In this paper we discuss the generalizations of the Kaotorovich inequality and obtain some generalized Kantorovich inequalities in the sense of matrix norm. We further illustrate how to use these inequalities to deter...In this paper we discuss the generalizations of the Kaotorovich inequality and obtain some generalized Kantorovich inequalities in the sense of matrix norm. We further illustrate how to use these inequalities to determine the lower bound of relative efficiency of the parameter estimate in linear model.展开更多
To avoid drawbacks of classic discrete Fourier transform(DFT)method,modern spectral estimation theory was introduced into harmonics and inter-harmonics analysis in electric power system.Idea of the subspace-based root...To avoid drawbacks of classic discrete Fourier transform(DFT)method,modern spectral estimation theory was introduced into harmonics and inter-harmonics analysis in electric power system.Idea of the subspace-based root-min-norm algorithm was described,but it is susceptive to noises with unstable performance in different SNRs.So the modified root-min-norm algorithm based on cross-spectral estimation was proposed,utilizing cross-correlation matrix and independence of different Gaussian noise series.Lots of simulation experiments were carried out to test performance of the algorithm in different conditions,and its statistical characteristics was presented.Simulation results show that the modified algorithm can efficiently suppress influence of the noises,and has high frequency resolution,high precision and high stability,and it is much superior to the classic DFT method.展开更多
针对目前大多数水印算法功能较单一,提出了一种可以同时进行版权保护和内容认证的多功能彩色图像双水印算法。首先,将原始图像由RGB空间转换到YUV空间,提取亮度分量的SURF(speeded up robust feature)特征点,根据特征点的主方向构造描...针对目前大多数水印算法功能较单一,提出了一种可以同时进行版权保护和内容认证的多功能彩色图像双水印算法。首先,将原始图像由RGB空间转换到YUV空间,提取亮度分量的SURF(speeded up robust feature)特征点,根据特征点的主方向构造描述向量并将其拆分为两个子向量,分别计算它们与一个参考向量之间的余弦夹角,通过比较夹角的大小关系来构造鲁棒零水印序列;然后对原始图像进行2×2分块,对子图像块奇异值范数进行异或运算来产生脆弱水印信息,并将其嵌入至图像空域的最低有效位。版权归属鉴定时通过计算原始鲁棒零水印序列和从待检测图像中提取的水印序列之间的BCR(bit correct rates)相关系数作为鉴定的依据,内容认证时通过比较图像最低有效位和脆弱水印信息是否一致来实现篡改的检测和定位。实验结果表明,本文算法具有良好的透明性和较高的计算效率,同时具备版权保护和内容认证的双重功能。展开更多
平滑范数(Smoothed l0,SL0)压缩感知重构算法通过引入平滑函数序列将求解最小l0范数问题转化为平滑函数优化问题,可以有效地用于稀疏信号重构。针对平滑函数的选取和算法稳健性问题,提出一种新的平滑函数序列近似范数,结合梯度投影法优...平滑范数(Smoothed l0,SL0)压缩感知重构算法通过引入平滑函数序列将求解最小l0范数问题转化为平滑函数优化问题,可以有效地用于稀疏信号重构。针对平滑函数的选取和算法稳健性问题,提出一种新的平滑函数序列近似范数,结合梯度投影法优化求解,并进一步提出采用奇异值分解(Singular value decomposition,SVD)方法改进算法的稳健性,实现稀疏度信号的精确重构。仿真结果表明,在相同的测试条件下,本文算法相比OMP算法、SL0算法以及L1-magic算法在重构精度、峰值信噪比方面都有较大改善。展开更多
文摘In this paper we discuss the generalizations of the Kaotorovich inequality and obtain some generalized Kantorovich inequalities in the sense of matrix norm. We further illustrate how to use these inequalities to determine the lower bound of relative efficiency of the parameter estimate in linear model.
基金Shandong University of Science and Technology Research Fund(No.2010KYTD101)
文摘To avoid drawbacks of classic discrete Fourier transform(DFT)method,modern spectral estimation theory was introduced into harmonics and inter-harmonics analysis in electric power system.Idea of the subspace-based root-min-norm algorithm was described,but it is susceptive to noises with unstable performance in different SNRs.So the modified root-min-norm algorithm based on cross-spectral estimation was proposed,utilizing cross-correlation matrix and independence of different Gaussian noise series.Lots of simulation experiments were carried out to test performance of the algorithm in different conditions,and its statistical characteristics was presented.Simulation results show that the modified algorithm can efficiently suppress influence of the noises,and has high frequency resolution,high precision and high stability,and it is much superior to the classic DFT method.
文摘平滑范数(Smoothed l0,SL0)压缩感知重构算法通过引入平滑函数序列将求解最小l0范数问题转化为平滑函数优化问题,可以有效地用于稀疏信号重构。针对平滑函数的选取和算法稳健性问题,提出一种新的平滑函数序列近似范数,结合梯度投影法优化求解,并进一步提出采用奇异值分解(Singular value decomposition,SVD)方法改进算法的稳健性,实现稀疏度信号的精确重构。仿真结果表明,在相同的测试条件下,本文算法相比OMP算法、SL0算法以及L1-magic算法在重构精度、峰值信噪比方面都有较大改善。
