Multiple kernel clustering based on local kernel alignment has achieved outstanding clustering performance by applying local kernel alignment on each sample.However,we observe that most of existing works usually assum...Multiple kernel clustering based on local kernel alignment has achieved outstanding clustering performance by applying local kernel alignment on each sample.However,we observe that most of existing works usually assume that each local kernel alignment has the equal contribution to clustering performance,while local kernel alignment on different sample actually has different contribution to clustering performance.Therefore this assumption could have a negative effective on clustering performance.To solve this issue,we design a multiple kernel clustering algorithm based on self-weighted local kernel alignment,which can learn a proper weight to clustering performance for each local kernel alignment.Specifically,we introduce a new optimization variable-weight-to denote the contribution of each local kernel alignment to clustering performance,and then,weight,kernel combination coefficients and cluster membership are alternately optimized under kernel alignment frame.In addition,we develop a three-step alternate iterative optimization algorithm to address the resultant optimization problem.Broad experiments on five benchmark data sets have been put into effect to evaluate the clustering performance of the proposed algorithm.The experimental results distinctly demonstrate that the proposed algorithm outperforms the typical multiple kernel clustering algorithms,which illustrates the effectiveness of the proposed algorithm.展开更多
Longitudinal cracks are common defects of continuous casting slabs and may lead to serious quality accidents. Image capturing and recognition of hot slabs is an effective way for on-line detection of cracks, and recog...Longitudinal cracks are common defects of continuous casting slabs and may lead to serious quality accidents. Image capturing and recognition of hot slabs is an effective way for on-line detection of cracks, and recognition of cracks is essential because the surface of hot slabs is very complicated. In order to detect the surface longitudinal cracks of the slabs, a new feature extraction method based on Curvelet transform and kernel locality preserving projections （KLPP） is proposed. First, sample images are decomposed into three levels by Curvelet transform. Second, Fourier transform is applied to all sub-band images and the Fourier amplitude spectrum of each sub-band is computed to get features with translational invariance. Third, five kinds of statistical features of the Fourier amplitude spectrum are computed and combined in different forms. Then, KLPP is employed for dimensionality reduction of the obtained 62 types of high-dimensional combined features. Finally, a support vector machine （SVM） is used for sample set classification. Experiments with samples from a real production line of continuous casting slabs show that the algorithm is effective to detect longitudinal cracks, and the classification rate is 91.89%.展开更多
Space object recognition plays an important role in spatial exploitation and surveillance, followed by two main problems： lacking of data and drastic changes in viewpoints. In this article, firstly, we build a three-...Space object recognition plays an important role in spatial exploitation and surveillance, followed by two main problems： lacking of data and drastic changes in viewpoints. In this article, firstly, we build a three-dimensional （3D） satellites dataset named BUAA Satellite Image Dataset （BUAA-SID 1.0） to supply data for 3D space object research. Then, based on the dataset, we propose to recognize full-viewpoint 3D space objects based on kernel locality preserving projections （KLPP）. To obtain more accurate and separable description of the objects, firstly, we build feature vectors employing moment invariants, Fourier descriptors, region covariance and histogram of oriented gradients. Then, we map the features into kernel space followed by dimensionality reduction using KLPP to obtain the submanifold of the features. At last, k-nearest neighbor （kNN） is used to accomplish the classification. Experimental results show that the proposed approach is more appropriate for space object recognition mainly considering changes of viewpoints. Encouraging recognition rate could be obtained based on images in BUAA-SID 1.0, and the highest recognition result could achieve 95.87%.展开更多
Based on a representation lemma. Riesz type kernels on the local field K and on the integer ring O in K are coitstructed. Furthermore, we discuss approximation theorems for the Lipschitz class Lip(L ;α) ana the Lp bo...Based on a representation lemma. Riesz type kernels on the local field K and on the integer ring O in K are coitstructed. Furthermore, we discuss approximation theorems for the Lipschitz class Lip(L ;α) ana the Lp boundedness of such operators motivated by the open problem: Does σηfa,s→f for f ∈L1(O) (see M. H. Taible-son [6] and [5])?展开更多
Since data services are penetrating into our daily life rapidly, the mobile network becomes more complicated, and the amount of data transmission is more and more increasing. In this case, the traditional statistical ...Since data services are penetrating into our daily life rapidly, the mobile network becomes more complicated, and the amount of data transmission is more and more increasing. In this case, the traditional statistical methods for anomalous cell detection cannot adapt to the evolution of networks, and data mining becomes the mainstream. In this paper, we propose a novel kernel density-based local outlier factor(KLOF) to assign a degree of being an outlier to each object. Firstly, the notion of KLOF is introduced, which captures exactly the relative degree of isolation. Then, by analyzing its properties, including the tightness of upper and lower bounds, sensitivity of density perturbation, we find that KLOF is much greater than 1 for outliers. Lastly, KLOFis applied on a real-world dataset to detect anomalous cells with abnormal key performance indicators(KPIs) to verify its reliability. The experiment shows that KLOF can find outliers efficiently. It can be a guideline for the operators to perform faster and more efficient trouble shooting.展开更多
Biolubricant was synthesized from Cameroon palm kernel oil (PKO) by double transesterification, producing methyl esters in the first stage which were then transesterified with trimethylolpropane (TMP) to give the PKO ...Biolubricant was synthesized from Cameroon palm kernel oil (PKO) by double transesterification, producing methyl esters in the first stage which were then transesterified with trimethylolpropane (TMP) to give the PKO biolubricant in the presence of a base catalyst obtained from plantain peelings (municipal waste). The yields from both catalysts were significantly similar (48% for the locally produced and 51% for the conventional) showing that the locally produced catalyst could be valorized. The synthesized biolubricant was characterized by measuring its physical and chemical properties. The specific gravity of 1.2, ASTM color of 1.5, cloud point of 0°C, pour point of -9°C, viscosities at 40°C of 509.80 cSt and at 100°C of 30.80 cSt, viscosity index of 120, flash point greater than 210°C and a fire point greater than 220°C were obtained. This synthesized biolubricant was found to be comparable to commercial T-46 petroleum lubricant sample produced industrially from mineral sources. We have therefore used local materials to produce a biolubricant using a cheap base catalyst produced from municipal waste.展开更多
Currently, some fault prognosis technology occasionally has relatively unsatisfied performance especially for in- cipient faults in nonlinear processes duo to their large time delay and complex internal connection. To...Currently, some fault prognosis technology occasionally has relatively unsatisfied performance especially for in- cipient faults in nonlinear processes duo to their large time delay and complex internal connection. To overcome this deficiency, multivariate time delay analysis is incorporated into the high sensitive local kernel principal component analysis. In this approach, mutual information estimation and Bayesian information criterion （BIC） are separately used to acquire the correlation degree and time delay of the process variables. Moreover, in order to achieve prediction, time series prediction by back propagation （BP） network is applied whose input is multivar- iate correlated time series other than the original time series. Then the multivariate time delayed series and future values obtained by time series prediction are combined to construct the input of local kernel principal component analysis （LKPCA） model for incipient fault prognosis. The new method has been exemplified in a sim- ple nonlinear process and the complicated Tennessee Eastman （TE） benchmark process. The results indicate that the new method has suoerioritv in the fault prognosis sensitivity over other traditional fault prognosis methods.展开更多
The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddi...The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddings,such as manifold learning.However,these methods are all based on manual intervention,which have some shortages in stability,and suppressing the disturbance noise.To extract features automatically,a manifold learning method with self-organization mapping is introduced for the first time.Under the non-uniform sample distribution reconstructed by the phase space,the expectation maximization(EM) iteration algorithm is used to divide the local neighborhoods adaptively without manual intervention.After that,the local tangent space alignment(LTSA) algorithm is adopted to compress the high-dimensional phase space into a more truthful low-dimensional representation.Finally,the signal is reconstructed by the kernel regression.Several typical states include the Lorenz system,engine fault with piston pin defect,and bearing fault with outer-race defect are analyzed.Compared with the LTSA and continuous wavelet transform,the results show that the background noise can be fully restrained and the entire periodic repetition of impact components is well separated and identified.A new way to automatically and precisely extract the impulsive components from mechanical signals is proposed.展开更多
In this paper, we try to find numerical solution of y'(x)= p(x)y(x)+g(x)+λ∫ba K(x, t)y(t)dt, y(a)=α. a≤x≤b, a≤t≤b or y'(x)= p(x)y(x)+g(x)+λ∫xa K(x, t)y(t)dt, y(a)=α. a≤x≤b, a≤t≤b by using Local p...In this paper, we try to find numerical solution of y'(x)= p(x)y(x)+g(x)+λ∫ba K(x, t)y(t)dt, y(a)=α. a≤x≤b, a≤t≤b or y'(x)= p(x)y(x)+g(x)+λ∫xa K(x, t)y(t)dt, y(a)=α. a≤x≤b, a≤t≤b by using Local polynomial regression (LPR) method. The numerical solution shows that this method is powerful in solving integro-differential equations. The method will be tested on three model problems in order to demonstrate its usefulness and accuracy.展开更多
【目的】跨视角对象级地理定位(CVOGL)旨在卫星影像上精确定位地面街景或无人机影像所观测目标的地理位置。现有方法多聚焦于图像级匹配,通过对整张影像全局处理实现跨视角关联,缺乏对特定目标的位置编码研究,导致无法将模型的注意力引...【目的】跨视角对象级地理定位(CVOGL)旨在卫星影像上精确定位地面街景或无人机影像所观测目标的地理位置。现有方法多聚焦于图像级匹配,通过对整张影像全局处理实现跨视角关联,缺乏对特定目标的位置编码研究,导致无法将模型的注意力引导到感兴趣目标。并且由于参考图像覆盖范围的变化,查询目标在对应卫星图像中的像素占比极低,精确定位较为困难。【方法】针对以上问题,本文提出了一种基于高斯核函数与异构空间对比损失的跨视角对象级地理定位方法(Cross-View Object-Level Geo-Localization Method with Gaussian Kernel Function and Heterogeneous Spatial Contrastive Loss,GHGeo),用于精确定位感兴趣目标位置。该方法首先通过高斯核函数对查询目标进行精确位置编码,实现了对目标中心点及其分布特征的精细化建模;此外还提出了动态注意力精细化融合模块来动态加权交叉感知全局上下文与局部几何特征的空间相似性,以概率密度预测查询目标在卫星影像中的精确位置;最后通过异构空间对比损失函数来约束其训练过程,缓解跨视角特征差异。【结果】本文在CVOGL数据集进行了实验,实验结果显示:GHGeo在该数据集的“无人机-卫星”任务中,当交并比(IoU)≥25%和≥50%时定位准确率分别达到67.73%和63.00%,相较于基准方法DetGeo分别提升了5.76%和5.34%;在“街景-卫星”定位任务中,对应IoU阈值下的定位准确率分别为48.41%和45.43%的定位准确率,相较于基准方法DetGeo分别提升了2.98%和3.19%。同时与TransGeo,SAFA和VAGeo等方法在CVOGL数据集上进行对比,GHGeo则展现出了更高的定位准确性。【结论】本文方法有效提升了跨视角对象级地理定位方法的精度,为城市规划监测,应急救援调度等应用领域提供关键技术支持和精确位置信息支撑。展开更多
基金This work was supported by the National Key R&D Program of China(No.2018YFB1003203)National Natural Science Foundation of China(Nos.61672528,61773392,61772561)+1 种基金Educational Commission of Hu Nan Province,China(No.14B193)the Key Research&Development Plan of Hunan Province(No.2018NK2012).
文摘Multiple kernel clustering based on local kernel alignment has achieved outstanding clustering performance by applying local kernel alignment on each sample.However,we observe that most of existing works usually assume that each local kernel alignment has the equal contribution to clustering performance,while local kernel alignment on different sample actually has different contribution to clustering performance.Therefore this assumption could have a negative effective on clustering performance.To solve this issue,we design a multiple kernel clustering algorithm based on self-weighted local kernel alignment,which can learn a proper weight to clustering performance for each local kernel alignment.Specifically,we introduce a new optimization variable-weight-to denote the contribution of each local kernel alignment to clustering performance,and then,weight,kernel combination coefficients and cluster membership are alternately optimized under kernel alignment frame.In addition,we develop a three-step alternate iterative optimization algorithm to address the resultant optimization problem.Broad experiments on five benchmark data sets have been put into effect to evaluate the clustering performance of the proposed algorithm.The experimental results distinctly demonstrate that the proposed algorithm outperforms the typical multiple kernel clustering algorithms,which illustrates the effectiveness of the proposed algorithm.
基金Sponsored by Program for New Century Excellent Talents in University of China(NCET-08-0726)Beijing Nova Program of China(2007B027)
文摘Longitudinal cracks are common defects of continuous casting slabs and may lead to serious quality accidents. Image capturing and recognition of hot slabs is an effective way for on-line detection of cracks, and recognition of cracks is essential because the surface of hot slabs is very complicated. In order to detect the surface longitudinal cracks of the slabs, a new feature extraction method based on Curvelet transform and kernel locality preserving projections （KLPP） is proposed. First, sample images are decomposed into three levels by Curvelet transform. Second, Fourier transform is applied to all sub-band images and the Fourier amplitude spectrum of each sub-band is computed to get features with translational invariance. Third, five kinds of statistical features of the Fourier amplitude spectrum are computed and combined in different forms. Then, KLPP is employed for dimensionality reduction of the obtained 62 types of high-dimensional combined features. Finally, a support vector machine （SVM） is used for sample set classification. Experiments with samples from a real production line of continuous casting slabs show that the algorithm is effective to detect longitudinal cracks, and the classification rate is 91.89%.
基金National Natural Science Foundation of China (60776793,60802043)National Basic Research Program of China (2010CB327900)
文摘Space object recognition plays an important role in spatial exploitation and surveillance, followed by two main problems： lacking of data and drastic changes in viewpoints. In this article, firstly, we build a three-dimensional （3D） satellites dataset named BUAA Satellite Image Dataset （BUAA-SID 1.0） to supply data for 3D space object research. Then, based on the dataset, we propose to recognize full-viewpoint 3D space objects based on kernel locality preserving projections （KLPP）. To obtain more accurate and separable description of the objects, firstly, we build feature vectors employing moment invariants, Fourier descriptors, region covariance and histogram of oriented gradients. Then, we map the features into kernel space followed by dimensionality reduction using KLPP to obtain the submanifold of the features. At last, k-nearest neighbor （kNN） is used to accomplish the classification. Experimental results show that the proposed approach is more appropriate for space object recognition mainly considering changes of viewpoints. Encouraging recognition rate could be obtained based on images in BUAA-SID 1.0, and the highest recognition result could achieve 95.87%.
文摘Based on a representation lemma. Riesz type kernels on the local field K and on the integer ring O in K are coitstructed. Furthermore, we discuss approximation theorems for the Lipschitz class Lip(L ;α) ana the Lp boundedness of such operators motivated by the open problem: Does σηfa,s→f for f ∈L1(O) (see M. H. Taible-son [6] and [5])?
基金supported by the National Basic Research Program of China (973 Program: 2013CB329004)
文摘Since data services are penetrating into our daily life rapidly, the mobile network becomes more complicated, and the amount of data transmission is more and more increasing. In this case, the traditional statistical methods for anomalous cell detection cannot adapt to the evolution of networks, and data mining becomes the mainstream. In this paper, we propose a novel kernel density-based local outlier factor(KLOF) to assign a degree of being an outlier to each object. Firstly, the notion of KLOF is introduced, which captures exactly the relative degree of isolation. Then, by analyzing its properties, including the tightness of upper and lower bounds, sensitivity of density perturbation, we find that KLOF is much greater than 1 for outliers. Lastly, KLOFis applied on a real-world dataset to detect anomalous cells with abnormal key performance indicators(KPIs) to verify its reliability. The experiment shows that KLOF can find outliers efficiently. It can be a guideline for the operators to perform faster and more efficient trouble shooting.
文摘Biolubricant was synthesized from Cameroon palm kernel oil (PKO) by double transesterification, producing methyl esters in the first stage which were then transesterified with trimethylolpropane (TMP) to give the PKO biolubricant in the presence of a base catalyst obtained from plantain peelings (municipal waste). The yields from both catalysts were significantly similar (48% for the locally produced and 51% for the conventional) showing that the locally produced catalyst could be valorized. The synthesized biolubricant was characterized by measuring its physical and chemical properties. The specific gravity of 1.2, ASTM color of 1.5, cloud point of 0°C, pour point of -9°C, viscosities at 40°C of 509.80 cSt and at 100°C of 30.80 cSt, viscosity index of 120, flash point greater than 210°C and a fire point greater than 220°C were obtained. This synthesized biolubricant was found to be comparable to commercial T-46 petroleum lubricant sample produced industrially from mineral sources. We have therefore used local materials to produce a biolubricant using a cheap base catalyst produced from municipal waste.
基金Supported by the National Natural Science Foundation of China(61573051,61472021)the Natural Science Foundation of Beijing(4142039)+1 种基金Open Fund of the State Key Laboratory of Software Development Environment(SKLSDE-2015KF-01)Fundamental Research Funds for the Central Universities(PT1613-05)
文摘Currently, some fault prognosis technology occasionally has relatively unsatisfied performance especially for in- cipient faults in nonlinear processes duo to their large time delay and complex internal connection. To overcome this deficiency, multivariate time delay analysis is incorporated into the high sensitive local kernel principal component analysis. In this approach, mutual information estimation and Bayesian information criterion （BIC） are separately used to acquire the correlation degree and time delay of the process variables. Moreover, in order to achieve prediction, time series prediction by back propagation （BP） network is applied whose input is multivar- iate correlated time series other than the original time series. Then the multivariate time delayed series and future values obtained by time series prediction are combined to construct the input of local kernel principal component analysis （LKPCA） model for incipient fault prognosis. The new method has been exemplified in a sim- ple nonlinear process and the complicated Tennessee Eastman （TE） benchmark process. The results indicate that the new method has suoerioritv in the fault prognosis sensitivity over other traditional fault prognosis methods.
基金supported by National Natural Science Foundation of China(Grant No.51075323)
文摘The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddings,such as manifold learning.However,these methods are all based on manual intervention,which have some shortages in stability,and suppressing the disturbance noise.To extract features automatically,a manifold learning method with self-organization mapping is introduced for the first time.Under the non-uniform sample distribution reconstructed by the phase space,the expectation maximization(EM) iteration algorithm is used to divide the local neighborhoods adaptively without manual intervention.After that,the local tangent space alignment(LTSA) algorithm is adopted to compress the high-dimensional phase space into a more truthful low-dimensional representation.Finally,the signal is reconstructed by the kernel regression.Several typical states include the Lorenz system,engine fault with piston pin defect,and bearing fault with outer-race defect are analyzed.Compared with the LTSA and continuous wavelet transform,the results show that the background noise can be fully restrained and the entire periodic repetition of impact components is well separated and identified.A new way to automatically and precisely extract the impulsive components from mechanical signals is proposed.
文摘In this paper, we try to find numerical solution of y'(x)= p(x)y(x)+g(x)+λ∫ba K(x, t)y(t)dt, y(a)=α. a≤x≤b, a≤t≤b or y'(x)= p(x)y(x)+g(x)+λ∫xa K(x, t)y(t)dt, y(a)=α. a≤x≤b, a≤t≤b by using Local polynomial regression (LPR) method. The numerical solution shows that this method is powerful in solving integro-differential equations. The method will be tested on three model problems in order to demonstrate its usefulness and accuracy.
文摘【目的】跨视角对象级地理定位(CVOGL)旨在卫星影像上精确定位地面街景或无人机影像所观测目标的地理位置。现有方法多聚焦于图像级匹配,通过对整张影像全局处理实现跨视角关联,缺乏对特定目标的位置编码研究,导致无法将模型的注意力引导到感兴趣目标。并且由于参考图像覆盖范围的变化,查询目标在对应卫星图像中的像素占比极低,精确定位较为困难。【方法】针对以上问题,本文提出了一种基于高斯核函数与异构空间对比损失的跨视角对象级地理定位方法(Cross-View Object-Level Geo-Localization Method with Gaussian Kernel Function and Heterogeneous Spatial Contrastive Loss,GHGeo),用于精确定位感兴趣目标位置。该方法首先通过高斯核函数对查询目标进行精确位置编码,实现了对目标中心点及其分布特征的精细化建模;此外还提出了动态注意力精细化融合模块来动态加权交叉感知全局上下文与局部几何特征的空间相似性,以概率密度预测查询目标在卫星影像中的精确位置;最后通过异构空间对比损失函数来约束其训练过程,缓解跨视角特征差异。【结果】本文在CVOGL数据集进行了实验,实验结果显示:GHGeo在该数据集的“无人机-卫星”任务中,当交并比(IoU)≥25%和≥50%时定位准确率分别达到67.73%和63.00%,相较于基准方法DetGeo分别提升了5.76%和5.34%;在“街景-卫星”定位任务中,对应IoU阈值下的定位准确率分别为48.41%和45.43%的定位准确率,相较于基准方法DetGeo分别提升了2.98%和3.19%。同时与TransGeo,SAFA和VAGeo等方法在CVOGL数据集上进行对比,GHGeo则展现出了更高的定位准确性。【结论】本文方法有效提升了跨视角对象级地理定位方法的精度,为城市规划监测,应急救援调度等应用领域提供关键技术支持和精确位置信息支撑。
