针对非监督线性差分投影(unsupervised linear differential projection,ULDP)在特征提取过程中存在的不足,提出了基于多流形的非监督线性差分投影(multi-manifold unsupervised linear differential projection,MULDP)算法,并将其应用...针对非监督线性差分投影(unsupervised linear differential projection,ULDP)在特征提取过程中存在的不足,提出了基于多流形的非监督线性差分投影(multi-manifold unsupervised linear differential projection,MULDP)算法,并将其应用于人脸识别中。MULDP首先构造出多流形局部近邻图和多流形最大全局方差,然后通过多目标最优化问题求解出嵌入在高维空间的低维流形。这种映射不仅能表示全局结构,还能表示局部结构。该算法可以得到嵌入在高维空间的低维流形,更好地实现了局部与全局结构信息的有效保持。在ORL、Yale及AR人脸库上的实验结果验证了所提算法的优越性。展开更多
Nowadays orthogonal arrays play important roles in statistics, computer science, coding theory and cryptography. The usual difference matrices are essential for the construction for many mixed orthogonal arrays. But t...Nowadays orthogonal arrays play important roles in statistics, computer science, coding theory and cryptography. The usual difference matrices are essential for the construction for many mixed orthogonal arrays. But there are also orthogonal arrays which cannot be obtained by the usual difference matrices, such as mixed orthogonal arrays of run size 60. In order to construct these mixed orthogonal arrays, a class of special so-called generalized difference matrices were discovered by Zhang (1989,1990,1993,2006) from the orthogonal decompositions of projection matrices. In this article, an interesting equivalent relationship between orthogonal arrays and the generalized difference matrices is presented and proved. As an application, a lot of new orthogonal arrays of run size 60 have been constructed.展开更多
文摘针对非监督线性差分投影(unsupervised linear differential projection,ULDP)在特征提取过程中存在的不足,提出了基于多流形的非监督线性差分投影(multi-manifold unsupervised linear differential projection,MULDP)算法,并将其应用于人脸识别中。MULDP首先构造出多流形局部近邻图和多流形最大全局方差,然后通过多目标最优化问题求解出嵌入在高维空间的低维流形。这种映射不仅能表示全局结构,还能表示局部结构。该算法可以得到嵌入在高维空间的低维流形,更好地实现了局部与全局结构信息的有效保持。在ORL、Yale及AR人脸库上的实验结果验证了所提算法的优越性。
基金supported by Visiting Scholar Foundation of Key Lab in University and by National Natural Science Foundation of China (Grant No. 10571045)Specialized Research Fund for the Doctoral Program of Higher Education of Ministry of Education of China (Grant No. 44k55050)
文摘Nowadays orthogonal arrays play important roles in statistics, computer science, coding theory and cryptography. The usual difference matrices are essential for the construction for many mixed orthogonal arrays. But there are also orthogonal arrays which cannot be obtained by the usual difference matrices, such as mixed orthogonal arrays of run size 60. In order to construct these mixed orthogonal arrays, a class of special so-called generalized difference matrices were discovered by Zhang (1989,1990,1993,2006) from the orthogonal decompositions of projection matrices. In this article, an interesting equivalent relationship between orthogonal arrays and the generalized difference matrices is presented and proved. As an application, a lot of new orthogonal arrays of run size 60 have been constructed.