Wavelet Approximation Coefficients Research Articles

Wavelet shrinkage in nonparametric regression models with positive noise

ABSTRACT Wavelet shrinkage estimators are widely applied to analyse datasets in wavelet domain from several fields of science. They typically act by reducing magnitudes of empirical coefficients in a discrete wavelet transformation to estimate wavelet coefficients. In nonparametric regression problems, most of the shrinkage rules are derived from models composed of an unknown function with additive Gaussian noise. Although Gaussian noise assumption is reasonable in several real data analysis, mainly for large sample sizes, it is not general. Contaminated data with positive noise can occur in practice and nonparametric regression models with positive noise bring challenges in wavelet shrinkage point of view. In this sense, this work proposes bayesian shrinkage rules to estimate wavelet coefficients from a nonparametric regression framework with additive and strictly positive noise under exponential and lognormal distributions. Computational aspects are discussed and simulation studies to analyse the performances of the proposed shrinkage rules and compare them with standard techniques are done. An application for winning times Boston Marathon dataset is also provided.

A Multimodal Neuroimaging Feature Extraction Framework for Biomarker Discovery in Myotonic Dystrophies (P8-4.004)

<h3>Objective:</h3> Myotonic dystrophies (DM) are neuromuscular conditions that cause widespread effects throughout the body. New therapeutics are being developed that attack the RNA mechanism underlying DM. Here, we address the urgent need of identifying DM neuroimaging biomarkers that can be used in trials. <h3>Background:</h3> DTI shows decreased white matter connectivity in DM patients and MRI reveals decreased parietal, frontal, and temporal resting state metabolism. Additionally, DTI studies of DM have shown that the white matter abnormalities are diffuse as opposed to multi-focal. <h3>Design/Methods:</h3> Extracting important features from neuroimaging data is challenging due to the high feature dimension and low sample-size problem. This work utilizes a sparse matrix decomposition method and a manifold regularization approach to identify the most discriminative features and improves spatial and temporal resolutions of neuroimaging data. To this end, neuroimaging data of 46 DM1, 17 DM2 patients and 92 healthy controls were analyzed using a ten-fold-cross-validation technique. In addition to neuroimaging data, additional clinical assessments including Quebec Muscular Impairment Rating Scale and Quebec Daytime Sleepiness Scale were calculated for DM patients. Wavelet approximation and detailed coefficients were calculated to consider both long term and short-term fluctuations caused by the disease process. <h3>Results:</h3> Normalized sub-band energy calculated based on wavelet coefficients can characterize DM1 and DM2. Patients with DM1 have widespread white matter volume reductions that are correlated with distal to proximal progression of the muscular involvement in DM1. A significant difference can be seen between DM1 and healthy controls in terms of executive function, motor speed, processing speed, and memory. <h3>Conclusions:</h3> The proposed method can potentially be used to develop and validate key predictors of DM type, severity, and progression of DM central nervous system abnormalities using brain morphometry. Identified features can also potentially provide multidisciplinary biomarkers that help unravel our understanding of fatigue, sleepiness, and attention. <b>Disclosure:</b> Ms. Kamali has nothing to disclose. Dr. Day has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Affinia Therapeutics. Dr. Day has received personal compensation in the range of $5,000-$9,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Novartis Gene Therapy. Dr. Day has received personal compensation in the range of $5,000-$9,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Biogen. Dr. Day has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Roche/Genentech. Dr. Day has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Sarepta. Dr. Day has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Scholar Rock. Dr. Day has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Avidity. Dr. Day has received personal compensation in the range of $10,000-$49,999 for serving as an officer or member of the Board of Directors for Muscular Dystrophy Association. The institution of Dr. Day has received research support from Audentes. The institution of Dr. Day has received research support from Novartis Gene Therapy. The institution of Dr. Day has received research support from Biogen. The institution of Dr. Day has received research support from Cytokinetics. The institution of Dr. Day has received research support from Roche/Genentech. The institution of Dr. Day has received research support from Sanofi/Genzyme. The institution of Dr. Day has received research support from Sarepta. The institution of Dr. Day has received research support from Scholar Rock. Dr. Day has received intellectual property interests from a discovery or technology relating to health care. Dr. Sampson has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Dyne Therapeutics. Dr. Sampson has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Expansion Therapeutics. Dr. Sampson has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Viking Therapeutics. Dr. Sampson has received research support from Marigold Foundation. Dr. Sampson has a non-compensated relationship as a Scientific Advisory Committee with Myotonic Dystrophy Foundation that is relevant to AAN interests or activities. Dr. Murad has received personal compensation for serving as an employee of UCB. Dr. Murad has received personal compensation for serving as an employee of Skyhawk Therapeutics. Dr. Murad has stock in Forma Therapeutics. Dr. Murad has stock in Skyhawk Therapeutics. Mr. Chaufty has received personal compensation for serving as an employee of UCB. Mr. Chaufty has stock in UCB. Mr. Chaufty has received intellectual property interests from a discovery or technology relating to health care.

Compressive Wavelet Domain Deep CNN for Image Classification Using Genetic Algorithm Based Sensing Mask Learning

Using a novel Genetic Algorithm-based Compressive Learning (GACL), a compressed domain-learning framework is proposed that is implemented on the Haar wavelet approximation coefficient images of the standard kaggle RGB cat dog dataset with every images resized to 256x256x3. The compressive sensing (CS) measurements on the selected dataset is achieved by using a simple reduced pixel scheme by retaining only P% of the pixels of the approximation coefficient images and forcing the remaining pixels to 0 using the Primitive Walsh Hadamard (PWH) binary mask and the masked images are used for further learning. A numerical experiment is conducted to analyze the image classification performance of deep convolution neural network (DCNN) learning on compressive sensing (CS) measurements of wavelet approximation coefficient image of the selected dataset. The unmasked wavelet approximation coefficients images are of size only one fourth of the original image, but they visually resembles the original image. The numerical experiment shows that when learning is done on this unmasked wavelet approximation coefficient images a training accuracy of 97% and validation accuracy of 77% are achieved which is remarkable and as good as using the complete spatial domain image. It is found from numerical experiment that, when PWH masking with P=10 is applied only to the test images the validation accuracy falls up to 58% and this fall is due to the fact that the training is done on unmasked images and tested on masked images. On the other hand in a compressive learning framework the DCNN is trained using masked images and when tested using masked images the validation accuracy rises up to 62% due to the fact both the trained images and test images are masked. Further, it is demonstrated that the best PWH masks can be learned by GACL in which the training accuracy increases up to 89% when vertical cross over is used in GACL and increases up to 96% when diagonal crossover is used in GACL. in this case of GACL using diagonal crossover, the 96% of training accuracy using only (10/4=2.5) 2.5% of the image is very remarkable and stand as proof of concept to implement compressive learning framework which need very less pixel thus very less measurement rate (MR) both in training and testing phase minimizing the bandwidth and storing requirement in applications related to IoT and cloud solutions. The average SSIM ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">S_avg</i> )and average PSNR ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">PSNR_avg</i> )are used as quality measurements, which reduce as P reduces, and it is demonstrated that the average SSIM and average PSNR improves when GACL is used to learn the mask, which is the key reason for the performance improvement.

A Resilient Protection Scheme for Common Shunt Fault and High Impedance Fault in Distribution Lines Using Wavelet Transform

Distribution networks are most prone to high impedance faults (HIFs). These types of faults occur whenever an energized conductor makes a contact with some surfaces that are poor conductors such as sand, concrete, branches of trees, etc. The consequential fault current has an absolute magnitude similar to that of load current therefore, the overcurrent relays usually fail to operate correctly during these faults. In the past few decades, several schemes have surfaced for detecting HIFs. The studies conducted in this article suggest an application of discrete wavelet transform to obtain discriminative features of faulty current signals and the development of a threshold-based scheme to discriminate between faulty and nonfaulty situations. The first-level approximate wavelet coefficients are extracted and used to calculate the standard deviation features using a recursive moving window. With the help of these features, faulted phase is also identified simultaneously apart from fault detection. The efficacy of the scheme is tested on a modified IEEE-13 bus test feeder system incorporated with distributed generators. The algorithm has been compared with the conventional HIF detection schemes and the results proved the efficiency and robustness of the proposed scheme to detect the fault under a variety of fault conditions. The main contributions of the proposed scheme are viz., it can correctly differentiate between HIF/LIF and switching events and works well on both grid-connected or islanding mode. Finally, validation with real field fault events captured by digital fault recorder confirms that the proposed framework can assist utility engineers in detecting the HIF/LIF and switching events securely.

Comparison of ECG Baseline Wander Removal Techniques and Improvement Based on Moving Average of Wavelet Approximation Coefficients

The baseline wander is among the artifacts that corrupt the ECG signal. This noise can affect some signal features, in particular the ST segment, which is an important marker for the diagnosis of ischemia. This paper presents a study on the effectiveness of several methods and techniques for suppressing the baseline wonder (BW) from the ECG signals. As a result, a new technique called moving average of wavelet approximation coefficients (DWT-MAV) is proposed. The techniques concerned are the moving average, the approximation of the baseline by polynomial fitting, the Savitzky-Golay filtering, and the discrete wavelet transform (DWT). The comparison of this techniques is performed using the main criteria for assessing the BW denoising quality criteria such mean square error (MSE), percent root mean square difference (PRD) and correlation coefficient (COR). In this paper, three other criteria of comparison are proposed namely the number of samples of the ECG signal, the baseline frequency variation and the time processing. Two of these new indices are related to possible real time ECG denoising. To improve the quality of BW suppression including the new indices, a new method is proposed. This technique is a combination of the DWT and the moving average methods. This new technique performs the best compromise in terms of MSE, PRD, coefficient correlation and the time processing. The simulations were performed on ECG recording from MIT-BIH database with synthetic and real baselines.

Bayesian damage recognition in document images based on a joint global and local homogeneity model

Physical damages (such as torn-offs and scratches) are commonly seen in historical documents. Recognition of such damages is currently absent in digitization-and-information-extraction (DIE) systems but crucial for automatic document comprehension and exploitation. In this paper we propose a generic damage recognition (DR) method based on a joint global and local modeling of the text homogeneity (TH) pattern exhibited in document images. More specifically, a connected component (CC) based formulation is developed as a global homogeneity measure, where TH is characterized using a probabilistic graph model for a coarse recognition of damaged regions. A multi-resolution analysis (MRA) of TH is further developed for a granular within-CC recognition of damage pixels, where the disparity between damage and text pixels is characterized by exploiting neighborhood transitions. This enables the formulation of a local homogeneity measure, where the neighborhood transition around an individual pixel is modeled using the propagation of the approximation coefficients of a stationary wavelet transform (SWT). The proposed global and local homogeneity measures are integrated as a joint likelihood in a Bayesian model with a Markov random field (MRF) prior, where DR is formulated as a maximum a posterior (MAP) inference which is addressed using Markov Chain Monte Carlo (MCMC) sampling. The resulting algorithm is tested on a set of real-life historical newspaper images containing damages of varying size and shape. The performance of the algorithm is evaluated using both F-measures and the Intersection-over-Union (IoU) metric, where test results demonstrate the promising potential of the proposed method.

A new method of power quality disturbance classification based on deep belief network

In view of the low accuracy of single disturbances under the problem of noise interference, a new method of power quality disturbance classification based on deep belief network was proposed. A smooth wavelet multiscale transform is performed on the power quality disturbance signal, and then the soft threshold function is used to process the estimated wavelet coefficients for reconstructing the original signal. Further, it is proposed to use deep confidence network to classify and recognize the reconstructed single disturbance signal. The simulation results demonstrate that the recognition rate of this method for seven typical single disturbances is high. Even under 20dB noise interference, the classification accuracy rate is as high as 93% or more, which proves that the method has a strong ability to resist noise interference.

Fault Detection and Classification Scheme for Transmission Lines Connecting Windfarm Using Single end Impedance

This paper proposes a new algorithm for the detection and classification of faults in transmission lines in the presence of distributed generation. The proposed algorithm uses wavelet transform-based detail and approximation coefficients of voltage and current signals over one cycle after fault inception at one end of the transmission line. The line impedance calculated from the ratio of the voltage and current approximation coefficients is fed to the fuzzy inference system to classify faults. The proposed algorithm has been tested successfully in a real-time digital simulator for nonlinear high impedance faults in the presence of wind farm, considering the effects of CT saturation. All types of faults with the variation of pre-fault loading condition, line parameters, line length and source impedance can be correctly classified using the proposed algorithm.

Run length encoding based wavelet features for COVID-19 detection in X-rays.

Objectives:Introduced in his paper is a novel approach for the recognition of COVID-19 cases in chest X-rays.Methods:The discrete Wavelet transform (DWT) is employed in the proposed system to obtain highly discriminative features from the input chest X-ray image. The selected features are then classified by a support vector machine (SVM) classifier as either normal or COVID-19 cases. The DWT is well-known for its energy compression power. The proposed system uses the DWT to decompose the chest X-ray image into a group of approximation coefficients that contain a small number of high-energy (high-magnitude) coefficients. The proposed system introduces a novel coefficient selection scheme that employs hard thresholding combined with run-length encoding to extract only high-magnitude Wavelet approximation coefficients. These coefficients are utilized as features symbolizing the chest X-ray input image. After applying zero-padding to unify their lengths, the feature vectors are introduced to a SVM which classifies them as either normal or COVID-19 cases.Results:The proposed system yields promising results in terms of classification accuracy, which justifies further work in this direction.Conclusion:The DWT can produce a few features that are highly discriminative. By reducing the dimensionality of the feature space, the proposed system is able to reduce the number of required training images and diminish the space and time complexities of the system.Advances in knowledge:Exploiting and reshaping the approximation coefficients can produce discriminative features representing the input image.

Electroencephalography Artifact Removal using Optimized Radial Basis Function Neural Networks

Electroencephalography (EEG) is a major clinical tool to diagnose, monitor and manage neurological disorders which is mostly affected by artifacts. Given the importance and the need for an automated method to remove artifacts, in this paper some intelligent automated methods are proposed which are composed of three main parts as extraction of effective input, filtering and filter optimization. Wavelet transform is utilized to extract the effective input, and the wavelet approximation coefficients are used as an effective input signal. In addition, Radial Basis Function Neural Network (RBFNN) has been used for filtering. The appropriate number of RBFs has been selected using extensive simulations, and the optimal value​​ of spread parameter has been achieved by Bees algorithm (BA). Finally, the proposed artifact removal schemes have been evaluated on some real contaminated EEG signals in Mashad Ghaem hospital database. The results show that the proposed artifact removal schemes are able to effectively remove artifacts from EEG signals with little underlying brain signal distortion.

Smart Classification Paradigms for Protein Samples from 1-D Electrophoresis Gel

Electrophoresis allows us to identify the types of proteins present in food, DNA, tissues and more. With the help of the molecular marker their weight is known, these markers are applied within the one-dimensional gel, and their protein value is known by means of marks. In this research, the molecular marker is obtained and the wavelet transform (WT) is obtained, generating approximation coefficients, which were taken to determine a molecular weight using three classification paradigms. The first paradigm is an approach in content-based image retrieval (CBIR) which makes a detection of the molecular weight in electrophoresis samples. The second approach is a neural network, thus two models are employed: self-organization maps (SOM) and back propagation in a supervised and unsupervised way, respectively. The third approach is based in a J48 decision tree. A comparison is made between the three paradigms for molecular weight computation. Neural networks obtained an improvement in the precision compared versus the CBIR-WT. Five parametric statistics were generated from the wavelet approximation coefficients. The CBIR-WT, SOM, back propagation and J48 were decisive for the classification and calculation of the molecular weight of each protein stain in the one-dimensional electrophoresis gel.

Fluid and rock heterogeneity assessment of gas condensate reservoirs by wavelet transform of pressure-transient responses

The response of gas condensate reservoirs is directly influenced by heterogeneity. In such reservoirs, condensate is created around wellbore areas, when the reservoir pressure falls below the dew point pressure. Consequently, the distribution of this condensate bank makes fluid flow in the reservoir even more complicated by changing rock-fluid properties (e.g., relative permeability). This alteration can be assumed as a new heterogeneity, called fluid heterogeneity. The separation of fluid heterogeneity from rock heterogeneity is a challenging task. The main idea of this study is to investigate the transient pressure responses of a gas condensate reservoir to separate rock and fluid heterogeneities. Different homogeneous and heterogeneous reservoir models of a reservoir were constructed by the geostatistical approach. A commercial reservoir simulator was used to simulate the behavior of different drawdown and buildup scenarios. The fluid is a lean gas condensate selected from one of the Middle East formations. The wavelet transform (WT) approach was implemented to characterize the behavior of condensate banks for both homogeneous and heterogeneous models. Analyzing the wavelet approximate and detail coefficients of the pressure-transient responses enabled us to distinguish the rock and fluid heterogeneities. The results showed that the wavelet detail coefficient could be a good indicator of reservoir heterogeneity. Moreover, the WT could be successfully used to distinguish different regions of a condensate bank inside the reservoir under study. The implemented approach showed promising results in analyzing the pressure-transient responses of gas condensate reservoirs.

MRI radiomics for early prediction of response to vaccine therapy in a transgenic mouse model of pancreatic ductal adenocarcinoma

BackgroundThere is a lack of well-established clinical tools for predicting dendritic cell (DC) vaccination response of pancreatic ductal adenocarcinoma (PDAC). DC vaccine treatment efficiency was demonstrated using histological analysis in pre-clinical studies; however, its usage was limited due to invasiveness. In this study, we aimed to investigate the potential of MRI texture features for detection of early immunotherapeutic response as well as overall survival (OS) of PDAC subjects following dendritic cell (DC) vaccine treatment in LSL-KrasG12D;LSL-Trp53R172H;Pdx-1-Cre (KPC) transgenic mouse model of pancreatic ductal adenocarcinoma (PDAC).Materials and methodsKPC mice were treated with DC vaccines, and tumor growth was dynamically monitored. A total of a hundred and fifty-two image features of T2-weighted MRI images were analyzed using a kernel-based support vector machine model to detect treatment effects following the first and third weeks of the treatment. Moreover, univariate analysis was performed to describe the association between MRI texture and survival of KPC mice as well as histological tumor biomarkers.ResultsOS for mice in the treatment group was 54.8 ± 22.54 days while the control group had 35.39 ± 17.17 days. A subset of three MRI features distinguished treatment effects starting from the first week with increasing accuracy throughout the treatment (75% to 94%). Besides, we observed that short-run emphasis of approximate wavelet coefficients had a positive correlation with the survival of the KPC mice (r = 0.78, p < 0.001). Additionally, tissue-specific MRI texture features showed positive association with fibrosis percentage (r = 0.84, p < 0.002), CK19 positive percentage (r = − 0.97, p < 0.001), and Ki67 positive cells (r = 0.81, p < 0.02) as histological disease biomarkers.ConclusionOur results demonstrate that MRI texture features can be used as imaging biomarkers for early detection of therapeutic response following DC vaccination in the KPC mouse model of PDAC. Besides, MRI texture can be utilized to characterize tumor microenvironment reflected with histology analysis.

Video Steganography using IWT, DWT, LBP Methods and its Comparative Research

Steganography is a method used for inserting the secret information into an another medium such as text, images, audio signals or video signals, without revealing its existence in the medium. In video steganography, a video file will be used as a cover medium and secret message embedded inside the cover medium. Video are picture perfect for embedding the secret message because of its large embedding capacity. We propose three secure steganography algorithms that embed a bit stream of the secret message into the approximation coefficients of the integer wavelet transform(IWT), DWT and using LBP method to form stego video. The geometric alteration between the cover video and the stego video is measured by using the Mean Square Error (MSE) and the PSNR. The new results show that, the proposed algorithms can hide the secret message with a great payload capacity with a high level of security and a greater invisibility.

Online Fault Detection and Isolation in Advanced Heavy Water Reactor Using Multiscale Principal Component Analysis

Signals from sensors placed at different locations are used for control and protection purposes in a nuclear reactor. It also requires in-service fault detection and isolation (FDI) for its safe operation. The sensor signals are generally a superimposition of low-frequency components representative of true values of the variables being monitored, occasional high-frequency periodical oscillations due to disturbances and faults, and sensor faults. Techniques like principal component analysis (PCA) can be used for FDI; however, it would be more meaningful if the FDI technique can also help for predictive maintenance of reactor internals through vibration spectra. To address these issues, a multiscale PCA, integrating wavelet transform with PCA and aiming to reduce the modeling cost by using only a fewer scales that contribute to the monitoring, has been proposed in this paper for online FDI of an advanced heavy water reactor (AHWR). A new mathematical formulation of the generalized likelihood ratio test for its use with wavelet approximation coefficients has also been proposed for better sensor-FDI outcomes. The proposed approach detects and isolates sensor faults and process faults using the signals from neutron detectors. Efficacy of the proposed technique is established on the simulated ex-core ion chamber data of AHWR considering different scenarios that involve localized frequency contents representative of process faults, slowly developing (incipient) sensor faults, and the simultaneous presence of two or more of these scenarios. Simulation results validate the effectiveness of the proposed scheme for online FDI in the reactor.

A Fast Fault Identification in a Grid-Connected Photovoltaic System Using Wavelet Multi-Resolution Singular Spectrum Entropy and Support Vector Machine

A new protection scheme based on applying a combination of wavelet multi-resolution singular spectrum entropy and support vector machine is proposed to identify different types of grid faults in a three-phase grid-tied photovoltaic system. In this technique, discrete wavelet transform with multi-resolution singular spectrum entropy is utilized to extract the unique features of three-phase voltage signals at the point of common coupling. The three-phase voltage signals are decomposed to provide detail and approximation coefficients of wavelet transform. Then, various features between different types of grid faults can be extracted by a combination of multi resolution analysis and spectrum analysis with entropy as the output. The constructed features vector is utilized as input data of a support vector machine classifier to identify and classify various types of faults. The results illustrate that the proposed intelligent technique not only recognizes different types of grid faults correctly, but also performs quickly in identifying grid faults in a grid-connected photovoltaic system. Apart from this, a graphical investigation is executed to observe the effects of different types of grid faults in photovoltaic (PV) operation which highlight the necessity of intelligent protection methods to protect PV systems.

Multifocus image fusion based on coefficient significance of redundant discrete wavelet transform

PurposeThis study aims to obtain the more precise decision map to fuse the source images by Coefficient significance method. In the area of multifocus image fusion, the better decision map is very important the fusion results. In the processing of distinguishing the well-focus part with blur part in an image, the edge between the parts is more difficult to be processed. Coefficient significance is very effective in generating the better decision map to fuse the multifocus images.Design/methodology/approachThe energy of Laplacian is used in the approximation coefficients of redundant discrete wavelet transform. On the other side, the coefficient significance based on statistic property of covariance is proposed to merge the detail coefficient.FindingsDue to the shift-variance of the redundant discrete wavelet and the effectiveness of fusion rule, the presented fusion method is superior to the region energy in harmonic cosine wavelet domain, pixel significance with the cross bilateral filter and multiscale geometry analysis method of Ripplet transform.Originality/valueIn redundant discrete wavelet domain, the coefficient significance based on statistic property of covariance is proposed to merge the detail coefficient of source images.

Intelligent schemes for fault classification in mutually coupled series-compensated parallel transmission lines

The protection of mutually coupled series capacitor-compensated (SCC) parallel transmission lines is a more complicated task than uncompensated lines due to the effect of mutual coupling, inter-circuit faults, and non-linearity of effective impedance of SCC line. A method that can overcome these issues and still work efficiently is a supervised learning-based method which is an adaptive technique. Hence, in this work, various supervised learning-based intelligent schemes like artificial neural network (ANN), support vector machines (SVM), and decision tree (DT) are employed to find a suitable method for the protection of series capacitor-compensated lines. Discrete wavelet transform has been used to process the three-phase current signals of the parallel lines measured at one terminal only. A moving window of 20 samples is selected, and approximate wavelet coefficient is calculated up to level 1 using DB-4 mother wavelet. The resultant is then given as input to the intelligent schemes (ANN, SVM, and DT). The proposed intelligent schemes have been tested with variety of fault conditions such as inter-circuit faults, cross-country faults, transforming faults, single-circuit operation, and high resistance faults. A large number of fault simulation studies corroborate that DT-based fault classification method is better than ANN and SVM. The accuracy of faulty phase and ground identification scheme is 100% for all the tested fault cases. Hence, the proposed supervised learning-based intelligent method can be implemented in real power system network effectively.

Integrated condition monitoring scheme for bearing fault diagnosis of a wind turbine gearbox

Rolling element bearing faults of a laboratory scale wind turbine gearbox operating under nonstationary loads have been diagnosed using condition monitoring (CM) techniques such as vibration analysis, acoustic analysis, and lubrication oil analysis. Two local bearing faults, namely, bearing inner race fault and bearing outer fault are seeded in the gearbox. The raw data from these techniques are decomposed and wavelet approximation coefficients of level four (a4) are extracted using discrete wavelet transform (DWT). A plethora of statistical features is computed from the wavelet approximation coefficients and the most significant features are being identified by implementing the decision tree algorithm. The classification efficiencies of each of these CM techniques are compared by using the support-vector machine algorithm. Furthermore, an integrated CM scheme is developed by combining the individual CM techniques and the fault diagnosing ability of the integrated CM scheme is compared with the individual CM techniques. A principal component analysis-based approach is used as a feature classification algorithm and an input feature matrix is formed by combining the significant features extracted from vibration, acoustic, and lubrication oil analysis. It has been observed that the integrated CM scheme has provided better classification interpretations than the single CM techniques and it can be extended for real time fault diagnosis of a wind turbine gearbox.

Role of approximation wavelet coefficients in blind image deconvolution

The role of approximation coefficients (ACs) in blind image deconvolution (BID) is examined by restoring each sub-band image individually in the wavelet domain, and excellent performance is achieved by an average strategy for the estimation of AC. The most important observation is that accurate deconvolution of AC is not necessary and some form of hybrid can reach a balance between sharpness and contrast. This finding leads to reappraising the common treatment of wavelet coefficients in deconvolution, which takes the coefficients in every sub-band as a unity. It is argued that the AC of the blurred image can be regarded as a good approximation of the ‘true’ image with lower resolution. Wavelet transform of one level is preferred in the implementation and the difference between the method and common multiscale approach is stressed. With specific treatment of AC, the proposed method can be combined with many other BID algorithms to improve deconvolution performance.