Multi-resolution Singular Value Decomposition Research Articles

Research on bearing fault diagnosis based on novel MRSVD-CWT and improved CNN-LSTM

As a critical component in mechanical equipment, rolling bearings play a vital role in industrial production. Effective bearing fault diagnosis provides a more reliable guarantee for the safe operation of the industrial output. Traditional data-driven bearing fault diagnosis methods often have problems such as insufficient fault feature extraction and poor model generalization capabilities, resulting in reduced diagnostic accuracy. To solve these problems and significantly improve the diagnosis accuracy, this paper proposes a novel fault diagnosis method based on multi-resolution singular value decomposition (MRSVD), continuous wavelet transform (CWT), improved convolutional neural network (CNN) enhanced by convolutional block attention module, and long short-term memory (LSTM). Through MRSVD, the vibration signal is decomposed layer by layer into multiple denoised signals, thus signal noise can be eliminated to the greatest extent to gain the optimal denoised signals; then through CWT, the optimal denoised signals are converted into two-dimensional time-frequency images so that the local and global characteristic information can be fully captured. Finally, through improved CNN-LSTM, feature extraction is greatly enhanced, resulting in high accuracy of fault diagnosis. Lots of experiments are organized to test the performance, and the experimental results show that the proposed method on various datasets has better diagnosis accuracy and generalization ability under different working conditions than other methods.

Design of fault location algorithm based on online distributed travelling wave for HV power cable.

To tackle the challenge of localization failure due to traveling wave dispersion during a mid-line fault in a long high-voltage cable, this study conducts an in-depth analysis of the refraction characteristics of the cumulative three-phase sheath currents at cross commutation points and direct grounding points, facilitated by detailed theoretical derivation. We introduce the novel concept of a 'first traveling wave attenuation index,' which quantifies the peak of the initial faulted traveling wave. We strategically place distributed traveling wave detection devices solely in the first direct grounding box of each cross-interconnected main section, effectively segmenting the line into five distinct zones. These zones are identified based on the unique transient polarities exhibited by the first traveling wave of fault current at each impedance mismatch point along the cable. To overcome the issue of weak traveling wave signals collected at the first end measurement point, we propose an innovative peak detection method for the first wave. This method harnesses the power of empirical wavelet transform (EWT) and multi-resolution singular value decomposition (MRSVD), providing a significant boost in ranging accuracy compared to traditional wavelet methods. Simulation results validate the efficacy of our proposed fault location method, which accurately pinpoints the fault section by contrasting the polarity of the first wave peak of the sheath current detected at each measurement point. Notably, our method ensures safety and convenience as the equipment does not require direct contact with high voltage.

A Cable Defect Localization Method Based on Multiresolution Singular Value Decomposition and Stochastic Subspace Identification

A Cable Defect Localization Method Based on Multiresolution Singular Value Decomposition and Stochastic Subspace Identification

Information fusion for fully automated segmentation of head and neck tumors from PET and CT images.

PET/CT images combining anatomic and metabolic data provide complementary information that can improve clinical task performance. PET image segmentation algorithms exploiting the multi-modal information available are still lacking. Our study aimed to assess the performance of PET and CT image fusion for gross tumor volume (GTV) segmentations of head and neck cancers (HNCs) utilizing conventional, deep learning (DL), and output-level voting-based fusions. The current study is based on a total of 328 histologically confirmed HNCs from six different centers. The images were automatically cropped to a 200 × 200 head and neck region box, and CT and PET images were normalized for further processing. Eighteen conventional image-level fusions were implemented. In addition, a modified U2-Net architecture as DL fusion model baseline was used. Three different input, layer, and decision-level information fusions were used. Simultaneous truth and performance level estimation (STAPLE) and majority voting to merge different segmentation outputs (from PET and image-level and network-level fusions), that is, output-level information fusion (voting-based fusions) were employed. Different networks were trained in a 2D manner with a batch size of 64. Twenty percent of the dataset with stratification concerning the centers (20% in each center) were used for final result reporting. Different standard segmentation metrics and conventional PET metrics, such as SUV, were calculated. In single modalities, PET had a reasonable performance with a Dice score of 0.77±0.09, while CT did not perform acceptably and reached a Dice score of only 0.38±0.22. Conventional fusion algorithms obtained a Dice score range of [0.76-0.81] with guided-filter-based context enhancement (GFCE) at the low-end, and anisotropic diffusion and Karhunen-Loeve transform fusion (ADF), multi-resolution singular value decomposition (MSVD), and multi-level image decomposition based on latent low-rank representation (MDLatLRR) at the high-end. All DL fusion models achieved Dice scores of 0.80. Output-level voting-based models outperformed all other models, achieving superior results with a Dice score of 0.84 for Majority_ImgFus, Majority_All, and Majority_Fast. A mean error of almost zero was achieved for all fusions using SUVpeak , SUVmean and SUVmedian . PET/CT information fusion adds significant value to segmentation tasks, considerably outperforming PET-only and CT-only methods. In addition, both conventional image-level and DL fusions achieve competitive results. Meanwhile, output-level voting-based fusion using majority voting of several algorithms results in statistically significant improvements in the segmentation of HNC.

A robust and secured fusion based hybrid medical image watermarking approach using RDWT-DWT-MSVD with Hyperchaotic system-Fibonacci Q Matrix encryption.

Digital image watermarking, the process of marking a host image with a watermark, is generally used to authenticate the data. In the medical field, it is of utmost importance to verify the authenticity of the data using Medical Image Watermarking (MIW), especially in e-healthcare applications. Recently, MIW with image fusion, the merging of multimodal images to improve image quality, is being widely utilized to make diagnosis more accessible and precise with the verified data. This paper offers a durable and secure fusion-based hybrid MIW approach. The method initially used Fast Filtering (FF) to merge two medical images from different modalities to form the cover image. A first-level Redundant Discrete Wavelet Transform (RDWT) is employed on this host image to locate the component with the highest entropy. Then a single-level Discrete Wavelet Transform (DWT) is applied to it. It performed a Multi-resolution Singular Value Decomposition (MSVD) on the wavelet decomposed component and the embedding watermark. Finally, a Hyperchaotic System-Fibonacci Q Matrix (HFQM) encryption system was utilized, which increases the watermarked image's security. Here, using various medical images, the performance of the proposed technique is evaluated. Without any attacks, the approach achieved a maximum Peak Signal to Noise Ratio (PSNR) of 90.31 dB and a Structural Similarity Index Matrix (SSIM) of value 1. Various watermarking assaults were employed to test the proposed method's resilience. The suggested technique achieved a perfect value of 1 for the Normalised Correlation (NC) for almost all attacks with acceptable imperceptibility, which substantially improves over current procedures. The suggested technique's average embedding and extraction times are 0.3958 and 0.4721 seconds, respectively, which are pretty short compared to existing approaches.

ViMDH: Visible-Imperceptible Medical Data Hiding for Internet of Medical Things

Over the recent years, volume of medical images and related digital records, called electronic medical records, generated, shared, and stored by different intelligent devices, sensors, and Internet of medical things networks, to name a few, has drastically increased. Such records are shared by cloud providers for storage and further processing. However, an increasingly serious concern is the illegal copying, modification, and forgery of medical records. This article presents a visible and imperceptible medical data hiding technique, namely ViMDH, which can prevent to intellectual property theft of medical records. The carrier image is visibly marked with logo mark, which is suitable for owner identification and avoid illegal duplication, and then an imperceptible data hiding based on nonsubsampled shearlet transform (NSST), redundant discrete wavelet transform (RDWT), and multiresolution singular value decomposition is introduced. Finally, key-based encryption scheme designed by RDWT-RSVD ensure the security of the watermarking system. Under the experimental evaluation, our ViMDH is not only visible and imperceptible, but also has a satisfactory advantage in robustness and security compared with the traditional watermarking schemes.

Fault Diagnosis and Verification of Rolling Bearings Based on Optimal Resonance Demodulation with Deconvolution MRSVD

As one of the important aviation equipment components, rolling bearings will bring hidden dangers to the safety of aviation equipment once they fail. In aviation equipment, the sensor is far away from the bearing, and the collected signal of the rolling bearing is relatively weak. In this paper, the deconvolution algorithm is used to compensate the transfer function, and the optimal resonance demodulation algorithm based on the improved multi-resolution singular value decomposition (MRSVD) is constructed to determine the resonance frequency band, and then the bearing fault state is determined by the normalized squared envelope spectrum. The analysis results of the rolling bearing seed fault experiment show that this algorithm can adaptively extract the periodic fault pulse components under strong noise; compared with the normal spectral analysis and fast spectral kurtosis methods, this algorithm can completely extract the bearing fault characteristic frequency components, and effectively improve the bearing fault diagnosis accuracy.

Research on a Denoising Method of Vibration Signals Based on IMRSVD and Effective Component Selection

This paper proposes a denoising method of vibration signal based on improved multiresolution singular value decomposition (IMRSVD) and effective component selection. A new construction method of trajectory matrix is used, which can enhance the oscillating component of the original signal. Next, based on the improved trajectory matrix, singular value decomposition (SVD), which plays the role of pre-decomposition, is used to obtain multiple one-dimensional components, and the further decomposition of that is achieved by multiresolution singular value decomposition (MRSVD). Finally, the effective components selection of a series of decomposed signal components is achieved based on the proposed feature evaluation index (FEI). The denoising experiments are carried out using the simulation signal and the vibration signal of planetary gear, respectively. The experimental results show that the proposed method performs better than the traditional SVD denoising method, and the weak fault feature in the vibration signal can be extracted successfully. In addition, the comparison between periodic modulation intensity (PMI) and FEI displays that the proposed method has better robustness and accuracy than the interference components with similar frequency. Thus, the proposed method is an effective weak fault feature extraction and denoising tool of vibration signals for fault diagnosis.

Geophysical data fusion of ground-penetrating radar and magnetic datasets using 2D wavelet transform and singular value decomposition

This work addresses the problem of the lack of perceptibility that geophysical data may have. Data fusion allows us to combine datasets, providing an improved and more informative source of information about structures buried in the ground. After testing different approaches, a strategy was developed using ground-penetrating radar and magnetic datasets collected over the same area. Data collected at the Roman Villa of Pisões (Beja, Portugal), which is a place of easy application of geophysical methods, were used to test the method, but with problems caused by the properties of the soil. The approach was based on processing operations that allow the fusion of images obtained by different equipment widely used in medical imaging for tumor detection and image processing. The goal is to create an improved image with data fusion that has higher quality than the input images, allowing a better understanding of the object of the study. The approach is composed of two stages: pre-processing and data fusion. Pre-processing is applied to enhance the input data. It consists of removing background noise through singular value decomposition applied in the spectral domain. Then the calculation of the data entropy will highlight the differences corresponding to the spatial alignments compatible with buried structures. Then, both entropy maps of the two datasets are fused in the second processing step to produce the final image. This step involves applying the 2D wavelet transform to each entropy map, decomposing them into sub-bands. Algorithms to calculate multiresolution singular value decomposition and the image gradient are applied to the sub-bands. The processed sub-band pairs are then fused using specific fusion rules. The fused image is obtained by applying the inverse of the wavelet transform. Data fusion with the proposed approach allows us to obtain a detailed image that is sharper and of better quality than the input datasets. The increase in sharpness and quality can be quantified through the sharpness index and the BRISQUE quality index in several steps of the processing. The obtained values confirm the graphical results. Images produced by the proposed data fusion approach suggest that the perceptibility has increased, allowing us to provide conclusions about the existence of buried structures.

Multiple Degradation Skilled Network for Infrared and Visible Image Fusion Based on Multi-Resolution SVD Updation

Existing infrared (IR)-visible (VIS) image fusion algorithms demand source images with the same resolution levels. However, IR images are always available with poor resolution due to hardware limitations and environmental conditions. In this correspondence, we develop a novel image fusion model that brings resolution consistency between IR-VIS source images and generates an accurate high-resolution fused image. We train a single deep convolutional neural network model by considering true degradations in real time and reconstruct IR images. The trained multiple degradation skilled network (MDSNet) increases the prominence of objects in fused images from the IR source image. In addition, we adopt multi-resolution singular value decomposition (MRSVD) to capture maximum information from source images and update IR image coefficients with that of VIS images at the finest level. This ensures uniform contrast along with clear textural information in our results. Experiments demonstrate the efficiency of the proposed method over nine state-of-the-art methods using five image quality assessment metrics.

Method improving low Signal-to-noise ratio of velocity test signals for Laser-induced shock waves

When laser-induced shock (LIS) wave characteristics testing is carried out using laser interference methods, the electrical noise, vibration noise, jitter or offset of the diffuse reflective surface during high-speed motion and other factors can lead to superposition or offset of the noise signal frequency with the low-frequency phase of the useful signal. As a result, problems such as phase shift of the useful signal and difficulty in extracting the initial position of the wavelet ridge may appear, and the initial position cannot be zeroed. By introducing a measurement factor through multi-resolution singular value decomposition and adaptively determining the optimal number of decomposition layers, zero-phase offset filtering is achieved and the signal-to-noise ratio (S/N) of the data is improved. According to the waveform decay characteristics of laser-induced shock wave velocity signals, the Shannon wavelet entropy method is proposed to accurately optimise the Morlet wavelet basis parameters to enhance the aggregation of wavelet scale spectrum time–frequency distribution. Simulation and test results show that: the proposed method overcomes the background noise interference and enables the initial velocity to be zeroed; the relative singular value points of the pulse are accurately symmetrically distributed and no phase shift occurs; the denoising performance is better than that of the wavelet method, the root mean square error is only 0.046, and the repeatability is strong; the error of the elastic wave velocity of the laser-induced shock wave at different power densities is within 3.2%; the relative velocity resolution remains constant at low/high frequencies, effectively avoiding false peaks and misclassification of the velocity profile.

A low-frequency fault detection method for low-speed planetary gearbox based on acoustic signals

In some practical application of construction machinery and equipment, the planetary gearbox in the walking parts is wrapped with tires, and its speed is very low. When this planetary gearbox fails, the fault frequency is very low (less than 1 Hz) due to its complex transmission relationship. Because the planetary gearbox has the characteristics of low speed, low fault frequency and surrounded by tires, it is difficult to detect its fault condition in practical engineering. This work proposes a new low-frequency fault detection method. First, the pre-processing is used for the measured acoustic signals. Second, the pre-processed acoustic signals are decomposed into components by the multi-resolution singular value decomposition. Third, the autoregressive model is constructed to select fault-related component. Finally, the output component is further analyzed by envelope spectrum analysis. The results indicate that the proposed method can accurately detect the characteristic information of low frequency faults of planetary gearbox in practical engineering.

Real-time identification of performance degradation stages of rolling element bearings by RVCFI

The rolling element bearing is a critical component, and its fault results in most of the faults in rotary machines. The performance condition monitoring of bearings can improve the safety of equipment and can provide a reasonable maintenance plan at the optimal time. To identify the initial fault time (IFT) and performance degradation stages of a bearing based on the sampling signal from the start to the current time in real time, the monitoring indicator of the initial fault (MIIF) and the monitoring indicator of the degradation stages (MIDS) are constructed. Firstly, multiple features with stable and high robustness are calculated based on the envelope spectrum of the vibration signal. And these features are standardized and integrated by the weighted-sum of multiple standardized features into a fused indicator. Then, the online variation coefficient of the fused indicator and its rate are calculated. Finally, the MIIF and MIDS can be obtained using the variation coefficient and cumulative sum based on the rate of the variation coefficient of the fused indicator, respectively. Meanwhile, the adaptivity and versatility of the multi-resolution singular value decomposition algorithm are also improved. The proposed methods are verified using two public tested data packets. It is shown from the results that the methods are able to efficiently identify the IFT and performance degradation stages of a bearing in a timely manner.

Image-based decision making for reliable and proper diagnosing in NIFTI format using watermarking

Nowadays, advancement in Magnetic Resonance Imaging (MRI) and Computed Tomography Scan (CT-Scan) technologies have defined modern neuroimaging and drastically change the diagnosing of disease in the world healthcare system. These imaging technologies generate NIFTI (Neuroimaging Informatics Technology Initiative) images. Due to COVID-19 last several months CT-Scan has been performed on millions of the CORONA patients, so billions of the NIFTI images have been produced and communicate over the internet for the diagnosing purpose to detect the coronavirus. The communication of these medical images over the internet yielding the major problem of integrity, copyright protection, and other ethical issues for the world health care system. Another critical problem is that; is doctor diagnose the impeccable medical image of the patient because a large amount of COVID-19 patient’s data exists. For proper diagnosing it is also necessary to identify impeccable medical image. Therefore, to address these problems a secure and robust watermarking scheme is needed for these images. Various watermarking schemes have been developed for bmp, .jpg, .png, DICOM, and other image formats but the noticeable contribution is not reported for the NIFTI images. In this paper a robust and hybrid watermarking scheme for NIFTI images based on Lifting Wavelet Transform (LWT), MSVD (Multiresolution Singular Value Decomposition) and QR factorization. The combination of LWT, QR, and MSVD helps in retaining the sensitivity of the NIFTI image and improve the robustness of the watermarking scheme. In this scheme, multiple watermarks are inserted across the first slice of the NIFTI image. The proposed watermarking scheme is sustained against various noise attacks and performance is measured in terms of PSNR, SNR, SSIM, Quality of image, and Normalized correlation. Quality of the image is much significant that lie between .99994 to .99998 and SSIM reported from .94 to .99. Whereas the PSNR of the proposed scheme lies between 56.76 to 57.28 db and NC values lie between .9993 to .9998. which shows that the results are better than the existing schemes where PSNR is lies between 32.66 to 52.02 db. Watermarking, NIFTI, MSVD, LWT, QR and Image.

Prediction of Heart Failure in Children with Congenital Heart Disease Based on Multichannel LSTM

Heart failure (HF) is a complicated clinical illness caused by a variety of primary and secondary causes, as well as increased infection pathways, that are associated with higher risk, illness, and costs. The overall incidence of congenital heart disease is approximately high and is the leading cause of death in infants and children. In this study, we present a novel computational model based on ECNN-LSTM for detecting congenital heart disease in real time and assessing its developing course objectively. The proposed model is a multiresolution singular value decomposition for congenital heart disease prediction in infants and children. Firstly, in the whole-life vibration time-domain signal of newborns and toddlers, the multiresolution singular value decomposition approach is employed to create approximation and detailed signals with different resolutions. Secondly, the health phases are split into the smooth running stage of newborns and toddlers, as well as the standard deviation of the earliest moments. Thirdly, the two-layer one-dimensional convolutional neural network structure divides the quick degradation stage, which can offer degradation information. Finally, the prediction of congenital heart disease is finished on LSTM utilizing the MSE loss function to unify the assessment scale. Validation with lifespan data demonstrates the feasibility and effectiveness of the proposed model. Moreover, the existing models have the insufficient ability for life span feature characterization in congenital heart disease prediction in infants and children.

A sparse representation based compression of fused images using WDR coding

In this paper, a sparse representation based compression of fused images is proposed using Multi-Resolution Singular Value Decomposition (MSVD). The main idea of this work is to identify significant and less significant details using MSVD. The core information is fused using the absolute maximum rule while the less significant information is fused using sparse representation. The fused significant information is compressed using wavelet difference reduction coding. On the flip side, the fused less significant information is compressed using quantization and Huffman encoding. On the receiver side, the proposed recovery algorithm can be used to obtain the fused image. The superiority of the proposed technique can be analyzed from the comparison of the proposed work with some related work.

Health Condition Identification of Rolling Element Bearing Based on Gradient of Features Matrix and MDDCs-MRSVD

Bearing is a key component in rotary machines, and the performance of the rotary machines mostly depends on the bearing health condition. In order to improve the safety and maintenance plan of the product based on the bearing condition, a monitoring indicator is constructed to identify the health condition of bearings in real time. Firstly, the vibration signal is processed by the proposed Maximal Difference of the Detail Components in Multi-Resolution Singular Value Decomposition (MDDCs-MRSVD) algorithm. Secondly, the features matrix is constructed by selected features to reflect the health condition of bearings. Then, the gradient standard deviation of each sampling time is obtained by the gradient in the amplitude direction of the features matrix. Finally, a monitoring indicator can be constructed to identify healthy stages of bearing. The proposed methods are verified via the tested datasets provided by Intelligent Maintenance Systems, and Xi’an Jiaotong University and the Changxing Sumyoung Technology Co., Ltd. (XJTU-SY). The results indicate that the proposed method is efficient and accurate to monitor and identify the health stages of bearing in real time.

Secure Watermarking Scheme for Color DICOM Images in Telemedicine Applications

Teleradiology plays a vital role in the medical field, which permits transmitting medical and imaging data over a communication network. It ensures data reliability and provides convenient communication for clinical interpretation and diagnostic purposes. The transmission of this medical data over a network raises the problems of legal, ethical issues, privacy, and copyright authenticity. The copyright protection of medical images is a significant issue in the medical field. Watermarking schemes are used to address these issues. A gray-level or binary image is used as a watermark frequently in color image watermarking schemes. In this paper, the authors propose a novel non-blind medical image watermarking scheme based on 2-D Lifting Wavelet Transform (LWT), Multiresolution Singular Value Decomposition (MSVD), and LU factorization to improve the robustness and authenticity of medical images. In this scheme, multiple color watermarks are embedded into the colored DICOM (Digital Imaging and Communications in Medicine) images obtained from Color Doppler images (DICOM format), and the average results achieved by our proposed scheme is 46.84 db for Peak Signal-to-Noise Ratio (PSNR), 37.46 db for Signal-to-Noise Ratio (SNR), 0.99 for Quality of Image and 0.998 for Normalized Correlation for various image processing attacks. These results make our watermarking technique an ideal candidate for medical image watermarking.

Denoising Application of Magnetotelluric Low-Frequency Signal Processing

As magnetotelluric (MT) is an important method for exploring the geoelectrical structure of the underground, it has motivated in-depth research and application by many geophysicists. Nevertheless, due to the influence of the environment, the collected data are interfered with strong humanistic noise, which might result in a loss of their authenticity. To solve the above problems, many time-frequency domain methods have emerged. Based on the advantages of singular value decomposition (SVD) denoising, we propose a method of magnetotelluric noisy data processing based on multiresolution singular value decomposition (MSVD) and improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), which overcomes the lack of flexibility in the construction of the matrix in SVD data processing. First, we introduce a new signal processing method by generalizing SVD to MSVD to obtain more accurate signal characteristics. Due to the difficulty of matrix selection, we suggest the singular value contribution rate as the standard to determine the suitable Hankel matrix and use MSVD to perform effective decomposition. Second, we propose the ICEEMDAN algorithm for removing impulse noise, which efficiently processes each modal component through adaptively decomposition of different thresholds. Experiments on synthetic and realistic data demonstrate that our proposed method can separate the large-scale contours of the magnetotelluric noisy data and improve the time-domain waveform quality of low-frequency signal. The apparent resistivity-phase curves, coherence and SNR are all obviously promoted.

Face recognition system based on the multi-resolution singular value decomposition fusion technique

This study proposes a Fusion, Feature-Level, Face Recognition System (FFLFRS) that is based on the Multi-Resolution, Singular Value Decomposition (MSVD) fusion technique. Face recognition in the FFLFRS is achieved via four processes: face detection, feature extraction, feature fusion, and face classification. In this system, the most significant face features (that is, the eyes, nose, and mouth) are first detected. Then, local and global features are extracted by the Local Binary Pattern (LBP) and Principal Component Analysis (PCA) extraction approaches. Afterwards, the extracted features are fused by the MSVD method and classified by the Artificial Neural Network (ANN). The proposed FFLFRS was verified on 10,000 face images drawn from the face images database of the Olivetti Research Laboratory (ORL). Face recognition performance of this system was contrasted with levels of performance of three state of the art, fusion-level, face recognition systems (FRSs) depending on the Frequency Partition (FP), Laplacian Pyramid (LP), and Covariance Intersection (CI) fusion methods. Ten-thousand images were employed to test the proposed model and assess its performance, which was evaluated in terms of changes in pose, illumination, and expression, besides low resolution and presence of occlusion. The face recognition results of the proposed FFLFRS are encouraging. This system proved to be effective in dealing with images having challenges to face recognition and it could achieve a recognition accuracy as high as 97.78%.