融合局部和全局结构的流形学习

黄鸿; 李见为; 冯海亮

doi:null

您当前的位置：

首页 >

文章列表页 >

融合局部和全局结构的流形学习

信息科学 | 更新时间：2020-08-12

- 融合局部和全局结构的流形学习
- Fusion of local and globle structures for manifold learning
- 光学精密工程 2009年17卷第3期页码：626-632
- 作者机构：
  
  1. 重庆大学光电技术及系统教育部重点实验室重庆,400044
  2. 重庆工学院教授流动站重庆,400050
- 作者简介：
- 基金信息：
  
  重庆市自然科学基金资助项目(No.CSTC2006BB215();No.CSTC2008BB2160)()
- DOI：
  中图分类号： TP391.4
- 收稿日期：2008-03-30，
  
  修回日期：2008-05-16，
  
  网络出版日期：2009-03-25，
  
  纸质出版日期：2009-03-25
- 稿件说明：
移动端阅览
黄鸿, 李见为, 冯海亮. 融合局部和全局结构的流形学习[J]. 光学精密工程, 2009,17(3): 626-632

HUANG Hong, LI Jian-wei, FENG Hai-liang. Fusion of local and globle structures for manifold learning[J]. Editorial Office of Optics and Precision Engineering, 2009,17(3): 626-632
黄鸿, 李见为, 冯海亮. 融合局部和全局结构的流形学习[J]. 光学精密工程, 2009,17(3): 626-632 DOI：

HUANG Hong, LI Jian-wei, FENG Hai-liang. Fusion of local and globle structures for manifold learning[J]. Editorial Office of Optics and Precision Engineering, 2009,17(3): 626-632 DOI：

摘要

提出了一种融合局部与全局结构的保持嵌入(LGPE)算法。该方法首先假定目标空间的整体映射函数

然后结合数据的全局几何结构分布信息进行数据重构

最后通过最小化准则函数来得到嵌入高维空间的低维子流形。LGPE方法在保持数据局部结构的同时保留了全局结构信息

在信噪比为10 dB的稀疏Swiss-roll(

=400)和COIL-20多姿态数据集上都取得了较好的维数约简效果。与其他局部流形学习方法相比

该方法在AT

T人脸图像库中

当嵌入特征矢量维数

40时

其识别率提高了约15%。在人工与真实数据库的实验结果表明

本文方法对噪声和稀疏数据具有较好的鲁棒性。

Abstract

A new method called Local and Global Preserving Embedding (LGPE) is proposed for manifold learning. This method assumes a global embedding function in low dimensional space

then incorporates the relative compactness information of the data distributions on the global geometry to reconstruct sample data. Finally

the global low dimensional submanifold is obtained by minimizing the cost function.The LGPE preserves the local and global structures of the data points simultaneously

and can obtain better dimensionality reduction on the sparse Swiss-roll dataset with noises (

=400

SNR=10 dB) and COIL-20 multi-poses dataset.When it is used in the AT

T face dateset

the recognition rate can be improved by 15% as compared with that of other local manifold methods under condition of embedding dimension lower than 40. The experimental results on both synthetic and real data sets show that proposed method is effectiveness and robustness for noise and sparse data.

关键词

Keywords

references

尹峻松, 肖健,周宗潭,等. 非线性流形学习方法的分析与应用[J]. 自然科学展,2007,17(8):1015-1025. YIN J S,XIAO J,ZHOU Z T,et al..Analysis and application of non-linear manifold learning method[J]. Progress in Natural Science, 2007,17(8):1015-1025.(in Chinese)[2] 赵连伟, 罗四维, 赵艳敞, 等. 高维数据流形的低维嵌入及嵌入维数研究[J]. 软件学报,2005,16(8):1423-1430. ZHAO L W, LUO S W, ZHAO Y CH, et al.. Study on the low-dimensional embedding and the embedding dimensionality of manifold of high-dimensional data [J]. Journal of Software, 2005,16(8):1423-1430.(in Chinese)[3] 李粉兰, 徐可欣. 一种应用于人脸正面图像的眼睛自动定位算法[J]. 光学精密工程,2006,14(2):320-326. LI F L, XU K X. An algorithm applied in frontal-view face images for automatically localizing eyes [J]. Opt. Precision Eng.,2006,14(2):320-326. (in Chinese)[4] 李粉兰, 唐文彦, 段海峰,等. 分数次幂多项式核函数在核直接判别式分析中的应用[J]. 光学精密工程,2007,15(9):1410-1414. LI F L, TANG W Y, DUAN H F, et al.. Application of fractional power polynomial kernel function to kernel direct discriminant analysis [J]. Opt. Precision Eng., 2007,15(9):1410-1414.(in Chinese)[5] TENENBAUM J B,SILVA V D,LANGFORD J C. A global geometric framework for nonlinear dimensionality reduction [J]. Science, 2000,290(5500):2319-2323.[6] ROWEIS S T,SAUL L K. Nonlinear dimensionality reduction by locally linear embedding [J]. Science, 2000,290(5500):2323-2326.[7] BELKIN M,NIYOGI P. Laplacian eigenmaps and spectral techniques for embedding and clustering . In Advances in Neural Information Processing Systems, Cambridge, MA, USA, 2002(14):585-591.[8] ZHANG Z Y, ZHA H Y. Principal manifolds and nonlinear dimensionality reduction via local tangent space alignment [J]. SIAM Journal of Scientific Computing, 2004,26(1):313-338.[9] HE X F,NIYOGI P. Locality preserving projections . In Proceedings of Neural Information Processing System, Vancouver, Canada,2003:153-160.[10] 曾宪华, 罗四维. 动态增殖流形学习算法[J]. 计算机研究与发展, 2007,44(9):1462-1468. ZENG X H, LUO S W. A dynamically incremental manifold learning algorithm [J]. Journal of Computer Research and Development, 2007,44(9):1462-1468.(in Chinese)[11] 曹顺茂, 叶世伟. 一种在源数据稀疏情况下的流形学习算法研究[J]. 计算机仿真, 2007,24(3):104-106. CAO SH M, YE SH W. A manifold learning algorithm in a state of sparse source data [J]. Computer Simulation, 2007,24(3):104-106.(in Chinese)[12] XUE H, CHEN S C. Alternative robust local embedding . In Proceedings of International Conference on Wavelet Analysis and Pattern Recognition, Beijing, China, 2007:591-596. [13] 梁毅雄, 龚卫国, 潘英俊,等. 基于奇异值分解的人脸识别方法[J]. 光学精密工程, 2004,12(5):543-549. LIANG Y X, GONG W G, PANG Y J, et al.. Singular value decomposition-based approach for face recognition [J]. Opt. Precision Eng., 2004,12(5):543-549.(in Chinese)

浏览量

432

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于半监督稀疏鉴别嵌入的高光谱遥感影像分类

基于稀疏鉴别嵌入的高光谱遥感影像分类

面向高光谱图像分类的半监督双流网络

面向高光谱影像分类的深度流形重构置信网络

面向高光谱影像分类的多特征流形鉴别嵌入