Local discriminant embedding and its variants

@inproceedings{chen05, abstract = {We present a new approach, called local discriminant embedding (LDE), to manifold learning and pattern classification. In our framework, the neighbor and class relations of data are used to construct the embedding for classification problems. The proposed algorithm learns the embedding for the submanifold of each class by solving an optimization problem. After being embedded into a low-dimensional subspace, data points of the same class maintain their intrinsic neighbor relations, whereas neighboring points of different classes no longer stick to one another. Via embedding, new test data are thus more reliably classified by the nearest neighbor rule, owing to the locally discriminating nature. We also describe two useful variants: two-dimensional LDE and kernel LDE. Comprehensive comparisons and extensive experiments on face recognition are included to demonstrate the effectiveness of our method.}, author = {Chen, Hwann-Tzong and Chang, Huang-Wei and Liu, Tyng-Luh }, booktitle = {Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on}, citeulike-article-id = {2310509}, doi = {10.1109/CVPR.2005.216}, journal = {Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on}, keywords = {chen05}, pages = {846--853 vol. 2}, posted-at = {2008-03-26 16:22:44}, priority = {0}, title = {Local discriminant embedding and its variants}, url = {http://dx.doi.org/10.1109/CVPR.2005.216}, volume = {2}, year = {2005} }

TY - CONF ID - chen05 L3 - citeulike-article-id:2310509 TI - Local discriminant embedding and its variants T2 - Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on JF - Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on VL - 2 SP - 846 EP - 853 vol. 2 N2 - We present a new approach, called local discriminant embedding (LDE), to manifold learning and pattern classification. In our framework, the neighbor and class relations of data are used to construct the embedding for classification problems. The proposed algorithm learns the embedding for the submanifold of each class by solving an optimization problem. After being embedded into a low-dimensional subspace, data points of the same class maintain their intrinsic neighbor relations, whereas neighboring points of different classes no longer stick to one another. Via embedding, new test data are thus more reliably classified by the nearest neighbor rule, owing to the locally discriminating nature. We also describe two useful variants: two-dimensional LDE and kernel LDE. Comprehensive comparisons and extensive experiments on face recognition are included to demonstrate the effectiveness of our method. KW - chen05 AU - Chen, Hwann-Tzong AU - Chang, Huang-Wei AU - Liu, Tyng-Luh PY - 2005/// UR - http://dx.doi.org/10.1109/CVPR.2005.216 ER -