SVD Approach Research Articles

Optimizing Robustness and Invisibility in Digital Image Watermarking: A SVM-Based Multi-Level DWT and SVD Approach

This research introduces a new digital image watermarking approach that utilizes discrete wave transformation (DWT), Support vector machine, and singular value decomposition. The method improves robustness under various assault situations by using the SVM classifier during watermark extraction. Multi-level DWT splits the host picture into sub-bands when embedding, and the coefficients are used as input for SVM. After SVD, the scaling factor embeds the watermark. Comparing the proposed approach to existing research under various attacks, the experimental findings demonstrate that it strikes an equilibrium between robustness and invisibility for watermarks of varying sizes. Support Vector Machine is a contemporary category of machine learning techniques that is extensively employed for the purpose of solving classification problems.

Three Phase Current Transformer Calibration Without External Reference IT Using Synchrophasors: An SVD Approach

Recently a pragmatic data centric method to calibrate substation capacitive voltage transformers (CVT) by exploiting local redundancy in substation CVT PMU data has been proposed. It does not require any external reference CVT. This paper extends the above philosophy to current transformer (CT) calibration problem. Once a line terminals CVTs are calibrated, then estimate of the three phase line currents can be deduced. Currents in a bus, for each of the three phases sum to zero. Hence, the corresponding measurement matrix will be rank deficient. Singular value decomposition's low rank approximation property is used to filter the noise and then a linear least squares problem is solved to calibrate remaining substation CTs. The method then can be intrinsically extended to calibrate other CTs of the network without using other series line impedances. Results are presented with both simulated and field PMU data of 400 kV lines.

An efficient reversible data hiding using SVD over a novel weighted iterative anisotropic total variation based denoised medical images

Computed tomography (CT) advancement and extensive usage have raised the public’s worry regarding the patient’s associated radiation dose. Reducing the radiation dose may lead to more noise and artifacts, which may harm the reputation of radiologists. The instability of low-dose CT reconstruction necessitates better image reconstruction, which increases the diagnostic performance. More modern low-dose CT tests have demonstrated outstanding results. Many times these low-dose denoised medical images with medical related information are also required to transmit over a network. Hence in this article, firstly is a novel denoising method is proposed to improve the quality of low-dose CT images that is based on the total variation method by utilizing whale optimization algorithm (WHA). WHA method is important for getting the best possible weighted function. Reduction of noise occurs by the comparison of a given output to the ground truth, although total variation tends to statistically migrate the data noise distribution from strong to weak. Following denoising, a reversible watermarking approach based on SVD and multi-local extrema (MLE) approaches is provided. Individual results of denoising and watermarking are excellent in terms of visual and performance metrics, according to comparative experimental investigation. Also it was also analyzed that if the watermark is embedded over the denoised CT images then the results of watermarking methods are impressive. So, resultant image offers us the chance to use our visual perception abilities to allow us to cut noise and keep vital and secure information.

Ionospheric Tomography: A Compressed Sensing Technique Based on Dictionary Learning

GNSS (Global Navigation Satellite System) observation insufficiency limits the development of the voxel-based computerized ionospheric tomography (CIT) technique. Electron densities of voxels without observation cannot be accurately estimated by the commonly used algebraic reconstruction techniques. In this study, we proposed a compressed sensing technique (CST) based on dictionary learning for ionospheric tomography. Specifically, the K-SVD (singular value decomposition) algorithm was used for dictionary learning based on training sets that are derived from the NeQuick model, hereafter referred to as the CST_NeQuick algorithm. K-SVD uses the orthogonal matching pursuit (OMP) for sparse coding and the SVD approach for dictionary updating. Both simulations and real experiments demonstrated the feasibility and superiority of the CST algorithm when compared to the widely used multiplicative algebraic reconstruction technique (MART). It was found that the CST_NeQuick algorithm’s tomographic performances were mostly superior to those of the MART algorithm in comparison with the independent slant total electron content (STEC) references. Another CST-based tomographic experiment was performed by using the MART-based solutions for dictionary learning, hereafter referred to as the CST_MART algorithm. It showed that the CST_MART algorithm can reduce the average root mean square (RMS) of the CIT-derived STEC by 37.3 % and 20.2 %, respectively, when compared to the MART and CST_NeQuick algorithms. Besides, the CST_MART algorithm’s electron density profiles also showed more agreement with the electron density profiles that were derived from radio occultation data.

A Hybrid Ontology-based Recommender System Utilizing Data Enrichment and SVD Approaches

A Hybrid Ontology-based Recommender System Utilizing Data Enrichment and SVD Approaches

Extreme Learning Machine (ELM) based Performance Analysis and Epilepsy Identification from EEG Signals

Epilepsy is a chronic brain condition affecting one in 100 patients. EEG (Electroencephalogram) is a signal representing the influence of combining various processes in the brain. This paper aims to identify the likelihood levels of epilepsy from the features extracted from the EEG signals and to compare them with the output of the Singular Value Decomposition and Extreme Learning Machine (ELM). The SVD approach extracts functionality from EEG signals (that is, to minimize dimensionality) while the ELM is used as a classification system. A study of twenty patients is documented in this paper. The criteria such as the Performance Index (PI), Sensitivity (Se), Specificity (Sp), Average Detection (AD), and Good Detection Ratio (GDR) are evaluated using ELM to identify epilepsy. Results show that when SVD is classified with ELM at sigmoid activation function, an average detection of 98.94% and a GDR of 97.83% are obtained.

R-AMMI-LM: Linear-fit Robust-AMMI model to analyze genotype-by environment interactions

Outliers are a common phenomenon when genotypes are evaluated over locations and years under field conditions and such outliers makes studying genotype-environment Interactions difficult. Robust-AMMI models which use a combination of robust fit and robust SVD approaches, denoted as ‘R-AMMI-RLM’ have been proposed to study GEI in presence of such outliers. Instead of ‘R-AMMI-RLM’ we propose a model which uses a combination of linear fit and robust SVD to study GEI in presence of outliers and we denote this model as ‘R-AMMI-LM’. Here we prove that ‘RAMMI-LM’ was superior over ‘R-AMMI-RLM’ as it recorded very low residual sum of squares and low RMSE values. Thus proposed, ‘R-AMMI-LM’ model could explain the GEI more precisely even in presence of outliers.

Calibrating CVTs of a Substation From Local PMU Data: An SVD Approach

We show that if all the line and bus CVTs within a substation are monitored by PMUs, then it is possible to perform CVT calibration locally. This requires 1) discovery of a set of three phase CVTs which can be used as the reference and 2) calculation of ratio correction factors. We show that the best set of three phase CVTs among the redundant sets can be discovered by minimizing the overall measured unbalance among all the three phase CVT combinations. All the CVTs measure the same bus voltage time series. Hence, the corresponding phase measurement matrix will be a small modification of a rank-1 matrix; each column is a scaled replica of the true voltage vector modified by white noise. A microscopic evaluation of the substation PMU data shows that rank-1 approximation by SVD retains 99.99% information. The magnitude correction factors can be obtained by projecting CVT measurement vectors on the first left singular vector. Phase angle correction factors correspond to angles of elements of the first right singular vector. Results with both simulated and field PMU data of 400 kV substations substantiate the claims made.

Application of Hilbert–Schmidt SVD approach to solve linear two-dimensional Fredholm integral equations of the second kind

Meshfree techniques based on infinitely smooth radial kernels have the great potential to provide spectrally accurate function approximations with irregular domain in high dimensions. The maximum accuracy can mostly be found when the RBF shape parameter is small, i.e., when the radial kernel is relatively smooth. However, as the shape parameter goes to zero, the standard RBF interpolant matrix will be very ill-conditioned. The ill-conditioning can be alleviated using alternate bases. One of these alternative bases is the Hilbert–Schmidt SVD basis. The Hilbert–Schmidt SVD approach suggests a stable mechanism for replacing a set of near-flat kernels with scattered centres to a well-conditioned base for exactly the same space. In this work, the Gaussian Hilbert–Schmidt SVD basis functions method is presented to numerically solve the linear two-dimensional Fredholm integral equations of the second kind. The method estimates the solution by the discrete collocation method based on Gaussian Hilbert–Schmidt SVD basis functions constructed on a set of scattered points. The emerged integrals in the scheme are approximately computed by the Gauss–Legendre quadrature rule. This approach reduces the problem under study to a linear system of algebraic equations which can be solved easily via applying an appropriate numerical technique. Also, the convergence of the proposed approach is established. Finally, numerical results are compared with standard RBF method to indicate the accuracy and efficiency of the suggested approach.

Application of least-squares fitting of ellipse and hyperbola for two dimensional data

Application of the least-square method of ellipse and hyperbola for two-dimensional data has been applied to analyze the spatial continuity of coal deposits in the mining field, by using the fitting method introduced by Fitzgibbon, Pilu, and Fisher in 1996. This method uses as a constrain function. Meanwhile, in 1994, Gander, Golub and Strebel have introduced ellipse and hyperbola fitting methods using the singular value decomposition approach. This SVD approach can be generalized into a three-dimensional fitting. In this research we, will discuss about those two fitting methods and apply it to four data content of coal that is in the form of ash, calorific value, sulfur and thickness of seam so as to produce form of ellipse or hyperbola. In addition, we compute the error difference resulting from each method and from that calculation, we conclude that although the errors are not much different, the error of the method introduced by Fitzgibbon et al is smaller than the fitting method that introduced by Golub et al.

Target Localization in Multistatic Passive Radar Using SVD Approach for Eliminating the Nuisance Parameters

In this paper, a weighted least squares (WLS) method is presented to find the location of a single target by the use of bistatic range (BR) measurements in multistatic passive radar. The proposed method is based on the idea of eliminating the nuisance parameter. Then, the method employs several WLS minimizations to obtain the target location estimate. The weighting matrix producing a significant improvement in the performance of the method is derived in two different conditions—one of them results in the maximum likelihood estimator (MLE) and the other one leads to the best linear unbiased estimator (BLUE). The accuracy properties of the method are analyzed, and the Cramer-Rao lower bound (CRLB) for target localization accuracy is also derived. In addition, in a two-dimensional space and in the case of localizing with one transmitter and the minimal number of receivers, the different placements of receivers are analytically determined, in each of which the location solution is not unique. Furthermore, from different aspects, the BR-based localization is compared to the hyperbolic localization derived from range difference (RD) measurements.

Block Based SVD approach for Additive White Gaussian Noise level Estimation in Satellite Images

In this digital era, estimation of noise and de-noising of digital image is one of the chart-topping fields. In real time, noise may happen at any stage starting from capturing to reception. Noise estimation is ever challenging field even till now! Additive White Gaussian Noise is the basic type, when dealing with images on statistical way of studying. So, this paper as a first step proposed an effective technique for the estimation of noise in an image by using Block based SVD domain approach. This paper addresses the following issues 1) Noise estimation by trail method of singular values 2) Corrupting the image by known noise and study the change 3) Divide the whole image into blocks and add the known noise to the image and finally retrieve the effect on original image. It is one of best techniques for estimating AWGN.

Tensor based geology preserving reservoir parameterization with Higher Order Singular Value Decomposition (HOSVD)

Parameter estimation through robust parameterization techniques has been addressed in many works associated with history matching and inverse problems. Reservoir models are in general complex, nonlinear, and large-scale with respect to the large number of states and unknown parameters. Thus, having a practical approach to replace the original set of highly correlated unknown parameters with non-correlated set of lower dimensionality, that captures the most significant features comparing to the original set, is of high importance. Furthermore, de-correlating system's parameters while keeping the geological description intact is critical to control the ill-posedness nature of such problems. We introduce the advantages of a new low dimensional parameterization approach for reservoir characterization applications utilizing multilinear algebra based techniques like higher order singular value decomposition (HOSVD). In tensor based approaches like HOSVD, 2D permeability images are treated as they are, i.e., the data structure is kept as it is, whereas in conventional dimensionality reduction algorithms like SVD data has to be vectorized. Hence, compared to classical methods, higher redundancy reduction with less information loss can be achieved through decreasing present redundancies in all dimensions. In other words, HOSVD approximation results in a better compact data representation with respect to least square sense and geological consistency in comparison with classical algorithms. We examined the performance of the proposed parameterization technique against SVD approach on the SPE10 benchmark reservoir model as well as synthetic channelized permeability maps to demonstrate the capability of the proposed method. Moreover, to acquire statistical consistency, we repeat all experiments for a set of 1000 unknown geological samples and provide comparison using RMSE analysis. Results prove that, for a fixed compression ratio, the performance of the proposed approach outperforms that of conventional methods perceptually and in terms of least square measure.

An SVD and Derivative Kernel Approach to Learning from Geometric Data

Motivated by problems such as molecular energy prediction, we derive an (improper) kernel between geometric inputs, that is able to capture the relevant rotational and translation invariances in geometric data. Since many physical simulations based upon geometric data produce derivatives of the output quantity with respect to the input positions, we derive an approach that incorporates derivative information into our kernel learning. We further show how to exploit the low rank structure of the resulting kernel matrices to speed up learning. Finally, we evaluated the method in the context of molecular energy prediction, showing good performance for modeling previously unseen molecular configurations. Integrating the approach into a Bayesian optimization, we show substantial improvement over the state of the art in molecular energy optimization.

A Preconditioned Hybrid SVD Method for Accurately Computing Singular Triplets of Large Matrices

The computation of a few singular triplets of large, sparse matrices is a challenging task, especially when the smallest magnitude singular values are needed in high accuracy. Most recent efforts try to address this problem through variations of the Lanczos bidiagonalization method, but they are still challenged even for medium matrix sizes due to the difficulty of the problem. We propose a novel SVD approach that can take advantage of preconditioning and of any well designed eigensolver to compute both largest and smallest singular triplets. Accuracy and efficiency is achieved through a hybrid, two-stage meta-method, PHSVDS. In the first stage, PHSVDS solves the normal equations up to the best achievable accuracy. If further accuracy is required, the method switches automatically to an eigenvalue problem with the augmented matrix. Thus it combines the advantages of the two stages, faster convergence and accuracy, respectively. For the augmented matrix, solving the interior eigenvalue is facilitated by a proper use of the good initial guesses from the first stage and an efficient implementation of the refined projection method. We also discuss how to precondition PHSVDS and to cope with some issues that arise. Numerical experiments illustrate the efficiency and robustness of the method.

Maximally Decimated Paraunitary Linear Phase FIR Filter Bank Design via Iterative SVD Approach

A necessary and sufficient condition for M-channel maximally decimated paraunitary FIR filter banks is derived when the lengths of all the filters are the same and they are integer multiples of M. Based on the derived condition, a family of this kind of filter banks is defined. These results are also applied to the corresponding linear phase filter banks. Then, the design of this type of filter banks with the linear phase property is formulated as two orthogonal Procrustes optimization problems. The analytical solution of each optimization problem is derived based on a singular value decomposition approach. Since no numerical optimization computer aided design tool is required for finding the analytical solution of each optimization problem, the computational power required for our proposed method is much lower than that for existing methods. Computer numerical simulation results show that our proposed method can obtain a design instantaneously.

Research on Concept Clustering of Order Task Ontology based on Modified PART

Clustering is the foundation of establishing the semantic model of order task ontology. The process of concept clustering of order manufacturing task was analyzed to solve the clustering problem. A three-index evaluation method of order task document was proposed to construct the corpus. And then SVD approach was employed to reduce the term-document matrix. On the basis of this, a modified PART clustering algorithm was proposed to reestablish similarity check function by integrating of latent semantic information of ontology concept to improve the clustering accuracy. Finally, the task documents in a building materials equipment enterprise were collected as sample data and the experiment was simulated by matlab. By comparison with ART and PART, the modified PART had better performances in recall and accuracy of concept clustering.

Imputing missing values for genetic interaction data

BackgroundEpistatic Miniarray Profiles (EMAP) enable the research of genetic interaction as an important method to construct large-scale genetic interaction networks. However, a high proportion of missing values frequently poses problems in EMAP data analysis since such missing values hinder downstream analysis. While some imputation approaches have been available to EMAP data, we adopted an improved SVD modeling procedure to impute the missing values in EMAP data which has resulted in a higher accuracy rate compared with existing methods. ResultsThe improved SVD imputation method adopts an effective soft-threshold to the SVD approach which has been shown to be the best model to impute genetic interaction data when compared with a number of advanced imputation methods. Imputation methods also improve the clustering results of EMAP datasets. Thus, after applying our imputation method on the EMAP dataset, more meaningful modules, known pathways and protein complexes could be detected. ConclusionWhile the phenomenon of missing data unavoidably complicates EMAP data, our results showed that we could complete the original dataset by the Soft-SVD approach to accurately recover genetic interactions.

Approaches of Buried Object Detection Technology

This paper describes the different art of buried object detection technology and algorithms. This detection of buried object finds application in many areas, importantly in the Landmine detection which is of growing concern due to the danger of buried landmines to people’s lives, economic growth and development. This paper describes and analyzes different technology available. The approaches discussed are Electrical Impedance Tomography, X-ray backscatter, Infrared Systems, Acoustics/seismic systems; Neutron based Method and finally Ground-Penetrating Radar with two commonly available approaches: Least squares and SVD approach. Finally, the paper concludes highlighting the need to improve the way this information is processed and compared.

Face Recognition through Combined SVD and LBP Features

number of global and local methods are available for the representation of face images, still no single approach is found to be suitable in most of the situations. As the information conveyed by these two feature sets, is different hence the techniques that combine the global and local features together are necessary to obtain the optimal results. In this paper, we have developed an approach to combine two feature sets obtained from SVD and LBP approaches. SVD approach is able to efficiently represent the global variations of face images whereas the LBP is one of the most useful descriptors to extract the local variation of face images. In order to analyse the effectiveness of the proposed approach obtained by the fusion of SVD and LBP approaches, various experiments have been carried out on ORL and Yale face databases. The proposed approach has also been compared to some existing techniques and from the detailed experiments it has been observed that the results obtained by the proposed method are far better than these approaches.