Discriminant Pairwise Local Embeddings

This paper introduces Discriminant Pairwise Local Embeddings (DPLE) a supervised dimensionality reduction technique that generates structure preserving discriminant subspaces. This objective is achieved through a convex optimization formulation where Euclidean distances between data pairs that belong to the same class are minimized, while those of pairs belonging to different classes are maximized. These pairwise relations are encoded in two matrices and weighted with the data affinity matrix to ensure local structure preservation. The discriminant efficiency of our technique is demonstrated in two popular applications, face and sketch recognition, where DPLE outperforms competitive manifold learning algorithms. A kernelized version of DPLE, that further enhances recognition accuracy, is also explained.