Abstract: In this paper,considering kernel and parameter-less nearest-neighbor graph,a novel method named parameter-less supervised kernel locality preserving projection algorithm which aims at discovering an embedding that preserves nonlinear information was proposed for face representation and recognition.In this algorithm,firstly,by changing the Euclidean distance to the Cosine distance which is more robust to outliner,and constructing a parameter-less nearest-neighbor graph,this algorithm uses the nonlinear kernel mapping to map the face data into an implicit feature space.And then a linear transformation is preformed to preserve locality geometric structures of the face image,which solves the difficulty of parameter selection in computing affinity matrix.Experiments based on both ORL and Yale face database demonstrate the effectiveness of the new algorithm.

Key words: Face recognition,Feature extraction,Locality preserving projection,Parameter-less nearest-neighbor graph,Kernel method