Face Recognition based on Multi-shape Morphological Profiles-based Covariance Descriptors and Log- Etuclidean Kernel SVM

Face recognition (FR) is one of the important areas of image processing that allows us to recognize individuals from gathered face images. It has been shown in the literature that contextual information of the face can improve the performance of FR algorithms. This paper proposes an efficient face recognition method based on multi-shape morphological profiles (MMPs), covariance descriptors, and log-Euclidean kernel SVM. In the first stage of this method, MMPs containing contextual information are generated from the original faces. The covariance descriptors of the MMPs are produced in the subsequent stage. Finally, these covariance descriptors are classified using log-Euclidean kernel SVM. To compare the findings, we ran our tests on ORL face dataset. The mean accuracy of the proposed method is 96.4 on the ORL face datasets, respectively, demonstrating that the proposed method outperforms some of the existing state-of-the-art FR methods.