Near-surface-related Nonstationary Coherent Noise Suppression Using a Physically Constrained Deep Neural Network

Real-time filtering adaptive algorithms for non-stationary noise in electrocardiograms

During seismic investigations, multiple and unexpected sources may cause serious interference on seismic records, and coherent noise generated by another unwanted active source could result in extremely poor data quality. Because airgun arrays have been widely used as the sound source in marine seismic surveys, the noise generated by another airgun array usually has similar characteristics to the primary signals in both frequency bands and wave forms, so the suppression of this type of coherent noise is very difficult. In practice, seismic crews try to avoid conducting multiple surveys simultaneously in a same area, so the source interference problem normally does not occur, and suppression of coherent noise from another active source has rarely been discussed and proposed before. This paper presents a dataset in which part of the records are contaminated by shot noise from another seismic vessel, and proposes a hybrid approach to suppress the coherent noise from that unwanted seismic source. Noise subtraction and primary signal preservation within different data properties are considered to begin the noise suppression. Based on different noise characteristics from various source directions and wave propagation paths, coherence noise can be separated from primary signals in frequency–wave number (F–K), frequency–time (F–T) and intercept time–slowness (tau–p) domains, respectively. This hybrid coherent noise suppression approach involves applying three different filters, F–K, F–T and tau–p, to the contaminated dataset. Our results show that most of the coherent noise generated by another seismic source could be suppressed, and seismic images could be substantially improved.

An illumination system for coherent noise suppression is envisioned. The principle of coherent noise suppression is introduced, and this illumination system is designed this way. The illumination system is simulated, and the hybrid sequence model of Zemax analyzes the irradiance. To verify the noise suppression effect of this illumination system, it is used as the illumination part of the Fizeau interferometer. The interference process is simulated under the nonsequential model of Zemax to obtain an off-axis light source with a ring radius of 0.32 mm and a numerical aperture of 0.14. The Fizeau interferometer with a conventional light source is also simulated. A comparison experiment is set up to generate the same noise point in the two interferometers using different illumination modes to trace and produce the same four interference fringes with the same interferometric cavity length of 20 mm and the same tilt angle of the measured surface of 0.057°. Compared with the interferograms in the conventional illumination mode, the interference fringes formed by the illumination of this study are almost undamaged, and the near-complete interference information can be retained. The interferometer system with this light source was built and the test results were verified, and it was found that it could achieve the measurement accuracy of 1/20 wavelength and the measurement stability of 1.219 nm, and it also had a good contrast of interference fringe.

Digital holographic microscopy (DHM) is a topographic measurement technique that permits full-field, nondestructive, dynamic, quantitative amplitude, and phase-contrast imaging. The technique may realize the lateral resolution with submicron scale and the longitudinal resolution with subnanometer scale, respectively. Improving imaging quality has always been the research focus in DHM since it has a direct effect on the precise topographic measurement. In this paper, the recent progress on phase aberration compensation and coherent noise suppression is reviewed. Included in this review are the hologram spectrum’s centering judgment methods of side band in tilt phase error compensation, the physical and numerical compensation methods in phase aberration compensation, and the single-shot digital process methods in coherent noise suppression. The summaries and analyses for these approaches can contribute to improving the imaging quality and reducing the measurement error of DHM, which will further promote the wider applications of DHM in the topographic measurement fields, such as biology and micro-electro mechanical systems.

The squelch problem of ultra-short wave communication under non-stationary noise and low Signal-to-Noise Ratio (SNR) in a complex electromagnetic environment is still challenging. To alleviate the problem, we proposed a squelch algorithm for ultra-short wave communication based on a deep neural network and the traditional energy decision method. The proposed algorithm first predicts the speech existence probability using a three-layer Gated Recurrent Unit (GRU) with the speech banding spectrum as the feature. Then it gets the final squelch result by combining the strength of the signal energy and the speech existence probability. Multiple simulations and experiments are done to verify the robustness and effectiveness of the proposed algorithm. We simulate the algorithm in three situations: the typical Amplitude Modulation (AM) and Frequency Modulation (FM) in the ultra-short wave communication under different SNR environments, the non-stationary burst-like noise environments, and the real received signal of the ultra-short wave radio. The experimental results show that the proposed algorithm performs better than the traditional squelch methods in all the simulations and experiments. In particular, the false alarm rate of the proposed squelch algorithm for non-stationary burst-like noise is significantly lower than that of traditional squelch methods.

Bayesian Estimators are very useful in speech enhancement and noise reduction. But, it is noted that the traditional estimators process only amplitudes and the phase is left unprocessed. Among the Bayesian estimators, Super- Gaussian based estimators provide improved noise reduction. Super-Gaussian Bayesian estimators, which uses processed phase information for estimation of amplitudes provides further improved results. In this work, the Complex speech coefficients given Uncertain Phase (CUP) based Bayesian estimators like CUP-GG (CUP Estimator with speech spectral coefficients assumed as Gamma and noise spectral coefficients as Generalized Gamma), CUP-NG (Speech as Nakagami) are compared under white noise, pink noise, Babble noise and Non-Stationary factory noise conditions. The statistical estimators show less effective results under completely non-stationary assumptions like non-stationary factory noise, babble noise etc. Non-negative Matrix Factorization (NMF) based algorithms show better performance for non stationary noises. The drawback of NMF is, it requires apriori knowledge about speech. This drawback can be overcome by taking the advantages of both statistical approaches and NMF approaches. NR-NMF and WR-NMF speech enhancement methods are developed by providing posteriori regularization based on statistical assumption of speech and noise DFT coefficients distribution. Also a speech enhancement method which uses CUP-GG estimator and NMF with online noise bases update are considered for comparison. The progress in neural network based approaches for speech enhancement further shown that with large dataset and better training, the speech enhancement algorithms results in improved results. In this work, the neural network approach for speech enhancement is implemented and compared the method with traditional estimators and NMF approaches. For generalization of unseen noise types the proposed neural network approach uses dropout. Also for training the network, the features obtained from apriori SNR and aposteriori SNR is used in this method. The objective of this paper is to analyze the performance of speech enhancement methods based on Neural Network, NMF and statistical based. The objective performance measures Perceptual Evaluation of Speech Quality (PESQ), Short-Time Objective Intelligibility (STOI), Signal to Noise Ratio (SNR), Segmental SNR (Seg SNR) are considered for comparison.

The paper studies the influence of coherent noises on the quality of images of particles reconstructed from digital holograms. Standard indicators (for example, signal-to-noise ratio) and such indicators as the boundary contrast and boundary intensity jump previously proposed by the authors are used to quantify the image quality. With the use of these parameters, for examples of some known methods of suppressing coherent noises in a holographic image (eliminating the mutual influence of virtual and real images in in-line holography, and time averaging), the features and ranges of applicability of such correction were determined. It was shown that the use of the complex field amplitude reconstruction method based on the Gerchberg–Saxton algorithm and the spatial-frequency method improves the quality of determining the particle image boundary (by boundary intensity jump) starting from the distance between a hologram and a particle, which is about twice the Rayleigh distance. In physical experiments with model particles, averaging methods were studied to suppress non-stationary coherent noises (speckles). It was also shown that averaging over three digital holograms or over three holographic images is sufficient to provide a quality of particle image boundary suitable for particle recognition. In the case of multiple scattering, when it is necessary to impose a limit on the working volume length (depth of scene) of the holographic camera, the paper provides estimates that allow selecting the optimal working volume length. The estimates were made using the example of a submersible digital holographic camera for plankton studies.

Speech-in-noise perception is a major problem for users of cochlear implants (CIs), especially with non-stationary background noise. Noise-reduction algorithms have produced benefits but relied on a priori information about the target speaker and/or background noise. A recurrent neural network (RNN) algorithm was developed for enhancing speech in non-stationary noise and its benefits were evaluated for speech perception, using both objective measures and experiments with CI simulations and CI users. The RNN was trained using speech from many talkers mixed with multi-talker or traffic noise recordings. Its performance was evaluated using speech from an unseen talker mixed with different noise recordings of the same class, either babble or traffic noise. Objective measures indicated benefits of using a recurrent over a feed-forward architecture, and predicted better speech intelligibility with than without the processing. The experimental results showed significantly improved intelligibility of speech in babble noise but not in traffic noise. CI subjects rated the processed stimuli as significantly better in terms of speech distortions, noise intrusiveness, and overall quality than unprocessed stimuli for both babble and traffic noise. These results extend previous findings for CI users to mostly unseen acoustic conditions with non-stationary noise.

The research subject of this article is the methods of locally adaptive filtering of non-stationary signals. The goal is to develop a locally-adaptive algorithm for non-stationary noise (from the viewpoint of its time-varying variance) suppression in signals characterized by a different behavior of the informative component, with restricted apriori information about the signal model and noise variance. The tasks are to investigate the effectiveness of the proposed local-adaptive myriad filter using numerical statistical estimates of processing quality for a complex model of one-dimensional process that contains different elementary signals in various additive Gaussian noise variance variations; to investigate the effectiveness of non-stationary noise suppression for model and real signals. The methods are integral and local indicators of filter quality according to the criteria of the mean square error have been obtained using numerical simulation (via Monte Carlo analysis). The following results have been obtained: a noise- and signal-adapting myriad filter for the suppressing of non-stationary noise with significantly varying variance in signals with different behaviors of the informative component is proposed. Statistical estimates of the filter quality, evaluated by numerical simulation, show a higher efficiency of the proposed local-adaptive myriad filter in conditions of different noise levels compared to the other highly efficient locally-adaptive filters. Practically, total preservation of a signal at very low noise levels, minimal dynamical errors caused by filtering at low and middle noise levels, and more effective noise suppression at high values of noise variance are demonstrated. The analysis of output signals and plots of parameters for local adaptation and adaptable parameters confirm the high efficiency and correct operation of the investigated locally-adaptive algorithms. The high robust properties of these nonlinear filters are shown, as well as the expedience of using to spike the elimination of the previous robust Hampel filter in which the median operation is replaced by a myriad one. Examples displaying the high quality of non-stationary noise suppression in a biomedical signal of electronystagmogram are presented. Conclusions. The scientific novelty of the obtained results is the development of locally-adaptive myriad filters with time-varying noise- and signal-dependent parameters for de-noising processes with non-stationary signal behavior and noise variance. This filter does not require time for parameter adaptation and their exact adjustment, a priori knowledge of the signal model and noise variance, and can be applied in a quasi-real-time mode. The proposed algorithm of noise- and signal-adapting myriad filtering algorithm improves the quality of signal processing in difficult conditions of significant noise non-stationarity (variance variation).

The research subject of the article is the methods of locally adaptive filtering of non-stationary (from the point of view of its variance) noise in long-term electrocardiogram (ECG) signals. The goal is to develop locally adaptive algorithms for filtering noise with different a priori unknown levels of variance in real-time for ECG signals recorded with a standard sampling rate of 500 Hz. The tasks to be solved are: to investigate the effectiveness of the developed adaptive ECG filtering algorithms using numerical statistical estimates of processing quality in a wide range of additive Gaussian noise variance variation, to investigate the suppression of real non-stationary electromyographic (EMG) noise, and to analyze the application for normal and pathological ECG signals. The methods are integral and local indicators of the filter quality according to the criteria of the mean square error and the signal-to-noise ratio was obtained using numerical simulation (via Monte Carlo analysis). The following results were obtained: an adaptive method for real-time suppression of non-stationary noise in the ECG is proposed, the one-pass and the two-pass algorithms, and the algorithm with selective depending on the preliminary estimates of noise levels re-filtering have been developed on the method basis. Statistical estimates of the filters' efficiency and analysis of their outputs show a high degree of suppression of the noise with different levels of variance in the ECGs. The distortions absence while processing QRS-complex and high efficiency of suppression of Gaussian and real EMG noise with varying variance are demonstrated. The analysis of the output signals and plots of the local adaptation parameters and the adaptable parameters of the proposed algorithms confirms the high efficiency of filtering. The developed algorithms have been successfully tested for normal and pathological ECG signals. Conclusions. The scientific novelty of the results is the development of a locally adaptive method with noise and signal-dependent filter parameters switching and of the adaptive algorithms based on this method for non-stationary noise reduction in the ECG in real-time. This method does not require time for filter parameters adaptation and a priori information about the noise variance, and it has a high-speed performance in real-time mode.

PreviousNext No AccessSecond International Meeting for Applied Geoscience & EnergyAiding self-supervised coherent noise suppression by the introduction of signal segmentation using blind-spot networksAuthors: Sixiu LiuClaire BirnieTariq AlkhalifahAndrey BakulinSixiu LiuKAUSTSearch for more papers by this author, Claire BirnieKAUSTSearch for more papers by this author, Tariq AlkhalifahKAUSTSearch for more papers by this author, and Andrey BakulinSaudi AramcoSearch for more papers by this authorhttps://doi.org/10.1190/image2022-3742756.1 SectionsSupplemental MaterialAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractBlind-spot networks have been shown to be natural noise suppressors under the assumption that noise is unpredictable based on the information fed into the network during training. Trained in a self-supervised manner, such approaches only utilise the original raw data to determine to remove the noise. In this work, we propose two novel elements for enhancing blind-spot denoising: (1) the introduction of a 2-class segmentation task to aid the network in identification of interest areas of signals that require particular attention during denoising, and; (2) the introduction of a trace-wise noise mask designed to obscure the coherency of noise from being observed by the network. The joint scheme is achieved by introducing a joint loss function to balance between the two deep learning tasks. As such, the final joint scheme is the combination of a self-supervised, blind-spot denoising procedure and a supervised segmentation procedure. We illustrate how the joint scheme can improve the denoising performance of the network, hypothesising that this is due to the introduction of prior information guiding the denoising procedure to areas of focus. Preliminary results from synthetic data contaminated by trace-wise noise, show an increase in the structural similarity index from 0.989 to 0.995, when comparing the optimal jointscheme versus the pure denoising procedure. Future work will extend the procedure to field data where rule-based approaches will be used to generate the segmentation labels.Keywords: self-supervised, learning, coherent noise suppression, blind-spot networks, segmentationPermalink: https://doi.org/10.1190/image2022-3742756.1FiguresReferencesRelatedDetails Second International Meeting for Applied Geoscience & EnergyISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2022 Pages: 3694 publication data© 2022 Published in electronic format with permission by the Society of Exploration Geophysicists and the American Association of Petroleum GeologistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 15 Aug 2022 CITATION INFORMATION Sixiu Liu, Claire Birnie, Tariq Alkhalifah, and Andrey Bakulin, (2022), "Aiding self-supervised coherent noise suppression by the introduction of signal segmentation using blind-spot networks," SEG Technical Program Expanded Abstracts : 2857-2861. https://doi.org/10.1190/image2022-3742756.1 Plain-Language Summary Keywordsself-supervised learningcoherent noise suppressionblind-spot networkssegmentationPDF DownloadLoading ...

Hearable is a recently emerging term that describes a wireless earpiece that enhances the user's listening experience in various acoustic environment. Another important feature of hearable devices is their capability to improve speech communication in difficult social settings, which usually consist of a mixture of different non-stationary noise. In this paper, we present techniques to suppress a combination of non-stationary noise and transient noise. This is achieved by employing a combined noise suppression filter based on prediction and masking to achieve impulsive noise suppression. Experimental results highlight the robustness of the proposed algorithm in suppressing the transient noise while maintaining the speech components, without requiring any prior information of the noise.

Nonlinear active noise control system based on correlated EMD and Chebyshev filter

An active noise control method of non-stationary noise under time-variant secondary path

We have developed a method for suppressing coherent noise from seismic data by using the morphological differences between the noise and the signal. This method consists of three steps: First, we applied a dictionary learning method on the data to extract a redundant dictionary in which the morphological diversity of the data is stored. Such a dictionary is a set of unit vectors called atoms that represent elementary patterns that are redundant in the data. Because the dictionary is learned on data contaminated by coherent noise, it is a mix of atoms representing signal patterns and atoms representing noise patterns. In the second step, we separate the noise atoms from the signal atoms using a statistical classification. Hence, the learned dictionary is divided into two subdictionaries: one describing the morphology of the noise and the other one describing the morphology of the signal. Finally, we separate the seismic signal and the coherent noise via morphological component analysis (MCA); it uses sparsity with respect to the two subdictionaries to identify the signal and the noise contributions in the mixture. Hence, the proposed method does not use prior information about the signal and the noise morphologies, but it entirely adapts to the signal and the noise of the data. It does not require a manual search for adequate transforms that may sparsify the signal and the noise, in contrast to existing MCA-based methods. We develop an application of the proposed method for removing the mechanical noise from a marine seismic data set. For mechanical noise that is coherent in space and time, the results show that our method provides better denoising in comparison with the standard FX-Decon, FX-Cadzow, and the curvelet-based denoising methods.