Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004. (2004)
Washington, D.C., USA
June 27, 2004 to July 2, 2004
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/CVPR.2004.152
JinHyeong Park , Pennsylvania State University
Zhenyue Zhang , Zhejiang University
Hongyuan Zha , Pennsylvania State University
Rangachar Kasturi , University of South Florida
We propose methods for outlier handling and noise reduction using weighted local linear smoothing for a set of noisy points sampled from a nonlinear manifold. Weighted PCA is used as a building block for our methods and we suggest an iterative weight selection scheme for robust local linear fitting together with an outlier detection method based on minimal spanning trees to further improve robustness. We also develop an efficient and effective bias-reduction method to deal with the "trim the peak and fill the valley" phenomenon in local linear smoothing. Synthetic examples along with several image data sets are presented to show that manifold learning methods combined with weighted local linear smoothing give more accurate results.
R. Kasturi, Z. Zhang, J. Park and H. Zha, "Local Smoothing for Manifold Learning," Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004.(CVPR), Washington, D.C., USA, 2004, pp. 452-459.