Issue No. 09 - September (2009 vol. 31)
Jingdong Wang , Microsoft Research Asia, Beijing
Fei Wang , Tsinghua University, Beijing
Changshui Zhang , Tsinghua University, Beijing
Helen C. Shen , The Hong Kong University of Science and Technology, Hong Kong
Long Quan , The Hong Kong University of Science and Technology, Hong Kong
In this paper, a novel graph-based transductive classification approach, called Linear Neighborhood Propagation, is proposed. The basic idea is to predict the label of a data point according to its neighbors in a linear way. This method can be cast into a second-order intrinsic Gaussian Markov random field framework. Its result corresponds to a solution to an approximate inhomogeneous biharmonic equation with Dirichlet boundary conditions. Different from existing approaches, our approach provides a novel graph structure construction method by introducing multiple-wise edges instead of pairwise edges, and presents an effective scheme to estimate the weights for such multiple-wise edges. To the best of our knowledge, these two contributions are novel for semi-supervised classification. The experimental results on image segmentation and transductive classification demonstrate the effectiveness and efficiency of the proposed approach.
Gaussian Markov random fields, linear neighborhood propagation, transductive classification, image segmentation.
J. Wang, H. C. Shen, F. Wang, L. Quan and C. Zhang, "Linear Neighborhood Propagation and Its Applications," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 31, no. , pp. 1600-1615, 2008.