大数据时代背景下,随着所获数据数量和维度的不断增加,高维数据的处理成为聚类分析的重点和难点.基于同一类别高维数据通常分布在高维环绕空间的低维子空间这一事实,子空间聚类成为高维数据聚类分析领域的重要方法.稀疏子空间聚类(Spars...大数据时代背景下,随着所获数据数量和维度的不断增加,高维数据的处理成为聚类分析的重点和难点.基于同一类别高维数据通常分布在高维环绕空间的低维子空间这一事实,子空间聚类成为高维数据聚类分析领域的重要方法.稀疏子空间聚类(Sparse Space Clustering,SSC)通过交替方向乘子法(Alternating Direction Method of Multipliers,ADMM)对数据矩阵的稀疏自表达系数进行求解,发现分布于低维子空间并集中的数据的稀疏表示并进行聚类.但是ADMM参数多、收敛速度慢,其效率难以满足对大规模数据库进行聚类分析的要求.针对这一问题提出了基于L_(0)约束的稀疏子空间聚类方法,该方法使用正交匹配追踪(Orthogonal Matching Pursuit,OMP)算法求解L_(0)约束的自表达稀疏重建问题,构建数据集中各数据之间的相关性矩阵,最终对相关性矩阵应用谱聚类方法得到聚类结果.根据OMP算法每次迭代之间的耦合关系对其进行优化,进一步降低了计算复杂度,提高了算法效率.在生成数据和Extended Yale B database人脸数据库的实验结果表明,该算法与SSC相比,在显著减少计算时间的基础上,取得了与SSC相当的聚类准确率.展开更多
l_(0)梯度最小化图像平滑算法可在保持边缘的同时滤除纹理和细节,但该算法使用图像梯度判决被平滑成分时会出现包含较小图像梯度(弱边缘)的区域会被平滑,而包含较大图像梯度(强纹理)的区域被保留的现象.为克服此缺陷,提出一种基于图像块...l_(0)梯度最小化图像平滑算法可在保持边缘的同时滤除纹理和细节,但该算法使用图像梯度判决被平滑成分时会出现包含较小图像梯度(弱边缘)的区域会被平滑,而包含较大图像梯度(强纹理)的区域被保留的现象.为克服此缺陷,提出一种基于图像块l_(0)梯度最小化算法(image-patch based l_(0)gradient minimization algorithm,简称IP-l_(0)算法)的图像平滑算法,通过对输入图像中的图像块而非整幅图像进行平滑,动态改变图像块目标函数中的权重参数,令主要包含强纹理的图像块以较大的力度进行平滑,而主要包含弱边缘的图像块以较小的力度进行平滑,再整合平滑后的图像块得到整个边缘保持平滑图像.对IP-l_(0)算法、原始的l_(0)梯度最小化算法、基于局部拉普拉斯滤波器的算法、基于相对全变差算法、基于树滤波的算法,以及2种基于深度学习的边缘保持算法进行仿真实验,结果表明,使用IP-l_(0)算法滤波后的图像能在保持较弱的边缘的同时平滑强纹理.展开更多
为了进一步提高极限学习机(extreme learning machine,ELM)的稳定性和稀疏性,在鲁棒ELM的基础上,引入l_(0)范数作为模型的正则项来提高稀疏性,建立了基于l_(0)范数正则项的稀疏鲁棒ELM。首先,通过一个凸差(difference of convex,DC)函...为了进一步提高极限学习机(extreme learning machine,ELM)的稳定性和稀疏性,在鲁棒ELM的基础上,引入l_(0)范数作为模型的正则项来提高稀疏性,建立了基于l_(0)范数正则项的稀疏鲁棒ELM。首先,通过一个凸差(difference of convex,DC)函数逼近l_(0)范数,得到一个DC规划的优化问题;然后,采用DC算法进行求解;最后,在人工数据集和基准数据集上进行实验。实验结果表明:基于l_(0)范数的鲁棒ELM能够同时实现稀疏性和鲁棒性的提升,尤其在稀疏性上表现出较大的优势。展开更多
Bioluminescence tomography(BLT)is an important noninvasive optical molecular imaging modality in preclinical research.To improve the image quality,reconstruction algorithms have to deal with the inherent ill-posedness...Bioluminescence tomography(BLT)is an important noninvasive optical molecular imaging modality in preclinical research.To improve the image quality,reconstruction algorithms have to deal with the inherent ill-posedness of BLT inverse problem.The sparse characteristic of bioluminescent sources in spatial distribution has been widely explored in BLT and many L1-regularized methods have been investigated due to the sparsity-inducing properties of L1 norm.In this paper,we present a reconstruction method based on L_(1/2) regularization to enhance sparsity of BLT solution and solve the nonconvex L_(1/2) norm problem by converting it to a series of weighted L1 homotopy minimization problems with iteratively updated weights.To assess the performance of the proposed reconstruction algorithm,simulations on a heterogeneous mouse model are designed to compare it with three representative sparse reconstruction algorithms,including the weighted interior-point,L1 homotopy,and the Stagewise Orthogonal Matching Pursuit algorithm.Simulation results show that the proposed method yield stable reconstruction results under different noise levels.Quantitative comparison results demonstrate that the proposed algorithm outperforms the competitor algorithms in location accuracy,multiple-source resolving and image quality.展开更多
Tomographic synthetic aperture radar(TomoSAR)imaging exploits the antenna array measurements taken at different elevation aperture to recover the reflectivity function along the elevation direction.In these years,for ...Tomographic synthetic aperture radar(TomoSAR)imaging exploits the antenna array measurements taken at different elevation aperture to recover the reflectivity function along the elevation direction.In these years,for the sparse elevation distribution,compressive sensing(CS)is a developed favorable technique for the high-resolution elevation reconstruction in TomoSAR by solving an L_(1) regularization problem.However,because the elevation distribution in the forested area is nonsparse,if we want to use CS in the recovery,some basis,such as wavelet,should be exploited in the sparse L_(1/2) representation of the elevation reflectivity function.This paper presents a novel wavelet-based L_(2) regularization CS-TomoSAR imaging method of the forested area.In the proposed method,we first construct a wavelet basis,which can sparsely represent the elevation reflectivity function of the forested area,and then reconstruct the elevation distribution by using the L_(1/2) regularization technique.Compared to the wavelet-based L_(1) regularization TomoSAR imaging,the proposed method can improve the elevation recovered quality efficiently.展开更多
文摘大数据时代背景下,随着所获数据数量和维度的不断增加,高维数据的处理成为聚类分析的重点和难点.基于同一类别高维数据通常分布在高维环绕空间的低维子空间这一事实,子空间聚类成为高维数据聚类分析领域的重要方法.稀疏子空间聚类(Sparse Space Clustering,SSC)通过交替方向乘子法(Alternating Direction Method of Multipliers,ADMM)对数据矩阵的稀疏自表达系数进行求解,发现分布于低维子空间并集中的数据的稀疏表示并进行聚类.但是ADMM参数多、收敛速度慢,其效率难以满足对大规模数据库进行聚类分析的要求.针对这一问题提出了基于L_(0)约束的稀疏子空间聚类方法,该方法使用正交匹配追踪(Orthogonal Matching Pursuit,OMP)算法求解L_(0)约束的自表达稀疏重建问题,构建数据集中各数据之间的相关性矩阵,最终对相关性矩阵应用谱聚类方法得到聚类结果.根据OMP算法每次迭代之间的耦合关系对其进行优化,进一步降低了计算复杂度,提高了算法效率.在生成数据和Extended Yale B database人脸数据库的实验结果表明,该算法与SSC相比,在显著减少计算时间的基础上,取得了与SSC相当的聚类准确率.
文摘l_(0)梯度最小化图像平滑算法可在保持边缘的同时滤除纹理和细节,但该算法使用图像梯度判决被平滑成分时会出现包含较小图像梯度(弱边缘)的区域会被平滑,而包含较大图像梯度(强纹理)的区域被保留的现象.为克服此缺陷,提出一种基于图像块l_(0)梯度最小化算法(image-patch based l_(0)gradient minimization algorithm,简称IP-l_(0)算法)的图像平滑算法,通过对输入图像中的图像块而非整幅图像进行平滑,动态改变图像块目标函数中的权重参数,令主要包含强纹理的图像块以较大的力度进行平滑,而主要包含弱边缘的图像块以较小的力度进行平滑,再整合平滑后的图像块得到整个边缘保持平滑图像.对IP-l_(0)算法、原始的l_(0)梯度最小化算法、基于局部拉普拉斯滤波器的算法、基于相对全变差算法、基于树滤波的算法,以及2种基于深度学习的边缘保持算法进行仿真实验,结果表明,使用IP-l_(0)算法滤波后的图像能在保持较弱的边缘的同时平滑强纹理.
文摘为了进一步提高极限学习机(extreme learning machine,ELM)的稳定性和稀疏性,在鲁棒ELM的基础上,引入l_(0)范数作为模型的正则项来提高稀疏性,建立了基于l_(0)范数正则项的稀疏鲁棒ELM。首先,通过一个凸差(difference of convex,DC)函数逼近l_(0)范数,得到一个DC规划的优化问题;然后,采用DC算法进行求解;最后,在人工数据集和基准数据集上进行实验。实验结果表明:基于l_(0)范数的鲁棒ELM能够同时实现稀疏性和鲁棒性的提升,尤其在稀疏性上表现出较大的优势。
基金supported by the National Natural Science Foundation of China(No.61401264,11574192)the Natural Science Research Plan Program in Shaanxi Province of China(No.2015JM6322)the Fundamental Research Funds for the Central Universities(No.GK201603025).
文摘Bioluminescence tomography(BLT)is an important noninvasive optical molecular imaging modality in preclinical research.To improve the image quality,reconstruction algorithms have to deal with the inherent ill-posedness of BLT inverse problem.The sparse characteristic of bioluminescent sources in spatial distribution has been widely explored in BLT and many L1-regularized methods have been investigated due to the sparsity-inducing properties of L1 norm.In this paper,we present a reconstruction method based on L_(1/2) regularization to enhance sparsity of BLT solution and solve the nonconvex L_(1/2) norm problem by converting it to a series of weighted L1 homotopy minimization problems with iteratively updated weights.To assess the performance of the proposed reconstruction algorithm,simulations on a heterogeneous mouse model are designed to compare it with three representative sparse reconstruction algorithms,including the weighted interior-point,L1 homotopy,and the Stagewise Orthogonal Matching Pursuit algorithm.Simulation results show that the proposed method yield stable reconstruction results under different noise levels.Quantitative comparison results demonstrate that the proposed algorithm outperforms the competitor algorithms in location accuracy,multiple-source resolving and image quality.
基金This work was supported by the Fundamental Research Funds for the Central Universities(NE2020004)the National Natural Science Foundation of China(61901213)+3 种基金the Natural Science Foundation of Jiangsu Province(BK20190397)the Aeronautical Science Foundation of China(201920052001)the Young Science and Technology Talent Support Project of Jiangsu Science and Technology Associationthe Foundation of Graduate Innovation Center in Nanjing University of Aeronautics and Astronautics(kfjj20200419).
文摘Tomographic synthetic aperture radar(TomoSAR)imaging exploits the antenna array measurements taken at different elevation aperture to recover the reflectivity function along the elevation direction.In these years,for the sparse elevation distribution,compressive sensing(CS)is a developed favorable technique for the high-resolution elevation reconstruction in TomoSAR by solving an L_(1) regularization problem.However,because the elevation distribution in the forested area is nonsparse,if we want to use CS in the recovery,some basis,such as wavelet,should be exploited in the sparse L_(1/2) representation of the elevation reflectivity function.This paper presents a novel wavelet-based L_(2) regularization CS-TomoSAR imaging method of the forested area.In the proposed method,we first construct a wavelet basis,which can sparsely represent the elevation reflectivity function of the forested area,and then reconstruct the elevation distribution by using the L_(1/2) regularization technique.Compared to the wavelet-based L_(1) regularization TomoSAR imaging,the proposed method can improve the elevation recovered quality efficiently.
