliable in identifying faces with masks. To further optimize the performance of the proposed method, the Gazelle Optimization Algorithm (GOA) is used to optimize both the KNN features and the number of k for KNN. Experimental results demonstrate that the proposed method outperforms existing methods in terms of accuracy and robustness to occlusion, achieving a recognition rate of 97%. This represents a significant improvement over existing methods for masked face recognition. The proposed method has the potential to be applied in a wide range of real-world scenarios, such as security systems, access control, and public health measures. The results of this study demonstrate that the integration of deep learning-based mask detection, landmark and oval face detection, and RPCA can improve the accuracy and reliability of masked face recognition, even in challenging and complex environments. The proposed method can be further improved and extended in future research to address other challenges in this field.