In this paper, an integrated method of Local Binary Pattern (LBP) and Discrete Cosine Transform (DCT) is used for face recognition under different variations with illumination, pose and facial expression. LBP is applied as a preprocessing step to generate the LBP image which is then analyzed on a block-by-block basis. Facial features are extracted from each block using the DCT algorithm. The objective is to explore the performance of LBP under different illumination, pose and expression variations. Face images in the Yale database and ORL database are used in our experiments. A comparison on the sensitivities of the LBP operator in illumination, pose and expression variations is provided.
展开▼
