In order to reduce the impact of changes in the expression of 3D face recognition, presents a coarse to fine identification methods. Takes depth data as the overall features of a human face, uses Fisherface to match these features. The facial rigid region as local feature is matched using the modified iterative closest point algorithm. The extracted matching results of the global and local features are fused. The experimental result shows that the method has better robustness to facial expression change.
3D Face RecognitionFisherface(PCA+LDA)Depth DataRigid RegionICP