Robust face recognition under partial occlusion based on support vector machine with local Gaussian summation kernel

This paper presents the use of Support Vector Machine (SVM) with local Gaussian summation kernel for robust face recognition under partial occlusion. In recent years, the effectiveness of SVM and local features has been reported. However, because conventional methods apply one kernel to global features and global features are influenced easily by noise or occlusion, the conventional methods are not robust to occlusion. The recognition method based on local features, however, is robust to occlusion because partial occlusion affects only specific local features. In order to utilize this property of local features in SVM, local kernels are applied to local features. The use of local kernels in SVM requires local kernel integration. The summation of local kernels is used as the integration method in this study. The effectiveness and robustness of the proposed method are shown by comparison with global kernel based SVM. The recognition rate of the proposed method is high under large occlusion, whereas the recognition rate of the SVM with the global Gaussian kernel decreases drastically. Furthermore, we investigate the robustness to practical occlusion in the real world using the AR face database. Although only face images with non-occlusion are used for training, faces wearing sunglasses or a scarf are classified with high accuracy.