Issue No. 09 - September (2011 vol. 33)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TPAMI.2011.39
Ruiping Wang , Chinese Academy of Sciences, Beijing
Shiguang Shan , Chinese Academy of Sciences, Beijing
Xilin Chen , Chinese Academy of Sciences, Beijing
Jie Chen , University of Oulu, Oulu
Wen Gao , Peking University, Beijing
Over the past few decades, dimensionality reduction has been widely exploited in computer vision and pattern analysis. This paper proposes a simple but effective nonlinear dimensionality reduction algorithm, named Maximal Linear Embedding (MLE). MLE learns a parametric mapping to recover a single global low-dimensional coordinate space and yields an isometric embedding for the manifold. Inspired by geometric intuition, we introduce a reasonable definition of locally linear patch, Maximal Linear Patch (MLP), which seeks to maximize the local neighborhood in which linearity holds. The input data are first decomposed into a collection of local linear models, each depicting an MLP. These local models are then aligned into a global coordinate space, which is achieved by applying MDS to some randomly selected landmarks. The proposed alignment method, called Landmarks-based Global Alignment (LGA), can efficiently produce a closed-form solution with no risk of local optima. It just involves some small-scale eigenvalue problems, while most previous aligning techniques employ time-consuming iterative optimization. Compared with traditional methods such as ISOMAP and LLE, our MLE yields an explicit modeling of the intrinsic variation modes of the observation data. Extensive experiments on both synthetic and real data indicate the effectivity and efficiency of the proposed algorithm.
Dimensionality reduction, manifold learning, maximal linear patch, landmarks-based global alignment.
X. Chen, S. Shan, W. Gao, J. Chen and R. Wang, "Maximal Linear Embedding for Dimensionality Reduction," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 33, no. , pp. 1776-1792, 2011.