计算机科学 ›› 2015, Vol. 42 ›› Issue (11): 90-93.doi: 10.11896/j.issn.1002-137X.2015.11.018

• 2014年全国高性能计算机学术年会 • 上一篇    下一篇

基于GPU并行加速的多特征融合的超图降维方法

洪朝群,陈旭辉,王晓栋,李士锦,吴克寿   

  1. 厦门理工学院计算机与信息工程学院 厦门361024,厦门理工学院计算机与信息工程学院 厦门361024,厦门理工学院计算机与信息工程学院 厦门361024,厦门理工学院计算机与信息工程学院 厦门361024,厦门理工学院计算机与信息工程学院 厦门361024
  • 出版日期:2018-11-14 发布日期:2018-11-14
  • 基金资助:
    本文受国家自然科学基金(61202145),福建省自然科学基金(2014J01256)资助

Hypergraph Dimensionality Reduction with Multiple Feature Fusion Based on GPU Parallel Acceleration

HONG Chao-qun, CHEN Xu-hui, WANG Xiao-dong, LI Shi-jin and WU Ke-shou   

  • Online:2018-11-14 Published:2018-11-14

摘要: 基于图的学习方法目前广泛用于降低特征维度。然而,对于多特征数据而言,不同特征之间的不同关联性很难结合到单个图中。针对多特征数据提出了新的半监督降维方法。首先,以超图中的超边作为片,使超图应用到片对齐框架中。然后,通过统计片中相邻的特征对的距离计算超边的权重,使得不同特征下的片得到结合。其次,由于欧氏距离和矩阵乘法的计算在拉普拉斯矩阵的构造过程中占用了大部分的时间,因此使用GPU对其进行加速。实验结果表明了所提方法在分类性能和学习速度上的提升效果。

关键词: 降维,多特征融合,片对齐框架,超图学习,基于GPU的并行加速

Abstract: Graph-based learning methods are currently popular for dimensionality reduction.However,for multiple feature data,different relationships from different features are hard to be integrated into a single graph.In this paper,a novel semi-supervised dimensionality reduction method was proposed for multiple feature data.First,the hyperedges in hypergraph are assumed as patches.In this way,hypergraph is applied to patch alignment framework.Then,the weights of hyperedges are computed with statistics of distances between neighboring pairs and the patches from different features are integrated.Second,the speed of computing Euclidean distances and matrix multiplication is improved by using GPU,since they take most of time in constructing the Laplacian matrix.The experimental results demonstrate the improvement on both classification performance and learning speed.

Key words: Dimensionality reduction,Multiple feature fusion,Patch alignment framework,Hypergraph learning,GPU-based parallel acceleration

[1] Yu Jun,Tao Da-cheng,Rui Yong,et al.Pairwise constraintsbased multiview features fusion for scene classification[J].Pattern Recognition,2013,46(2):483-496
[2] Yu Jun,Wang Meng,Tao Da-cheng.Semisupervised multiview distance metric learning for cartoon synthesis[J].IEEE Transactions on Image Process,2012,21(11):4636-4648
[3] Joliffe I.Principal component analysis[M].New York:Springer-Verlag,1986
[4] Belhumeur P N,Hespanha J P,Kriegman D J.Eigenfaces vs.Fisherfaces:Recognition using class specific linear projection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1997,19(7):711-720
[5] Roweis S,Saul L.Nonlinear dimensionality reduction by locally linear embedding[J].Science,2000,290(5500):2323-2326
[6] Tenenbaum V S J,Langford J.A global geometric framework for nonlinear dimensionality reduction[J].Science,2000,290(5500):2319-2323
[7] Belkin M,Niyogi P.Laplacian eigenmaps and spectral techniques for embedding and clustering[M]∥Advances in Neural Information Processing System 14.Cambridge:MIT Press,2001:585-591
[8] Donoho D L,Grimes C.Hessian Eigenmaps:New Locally Linear Embedding Techniques for High-dimensional Data[J].National Academy of Sciences,2003,100(10):5591-5596
[9] Zhang Z,Zha H.Principal manifolds and nonlinear dimension reduction via local tangent space alignment[J].SIAM J.Scientific Computing,2005,26(1):313-338
[10] Zhou Deng-yong,Huang Jia-yuan,Scholkopf B.Learning withhypergraphs:clustering,classification,and embedding[C]∥Proc.of Advances in Neural Information Processing Systems.Cambridge:MIT Press,2006:1601-1608
[11] Zhang Tian-hao,Tao Da-cheng,Li Xue-long,et al.Patch alignment for dimensionality reduction[J].IEEE Transactions on Knowledge and Data Engineering,2009,21(9):1299-1313
[12] NVidia Corporation.CUDA C PROGRAMMING GUIDE v6.0[M].2014
[13] Everingham M,Van Gool L,et al.The PASCAL Visual Object Classes Challenge 2007(VOC2007) Results.http://www.pascal-network.org/challenges/VOC/voc2007/workshop/index.html
[14] Vapnik V N.Statistical Learning Theory[M].New York:Wiley-Interscience,1998
[15] Agoston M K.Computer Graphics and Geometric Modeling:Implementation and Algorithms[M].London:Springer,2005:300-306
[16] Bosch A,Zisserman A,Munoz X.Image classification using random forests and ferns[C]∥Proceedings of International Confe-rence on Computer Vision.New York:IEEE,2007:1-8
[17] Siagian C,Itti L.Rapid Biologically-Inspired Scene Classification Using Features Shared with Visual Attention[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2007,29(2):300-312
[18] Xia Tian,Tao Da-cheng,Mei Tao,et al.Multiview spectral embedding[J].IEEE Transactions on Systems,Man,and Cyberne-tics,Part B:Cybernetics,2010,40(6):1438-1446
[19] Cheng Bin,Liu Guang-can,Wang Jing-dong,et al.Multi-task Low-rank Affinity Pursuit for Image Segmentation[C]∥Proceedings of IEEE International Conference on Computer Vision.New York:IEEE,2011:2439-2446

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!