Face recognition algorithm based on multi-scale CLBP

Digital surveillance systems are extensively being used in our daily lives, and there is an increasingly demand for face recognition technology based on monitoring scenarios. In this paper, we propose a recognition algorithm based on multi-scale completed local binary pattern (CLBP) operators which can be used for face recognition tasks. The CLBP operator which is initially used in texture recognition has achieved great performance. The key idea of CLBP is using three CLPB operators to construct mixed feature representation. On this basis, the concept of multi-scale CLBP mixed-feature construction is proposed, which extracts the same feature on different pixel-scale levels and fused as one optimal feature. The multi-scale CLBP descriptor contains much more information and has great discriminative power. We investigate the performance of our algorithm on public dataset. Extensive experiments results demonstrate that our multi-scale CLBP algorithm outperforms other state-of-art operators.