De-noising Method Research Articles

An improved attentive residue multi-dilated network for thermal noise removal in magnetic resonance images

Magnetic resonance imaging (MRI) technology is crucial in the medical field, but the thermal noise in the reconstructed MR images may interfere with the clinical diagnosis. Removing the thermal noise in MR images mainly contains two challenges. First, thermal noise in an MR image obeys Rician distribution, where the statistical features are not consistent in different regions of the image. In this case, conventional denoising methods like spatial convolutional filtering will not be appropriate to deal with it. Second, details and edge information in the image may get damaged while smoothing the noise. This paper proposes a novel deep-learning model to denoise MR images. First, the model learns a binary mask to separate the background and signal regions of the noised image, making the noise left in the signal region obey a unified statistical distribution. Second, the model is designed as an attentive residual multi-dilated network (ARM-Net), composed of a multi-branch structure, and supplemented with a frequency-domain-optimizable discrete cosine transform module. In this way, the deep-learning model will be more effective in removing the noise while maintaining the details of the original image. Furthermore, we have also made improvements on the original ARM-Net baseline to establish a new model called ARM-Net v2, which is more efficient and effective. Experimental results illustrate that over the BraTS 2018 dataset, our method achieves the PSNR of 39.7087 and 32.6005 at noise levels of 5% and 20%, which realizes the state-of-the-art performance among existing MR image denoising methods.

The Denoising Method With Variational Mode Decomposition for Signal-Dependent Counting Noise in Molecular Communications

The Denoising Method With Variational Mode Decomposition for Signal-Dependent Counting Noise in Molecular Communications

An extraction method for early fault features of rolling bearings based on resonance sparse signal decomposition and non-convex second-order total variation denoising

A primary challenge in the field of fault diagnosis is extracting weak fault characteristics of bearings under large background noise and non-stationary conditions. Due to significant noise interference in the signals of rolling bearings, resonance sparse signal decomposition (RSSD) cannot efficiently extract the transient impact components in the early failure stage and the total variation denoising (TVD) method distorts signal waveforms. In this study, a combined extraction method for early fault features of rolling bearings based on RSSD and non-convex second-order total variation denoising (NCSOTVD) is proposed. Firstly, a non-convex function is introduced to define the regularisation term in the second-order TVD method, and the regularisation parameter and the convexity parameter in the NCSOTVD method are respectively screened using the noise standard deviation and the permutation entropy value to enhance the impact characteristics of signals and induce signal sparsity. The NCSOTVD model is solved using the optimisation-minimisation algorithm, so as to achieve noise reduction and feature enhancement of the vibration signals. Then, the low-resonance components of RSSD are denoised using the NCSOTVD method to highlight periodic pulse signals and extract the fault features of the rolling bearings. The simulation results and experimental data show that the method largely suppresses the noise interference, highlights the fault characteristics and reduces the problems of waveform distortion and poor sparsity of the TVD method in the denoising process.

Field Deployment of Natural Gas Pipeline Pre-Warning System With CEEMDAN Denoising Method

Field Deployment of Natural Gas Pipeline Pre-Warning System With CEEMDAN Denoising Method

Cascaded UNet for progressive noise residual prediction for structure-preserving video denoising

The prominence of high-quality video services has become so substantial that by 2030, it is estimated that approximately 80% of internet traffic will consist of videos. On the contrary, video denoising remains a relatively unexplored and intricate field, presenting more substantial challenges compared to image denoising. Many published deep learning video denoising algorithms typically rely on simple, efficient single encoder–decoder networks, but they have inherent limitations in preserving intricate image details and effectively managing noise information propagation for noise residue modelling. In response to these challenges, the proposed work introduces an innovative approach; in terms of utilization of cascaded UNets for progressive noise residual prediction in video denoising. This multi-stage encoder–decoder architecture is meticulously designed to accurately predict noise residual maps, thereby preserving the locally fine details within video content as represented by SSIM. The proposed network has undergone extensive end-to-end training from scratch without explicit motion compensation to reduce complexity. In terms of the more rigorous SSIM metric, the proposed network outperformed all video denoising methods while maintaining a comparable PSNR.

Research on remote reference denoising method based on non-coaxial and non-coplanar tunnel NMR detection

Abstract The limited space within the tunnel constrains the size of the antenna for NMR detection, thereby significantly impacting the signal-to-noise ratio (SNR) of NMR signals. Insufficient SNR data poses substantial challenges to obtaining reliable NMR signals. The paper presents a novel approach to address the challenge of strong background noise in tunnel environments and low SNR data by incorporating the ground multi-channel remote reference denoising method into tunnel NMR advance detection. Specifically designed for narrow tunnels, a multi-channel non-coaxial and non-coplanar remote reference denoising method is proposed. Firstly, the effectiveness of the non-coaxial, non-coplanar remote reference denoising method is verified in the laboratory environment. Secondly, the correlation between the detector antenna and the reference antenna is calculated theoretically to ensure the significant correlation between the detector antenna and the reference antenna. Finally, two processing methods of reference denoising and non-reference denoising are carried out respectively by combining the tunnel detection data. By comparing the inversion results and the engineering construction results, the effectiveness of non-coaxial and non-coplanar remote reference denoising methods in tunnel NMR detection is proved, which provides relevant research support for expanding the application of tunnel NMR detection technology.

Noise-reduction techniques for 1H-FID-MRSI at 14.1 T: Monte Carlo validation and in vivo application.

Proton magnetic resonance spectroscopic imaging (1H-MRSI) is a powerful tool that enables the multidimensional non-invasive mapping of the neurochemical profile at high resolution over the entire brain. The constant demand for higher spatial resolution in 1H-MRSI has led to increased interest in post-processing-based denoising methods aimed at reducing noise variance. The aim of the present study was to implement two noise-reduction techniques, Marchenko-Pastur principal component analysis (MP-PCA) based denoising and low-rank total generalized variation (LR-TGV) reconstruction, and to test their potential with and impact on preclinical 14.1 T fast in vivo 1H-FID-MRSI datasets. Since there is no known ground truth for in vivo metabolite maps, additional evaluations of the performance of both noise-reduction strategies were conducted using Monte Carlo simulations. Results showed that both denoising techniques increased the apparent signal-to-noise ratio (SNR) while preserving noise properties in each spectrum for both in vivo and Monte Carlo datasets. Relative metabolite concentrations were not significantly altered by either method and brain regional differences were preserved in both synthetic and in vivo datasets. Increased precision of metabolite estimates was observed for the two methods, with inconsistencies noted for lower-concentration metabolites. Our study provided a framework for how to evaluate the performance of MP-PCA and LR-TGV methods for preclinical 1H-FID MRSI data at 14.1 T. While gains in apparent SNR and precision were observed, concentration estimations ought to be treated with care, especially for low-concentration metabolites.

Fourth-stable stochastic resonance and its application to weak fault signal detection of rotating machinery

Weak characteristic extraction is vital for weak fault signal detection of machinery. Stochastic resonance (SR) is able to transfer noise energy into weak fault characteristic frequency excited by a defect of machines. However, the potential function in SR is vital to enhance weak fault characteristic frequency and determines the capability of SR to improve the signal-to-noise ratio (SNR) of a noisy signal. Now, common potential functions include monostable, bistable and even tri-stable potentials but fourth-stable SR has not been studied and applied to detect early fault characteristic frequency. In this paper, thus, we would investigate the behaviors of SR with a fourth-stable potential subject to additive noise, in which the approximate theoretical expression of the power done by SR is derived to demonstrate the fourth-stable Sr Then, a SR method with the fourth-stable potential is proposed to enhance weak fault characteristic frequency, in which these system parameters are adjusted by using SNR as the objective function and using genetic algorithms adaptively. In this paper, thus, Finally, the proposed method is verified by using a simulated signal with noise and two early fault experiment of rolling element bearings with different levels of defects on the outer and inner races. Moreover, the proposed method is compared with wavelet denoising and fast kurtogram methods. The comparisons indicate that the proposed method has the better performance for enhancing weak fault characteristic frequency or weak useful signals than other two methods and is available to weak fault signal detection of machinery.

Unsupervised Domain Adaptation for EM Image Denoising with Invertible Networks.

Electron microscopy (EM) image denoising is critical for visualization and subsequent analysis. Despite the remarkable achievements of deep learning-based non-blind denoising methods, their performance drops significantly when domain shifts exist between the training and testing data. To address this issue, unpaired blind denoising methods have been proposed. However, these methods heavily rely on image-to-image translation and neglect the inherent characteristics of EM images, limiting their overall denoising performance. In this paper, we propose the first unsupervised domain adaptive EM image denoising method, which is grounded in the observation that EM images from similar samples share common content characteristics. Specifically, we first disentangle the content representations and the noise components from noisy images and establish a shared domain-agnostic content space via domain alignment to bridge the synthetic images (source domain) and the real images (target domain). To ensure precise domain alignment, we further incorporate domain regularization by enforcing that: the pseudo-noisy images, reconstructed using both content representations and noise components, accurately capture the characteristics of the noisy images from which the noise components originate, all while maintaining semantic consistency with the noisy images from which the content representations originate. To guarantee lossless representation decomposition and image reconstruction, we introduce disentanglement-reconstruction invertible networks. Finally, the reconstructed pseudo-noisy images, paired with their corresponding clean counterparts, serve as valuable training data for the denoising network. Extensive experiments on synthetic and real EM datasets demonstrate the superiority of our method in terms of image restoration quality and downstream neuron segmentation accuracy. Our code is publicly available at https://github.com/sydeng99/DADn.

Triple-0: Zero-shot denoising and dereverberation on an end-to-end frozen anechoic speech separation network.

Speech enhancement is crucial both for human and machine listening applications. Over the last decade, the use of deep learning for speech enhancement has resulted in tremendous improvement over the classical signal processing and machine learning methods. However, training a deep neural network is not only time-consuming; it also requires extensive computational resources and a large training dataset. Transfer learning, i.e. using a pretrained network for a new task, comes to the rescue by reducing the amount of training time, computational resources, and the required dataset, but the network still needs to be fine-tuned for the new task. This paper presents a novel method of speech denoising and dereverberation (SD&D) on an end-to-end frozen binaural anechoic speech separation network. The frozen network requires neither any architectural change nor any fine-tuning for the new task, as is usually required for transfer learning. The interaural cues of a source placed inside noisy and echoic surroundings are given as input to this pretrained network to extract the target speech from noise and reverberation. Although the pretrained model used in this paper has never seen noisy reverberant conditions during its training, it performs satisfactorily for zero-shot testing (ZST) under these conditions. It is because the pretrained model used here has been trained on the direct-path interaural cues of an active source and so it can recognize them even in the presence of echoes and noise. ZST on the same dataset on which the pretrained network was trained (homo-corpus) for the unseen class of interference, has shown considerable improvement over the weighted prediction error (WPE) algorithm in terms of four objective speech quality and intelligibility metrics. Also, the proposed model offers similar performance provided by a deep learning SD&D algorithm for this dataset under varying conditions of noise and reverberations. Similarly, ZST on a different dataset has provided an improvement in intelligibility and almost equivalent quality as provided by the WPE algorithm.

Seismic data denoising using the 3D curvelet transform method

One of the major difficulties in processing and interpreting seismic data is the contamination of seismic signals by noise from numerous sources. Conventional denoising methods are mostly used for processing 2D seismic data, but noise also exists in the 3D data space, resulting in poor results using conventional denoising methods. Consequently, here, for seismic data denoising, a multiscale and multidirectional 3D curvelet transform was adopted. Our study combined the core principle of the 3D curvelet approach with the threshold iteration method to denoise simulated and actual seismic data with varying signal-to-noise ratios, and the results were quantitatively compared with the processing outcomes of existing denoising methods. The effectiveness of our method is illustrated using both genuine 2D and 3D post-stack seismic data, and synthetic 2D sections with added white and colored noise. Finally, we demonstrate how to prepare the data for frequency-domain full-waveform inversion using curvelet denoising. Despite the complexity of the procedure used to create the training samples, testing results using synthetic and real seismic data demonstrate that this method has mastered the capacity to suppress Gaussian and super-Gaussian noise from various training samples.

Self-supervised tomographic image noise suppression via residual image prior network

Computed tomography (CT) denoising is a challenging task in medical imaging that has garnered considerable attention. Supervised networks require a lot of noisy-clean image pairs, which are always unavailable in clinical settings. Existing self-supervised algorithms for suppressing noise with paired noisy images have limitations, such as ignoring the residual between similar image pairs during training and insufficiently learning the spectrum information of images. In this study, we propose a Residual Image Prior Network (RIP-Net) to sufficiently model the residual between the paired similar noisy images. Our approach offers new insights into the field by addressing the limitations of existing methods. We first establish a mathematical theorem clarifying the non-equivalence between similar-image-based self-supervised learning and supervised learning. It helps us better understand the strengths and limitations of self-supervised learning. Secondly, we introduce a novel regularization term to model a low-frequency residual image prior. This can improve the accuracy and robustness of our model. Finally, we design a well-structured denoising network capable of exploring spectrum information while simultaneously sensing context messages. The network has dual paths for modeling high and low-frequency compositions in the raw noisy image. Additionally, context perception modules capture local and global interactions to produce high-quality images. The comprehensive experiments on preclinical photon-counting CT, clinical brain CT, and low-dose CT datasets, demonstrate that our RIP-Net is superior to other unsupervised denoising methods.

Ground-truth-free Deep Learning for 3-D Seismic Denoising and Reconstruction with Channel Attention Mechanism

Seismic denoising methods using supervised methods rely on a large number of high-quality paired training datasets to reach satisfactory performances. There are two ways to generate labels for network training: one is to simulate the synthetic data using the wave equation, and the other is to utilize denoised data obtained via conventional methods. However, using these labels will limit the networks' noise attenuation performance compared with using large volumes of noise-free data as labels. Here, we propose a ground-truth-free way for three-dimensional (3-D) seismic data processing. First, we use the 3-D patch scheme to divide the noisy seismic data into many fixed-size blocks and then flatten the obtained 3-D patches to expand the training set and capture more higher-order waveform characteristics from the input noisy data. Next, the obtained training dataset is sent into the proposed deep learning (DL) network, where the encoder blocks compress the feature map to extract the waveform features, and the decoder blocks reconstruct the denoised feature map. Notably, the convolutional bottleneck attention module (CBAM) and efficient channel attention (ECA) module are applied to guide the network to focus on signal fluctuation features with fewer network parameters. In addition, the concatenation mechanism is used to enable deep networks to reuse shallow-layer waveform features and mitigate overfitting during training. Finally, the unpatching scheme is used to reconstruct the denoised 3-D seismic data. Numerical experiments demonstrate that the proposed method outperforms benchmark approaches in terms of signal-to-noise ratio (SNR) improvement and useful signal preservation.

A Comprehensive Study of Object Tracking in Low-Light Environments.

Accurate object tracking in low-light environments is crucial, particularly in surveillance, ethology applications, and biometric recognition systems. However, achieving this is significantly challenging due to the poor quality of captured sequences. Factors such as noise, color imbalance, and low contrast contribute to these challenges. This paper presents a comprehensive study examining the impact of these distortions on automatic object trackers. Additionally, we propose a solution to enhance the tracking performance by integrating denoising and low-light enhancement methods into the transformer-based object tracking system. Experimental results show that the proposed tracker, trained with low-light synthetic datasets, outperforms both the vanilla MixFormer and Siam R-CNN.

Complete Ensemble Empirical Mode Decomposition and Wavelet Algorithm Denoising Method for Bridge Monitoring Signals

Complete Ensemble Empirical Mode Decomposition and Wavelet Algorithm Denoising Method for Bridge Monitoring Signals

Progressive Feature Fusion Attention Dense Network for Speckle Noise Removal in OCT Images.

Although deep learning for Big Data analytics has achieved promising results in the field of optical coherence tomography (OCT) image denoising, the low recognition rate caused by complex noise distribution and a large number of redundant features is still a challenge faced by deep learning-based denoising methods. Moreover, the network with large depth will bring high computational complexity. To this end, we propose a progressive feature fusion attention dense network (PFFADN) for speckle noise removal in OCT images. We arrange densely connected dense blocks in the deep convolution network, and sequentially connect the shallow convolution feature map with the deep one extracted from each dense block to form a residual block. We add attention mechanism to the network to extract the key features and suppress the irrelevant ones. We fuse the output feature maps from all dense blocks and input them to the reconstruction output layer. We compare PFFADN with the state-of-the-art denoising algorithms on retinal OCT images. Experiments show that our method has better improvement in denoising performance.

Stock Prices Forecasting and Optimization Strategies Based on Support Vector Machines

With the global trend of digitization gaining prominence, the usage of machine learning methods such as Support Vector Machines and Reinforcement Learning for stock price prediction is becoming a hot topic. Over the past 40 years, China's economic market has undergone significant changes since the country's reform and opening up. In this study, the closing price and return of China's CSI 300 stock index are used as the database, and various data processing methods such as wavelet domain denoising, RSI screening and various SVM model optimization methods such as grid search and cross-validation are used to predict the upward or downward trend of stocks on the day after. The results of the study are presented by the model evaluation report and the heat map of the confusion matrix, which shows that the model prediction accuracy is 61% with the default parameters, and the accuracy improves to 67% after optimization. The results indicate that support vector machines are effective in stock price prediction, but there is still room for further improvement. This paper offers a potential approach that can increase return on investment and assist investors and financial institutions in making more informed investment decisions.

Combining optimal SASS (Sparsity Assisted Signal Smoothing) and notch filters for transient measurement signals denoising of large power generating unit

Denoising and harmonic interference removal from the measurement signal is crucial for accurate parameters calculation of a generating unit model. Determining the actual parameters values requires various unit transient signals data recorded during a well-defined testing procedure. However, the recorded signals usually contain noise and harmonic interferences which are difficult to remove due to a presence of different kinds of discontinuities.The paper presents a proposal of optimal denoising and harmonic interference removal method from transient measurement signals of 200 MW power generating unit. The method combines notch filters for harmonic rejection followed by SASS (Sparsity Assisted Signal Smoothing) algorithm for wideband noise rejection. Parameters of the notch filters as well as the SASS are determined using popular optimization algorithms (simulated annealing, Nelder-Mead simplex and gradient algorithm).Performance of the proposed method is compared to its direct Linear Time-Invariant (LTI) competitors such as low-pass and notch filters alone, notch and low-pass filters combined and to the optimal SASS method alone. The performance evaluation is based on a set of specially generated test signals which are using the developed mathematical model of the signals (to set the level of noise and interferences) as well as the mathematical model of the power generating unit (to provide the reference clean signals). The obtained results confirm that the optimal notch-SASS method might be a valuable noise removal tool for considered class of signals and leads to superior results than the LTI filters or the SASS method alone.

Attenuation of non-stationary random noise in ground penetrating radar data based on time-varying filtering

The attenuation of random noise is significant in the enhancement of the signal-to-noise ratio in ground penetrating radar (GPR) data. However, the non-stationary nature of random noise poses significant challenges to most conventional denoising methods. We propose a time-varying bandpass filter (TVBF) based on the local maxima multiple synchrosqueezing transform (LMMSST), referred to as TVBFLMMSST, for the attenuation of non-stationary random noise in GPR data. The LMMSST method is introduced to achieve highly compressed time–frequency representation (TFR). By integrating an instantaneous frequency trajectory detection algorithm, we design the TVBF to successfully separate useful signals from random noise. The efficacy of the TVBFLMMSST is validated through synthetic and field experiments, demonstrating its capability to effectively isolate components of useful signals and noise signals. Compared to classical denoising methods, TVBFLMMSST exhibits robust performance in attenuating random noise while preserving the amplitude of useful signals.

A Denoising Convolutional Autoencoder for SNR Enhancement in Chemical Exchange Saturation Transfer imaging: (DCAE-CEST).

To develop a SNR enhancement method for chemical exchange saturation transfer (CEST) imaging using a denoising convolutional autoencoder (DCAE), and compare its performance with state-of-the-art denoising methods. The DCAE-CEST model encompasses an encoder and a decoder network. The encoder learns features from the input CEST Z-spectrum via a series of 1D convolutions, nonlinearity applications and pooling. Subsequently, the decoder reconstructs an output denoised Z-spectrum using a series of up-sampling and convolution layers. The DCAE-CEST model underwent multistage training in an environment constrained by Kullback-Leibler divergence, while ensuring data adaptability through context learning using Principal Component Analysis processed Z-spectrum as a reference. The model was trained using simulated Z-spectra, and its performance was evaluated using both simulated data and in-vivo data from an animal tumor model. Maps of amide proton transfer (APT) and nuclear Overhauser enhancement (NOE) effects were quantified using the multiple-pool Lorentzian fit, along with an apparent exchange-dependent relaxation metric. In digital phantom experiments, the DCAE-CEST method exhibited superior performance, surpassing existing denoising techniques, as indicated by the peak SNR and Structural Similarity Index. Additionally, in vivo data further confirms the effectiveness of the DCAE-CEST in denoising the APT and NOE maps when compared to other methods. While no significant difference was observed in APT between tumors and normal tissues, there was a significant difference in NOE, consistent with previous findings. The DCAE-CEST can learn the most important features of the CEST Z-spectrum and provide the most effective denoising solution compared to other methods.