期刊文献+

一种增量式类内局部保持降维算法

Incremental Within-class Locality Preserving Dimension Reduction Algorithm
下载PDF
导出
摘要 针对在线学习中的算法效率问题,提出了一种增量式类内局部保持降维算法。该算法综合考虑了基于QR分解的降维算法与保类内Fisher判别分析法的优点,根据训练过程中新增的样本进行投影矩阵在线更新,克服了传统的批量式训练方法在线学习时计算量过分冗余的缺陷。同时,通过兼顾输入样本的局部结构和全局分布状态,使得该算法能够有效地应用于多簇、重叠的数据形态。在ORL人脸库和COIL20图像库上的实验表明,该增量式算法不仅在降维效果上基本与批量式算法保持一致,而且具有较大的效率优势。 针对在线学习中的算法效率问题,提出了一种增量式类内局部保持降维算法。该算法综合考虑了基于QR分解的降维算法与保类内Fisher判别分析法的优点,根据训练过程中新增的样本进行投影矩阵在线更新,克服了传统的批量式训练方法在线学习时计算量过分冗余的缺陷。同时,通过兼顾输入样本的局部结构和全局分布状态,使得该算法能够有效地应用于多簇、重叠的数据形态。在ORL人脸库和COIL20图像库上的实验表明,该增量式算法不仅在降维效果上基本与批量式算法保持一致,而且具有较大的效率优势。
出处 《计算机科学》 CSCD 北大核心 2012年第S3期154-158,共5页 Computer Science
基金 国家自然科学基金项目(61070043)资助
关键词 在线学习 局部保持 特征降维 On-line learning Local features preservation Dimension reduction
  • 引文网络
  • 相关文献

参考文献12

  • 1郑建炜,王万良.一种新的保类内核Fisher判别法及说话人辨别应用[J].计算机科学,2010,37(7):243-247. 被引量:1
  • 2Haesun Park,Moongu Jeon,J. Ben Rosen.Lower Dimensional Representation of Text Data Based on Centroids and Least Squares[J]. Bit Numerical Mathematics . 2003 (2)
  • 3Turk M,Pentland A.Eigenfaces for recognition. Journal of Cognitive Neuroscience . 1991
  • 4Mika S,R(?)tsch G,Weston J,et al.Fisher discriminant analysis with kernels. Nerual Networks for Signal Processing Workshop . 1999
  • 5Swets DL,Weng J.Using discriminant eigenfeatures for image retrieval. IEEE Transactions on Pattern Analysis and Machine Intelligence . 1996
  • 6Sugiyama M.Dimensionality reduction of multimodal labeled data by local Fisher discriminant analysis. Journal of Machine Learning Research . 2007
  • 7Huang H,Li J W,Liu J M.Enhanced semi-supervised local Fisher discriminantanalysis for face recognition. Future Generation Computer Systems . 2012
  • 8Orriols-Puig A,Casillas J.Fuzzy knowledge representation studyfor incremental learning in data streams and classification prob-lems. Soft Computing . 2011
  • 9Clemmensen LH,Hansen M E,Ersboll B K.A comparison of di-mension reduction methods with application to multi-spectralimages of sand used in concrete. Machine Vision and Applications . 2010
  • 10Kim T K,Stenger B,Kittler J,et al.Incremental Linear Discrimin-ant Analysis Using Sufficient Spanning Sets and Its Applications. International Journal of Computer Vision . 2011

二级参考文献16

  • 1高秀梅,杨静宇,杨健.一种最优的核Fisher鉴别分析与人脸识别[J].系统仿真学报,2004,16(12):2864-2868. 被引量:13
  • 2徐勇,杨静宇,金忠,娄震.一种基于核的快速非线性鉴别分析方法[J].计算机研究与发展,2005,42(3):367-374. 被引量:9
  • 3Hinton G E,Salakhutdinov R R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5786):504-507.
  • 4Baudat G,Anouar F.Generalized discriminate analysis using a kernel approach[J].Neural Computation,2000,12(10):2385-2404.
  • 5Belhumeur P N,Hespanha J P,Kriegam D J.Eigenfaces vs.Fisherfaces:Recognition using class specific linear projection[J].IEEE Trans.on Pattern Anal.Machine Intell,1997,19(7):711-720.
  • 6Yang J,Yang J-Y.Why can LDA be performed in PCA transformed space?[J].Pattern Recognition,2003,36(2):563-566.
  • 7He X F,Niyogi P.Locality Preserving Projections[A] ∥Thrun S,Saul L,Scholkopf B,eds.Advances in Neural Information Processing Systems 16[C].Cambridge,MA:MIT Press,2004.
  • 8Sugiyama M.Dimensionality Reduction of Multimodal Labeled Data by Local Fisher Discriminant Analysis[J].Machine Lear-ning Research,2007,8:1027-1061.
  • 9Liu Qing-Shan,Huang Rui.Face recognition using kernel based fisher discriminant analysis[A] ∥Proceedings of the 5th International Conference on Automatic Face and Gesture Recognition[C].Washington,DC,USA,2002:187-191.
  • 10Liu Wei,Wang Yun-hong.Null space based kernel fisher discriminant analysis for face recognition[A] ∥Proceeding of the 6th International Conference on Automatic Face and Gesture Recognition[C].Seoul,Korea,2004:369-374.
;
使用帮助 返回顶部