Periocular Biometric Recognition for Masked Faces

Since the outbreak of Coronavirus Disease 2019 (COVID-19), people are recommended to wear facial masks limit spread virus. Under circumstances, traditional face recognition technologies cannot achieve satisfactory results. In this paper, we propose a algorithm that combines features and deep masked faces. For features, extract Local Binary Pattern (LBP), Scale-Invariant Feature Transform (SIFT) Histogram Oriented Gradient (HOG) from periocular region, use Support Vector Machines (SVM) classifier perform personal identification. We also an improved Convolutional Neural Network (CNN) model Angular Visual Geometry Group (A-VGG) learn features. Then decision-level fusion combine four Comprehensive experiments were carried out on databases real faces simulated faces, including frontal side taken at different angles. Images with motion blur tested evaluate robustness algorithm. Besides, experiment matching corresponding full is accomplished. The experimental results show proposed has state-of-the-art performance in recognition, region rich biological high discrimination.