Showing papers on "Noise measurement published in 2020"
14 Jun 2020
TL;DR: A frequency-based decompositionand- enhancement model that first learns to recover image objects in the low-frequency layer and then enhances high-frequency details based on the recovered image objects and outperforms state-of-the-art approaches in enhancing practical noisy low-light images.
Abstract: Low-light images typically suffer from two problems. First, they have low visibility (i.e., small pixel values). Second, noise becomes significant and disrupts the image content, due to low signal-to-noise ratio. Most existing lowlight image enhancement methods, however, learn from noise-negligible datasets. They rely on users having good photographic skills in taking images with low noise. Unfortunately, this is not the case for majority of the low-light images. While concurrently enhancing a low-light image and removing its noise is ill-posed, we observe that noise exhibits different levels of contrast in different frequency layers, and it is much easier to detect noise in the lowfrequency layer than in the high one. Inspired by this observation, we propose a frequency-based decompositionand- enhancement model for low-light image enhancement. Based on this model, we present a novel network that first learns to recover image objects in the low-frequency layer and then enhances high-frequency details based on the recovered image objects. In addition, we have prepared a new low-light image dataset with real noise to facilitate learning. Finally, we have conducted extensive experiments to show that the proposed method outperforms state-of-the-art approaches in enhancing practical noisy low-light images.
167 citations
14 Jun 2020
TL;DR: This work presents a method for training a neural network to perform image denoising without access to clean training examples or access to paired noisy training examples, which produces results which are competitive with other learned methods which require richer training data, and outperforms traditional non-learned Denoising methods.
Abstract: We present a method for training a neural network to perform image denoising without access to clean training examples or access to paired noisy training examples. Our method requires only a single noisy realization of each training example and a statistical model of the noise distribution, and is applicable to a wide variety of noise models, including spatially structured noise. Our model produces results which are competitive with other learned methods which require richer training data, and outperforms traditional non-learned denoising methods. We present derivations of our method for arbitrary additive noise, an improvement specific to Gaussian additive noise, and an extension to multiplicative Bernoulli noise.
136 citations
14 Jun 2020
TL;DR: This paper presents a holistic framework to train deep neural networks in a way that is highly invulnerable to label noise and achieves excellent performance on large-scale datasets with real-world label noise.
Abstract: Collecting large-scale data with clean labels for supervised training of neural networks is practically challenging. Although noisy labels are usually cheap to acquire, existing methods suffer a lot from label noise. This paper targets at the challenge of robust training at high label noise regimes. The key insight to achieve this goal is to wisely leverage a small trusted set to estimate exemplar weights and pseudo labels for noisy data in order to reuse them for supervised training. We present a holistic framework to train deep neural networks in a way that is highly invulnerable to label noise. Our method sets the new state of the art on various types of label noise and achieves excellent performance on large-scale datasets with real-world label noise. For instance, on CIFAR100 with a 40% uniform noise ratio and only 10 trusted labeled data per class, our method achieves 80.2% classification accuracy, where the error rate is only 1.4% higher than a neural network trained without label noise. Moreover, increasing the noise ratio to 80%, our method still maintains a high accuracy of 75.5%, compared to the previous best accuracy 48.2%.
135 citations
14 Jun 2020
TL;DR: A highly accurate noise formation model based on the characteristics of CMOS photosensors is presented, thereby enabling us to synthesize realistic samples that better match the physics of image formation process.
Abstract: Lacking rich and realistic data, learned single image denoising algorithms generalize poorly in real raw images that not resemble the data used for training. Although the problem can be alleviated by the heteroscedastic Gaussian noise model, the noise sources caused by digital camera electronics are still largely overlooked, despite their significant effect on raw measurement, especially under extremely low-light condition. To address this issue, we present a highly accurate noise formation model based on the characteristics of CMOS photosensors, thereby enabling us to synthesize realistic samples that better match the physics of image formation process. Given the proposed noise model, we additionally propose a method to calibrate the noise parameters for available modern digital cameras, which is simple and reproducible for any new device. We systematically study the generalizability of a neural network trained with existing schemes, by introducing a new low-light denoising dataset that covers many modern digital cameras from diverse brands. Extensive empirical results collectively show that by utilizing our proposed noise formation model, a network can reach the capability as if it had been trained with rich real data, which demonstrates the effectiveness of our noise formation model.
129 citations
TL;DR: A novel “Noisy-As-Clean” (NAC) strategy of training self-supervised denoising networks, where the corrupted test image is directly taken as the “clean” target, while the inputs are synthetic images consisted of this corrupted image and a second yet similar corruption.
Abstract: Supervised deep networks have achieved promising performance on image denoising, by learning image priors and noise statistics on plenty pairs of noisy and clean images. Unsupervised denoising networks are trained with only noisy images. However, for an unseen corrupted image, both supervised and unsupervised networks ignore either its particular image prior, the noise statistics, or both. That is, the networks learned from external images inherently suffer from a domain gap problem: the image priors and noise statistics are very different between the training and test images. This problem becomes more clear when dealing with the signal dependent realistic noise. To circumvent this problem, in this work, we propose a novel “Noisy-As-Clean” (NAC) strategy of training self-supervised denoising networks. Specifically, the corrupted test image is directly taken as the “clean” target, while the inputs are synthetic images consisted of this corrupted image and a second yet similar corruption. A simple but useful observation on our NAC is: as long as the noise is weak, it is feasible to learn a self-supervised network only with the corrupted image, approximating the optimal parameters of a supervised network learned with pairs of noisy and clean images. Experiments on synthetic and realistic noise removal demonstrate that, the DnCNN and ResNet networks trained with our self-supervised NAC strategy achieve comparable or better performance than the original ones and previous supervised/unsupervised/self-supervised networks. The code is publicly available at https://github.com/csjunxu/Noisy-As-Clean .
109 citations
TL;DR: A novel trilateration algorithm for indoor localization based on received signal strength indication (RSSI) based on the extreme value theory, which constructs a nonlinear error function depending on distances and anchor nodes position is proposed.
Abstract: This paper proposed a novel trilateration algorithm for indoor localization based on received signal strength indication (RSSI). Firstly, all the raw measurement data are preprocessed by a Gaussian filter to reducing the influence of measurement noise. Secondly, the transmit power and the path loss exponent are estimated by a novel least-squares curve fitting (LSCF) method in the RSSI-based localization. Thirdly, a novel trilateration algorithm is proposed based on the extreme value theory, which constructs a nonlinear error function depending on distances and anchor nodes position. To minimize the function, a Taylor series approximation can be used for reduce the computational complexity. And, an iteration condition is designed to further improve the positioning accuracy. Afterward, Bayesian filtering is used to smoothing the localization error, and decrease the influence of the process noise. Both the simulation and experimental results demonstrate the effectiveness of the proposed methodology.
95 citations
TL;DR: The proposed noise PSD tracker, called DeepMMSE makes no assumptions about the characteristics of the noise or the speech, exhibits no tracking delay, and produces an accurate estimate that requires no bias correction, and when employed in a speech enhancement framework is able to outperform state-of-the-art noise PSd trackers, as well as multiple deep learning approaches to speech enhancement.
Abstract: An accurate noise power spectral density (PSD) tracker is an indispensable component of a single-channel speech enhancement system. Bayesian-motivated minimum mean-square error (MMSE)-based noise PSD estimators have been the most prominent in recent time. However, they lack the ability to track highly non-stationary noise sources due to current methods of a priori signal-to-noise (SNR) estimation. This is caused by the underlying assumption that the noise signal changes at a slower rate than the speech signal. As a result, MMSE-based noise PSD trackers exhibit a large tracking delay and produce noise PSD estimates that require bias compensation. Motivated by this, we propose an MMSE-based noise PSD tracker that employs a temporal convolutional network (TCN) a priori SNR estimator. The proposed noise PSD tracker, called DeepMMSE makes no assumptions about the characteristics of the noise or the speech, exhibits no tracking delay, and produces an accurate estimate that requires no bias correction. Our extensive experimental investigation shows that the proposed DeepMMSE method outperforms state-of-the-art noise PSD trackers and demonstrates the ability to track abrupt changes in the noise level. Furthermore, when employed in a speech enhancement framework, the proposed DeepMMSE method is able to outperform state-of-the-art noise PSD trackers, as well as multiple deep learning approaches to speech enhancement. Availability: DeepMMSE is available at: https://github.com/anicolson/DeepXi .
88 citations
TL;DR: This article proposes a novel fixed-time convergent nonsmooth backstepping control scheme for FAHV via augmented sliding mode observers (ASMOs) to overcome these obstacles of uncertainty and measurement noise.
Abstract: Uncertainty and measurement noise are main obstacles that limit the tracking control performances of flexible air-breathing hypersonic vehicles (FAHVs). In this article, we propose a novel fixed-time convergent nonsmooth backstepping control scheme for FAHV via augmented sliding mode observers (ASMOs) to overcome these obstacles. The ASMOs are first designed for the FAHV dynamics by employing the measured states corrupted by noises as inputs. On one hand, the ASMOs can simultaneously estimate the uncertainties and filter out the measurement noises. On the other hand, the observation error of each ASMO can be convergent within a fixed time independent of its initial observation error. Then, based on the estimation results, the altitude and velocity tracking controllers are developed by using fixed-time nonsmooth backstepping technique. Afterwards, a Lyapunov-based stability analysis is given to illustrate the fixed-time convergence of the closed-loop signals of the FAHV control system. Finally, comparative simulations are conducted to illustrate the superiority of the proposed control scheme.
70 citations
TL;DR: This article proposes an adaptive linear active disturbance rejection control (LADRC) controller to achieve strong antidisturbance performance and reduce noise sensitivity for EMAs, and proposes a novel parallel structure to improve dynamic responses.
Abstract: Electromechanical actuator (EMA) exhibits advanced performance in industry, but its dynamic servo responses are constrained by parametric perturbations, load torque variations, and measurement noise. A strong disturbance rejection ability is necessary for EMAs. However, this usually makes them more sensitive to the measurement noise, reducing the steady-state precision. In this article, an adaptive linear active disturbance rejection control (LADRC) controller is proposed to achieve strong antidisturbance performance and reduce noise sensitivity for EMAs. A novel parallel structure is proposed to improve dynamic responses, which replaces the traditional cascade structure of position and speed loops. Aiming to improve the antidisturbance performance, a linear full-order-extended state observer is integrated with the parallel controller, called the LADRC controller. To reduce the difficulty of parameter tuning, the number of tuning parameters of LADRC is reduced to two by a pole placement design. And these two parameters of LADRC can be adjusted adaptively by the hyperbolic tangent function. Finally, the simulation and experimental results are provided to verify the effectiveness of the proposed strategy for EMAs.
70 citations
14 Jun 2020
TL;DR: An effective automatic label noise cleansing framework for face recognition datasets, FaceGraph, which performs global-to-local discrimination to select useful data in a noisy environment and surpasses state-of-the-art performance on the IJB-C benchmark.
Abstract: In the field of face recognition, large-scale web-collected datasets are essential for learning discriminative representations, but they suffer from noisy identity labels, such as outliers and label flips. It is beneficial to automatically cleanse their label noise for improving recognition accuracy. Unfortunately, existing cleansing methods cannot accurately identify noise in the wild. To solve this problem, we propose an effective automatic label noise cleansing framework for face recognition datasets, FaceGraph. Using two cascaded graph convolutional networks, FaceGraph performs global-to-local discrimination to select useful data in a noisy environment. Extensive experiments show that cleansing widely used datasets, such as CASIA-WebFace, VGGFace2, MegaFace2, and MS-Celeb-1M, using the proposed method can improve the recognition performance of state-of-the-art representation learning methods like Arcface. Further, we cleanse massive self-collected celebrity data, namely MillionCelebs, to provide 18.8M images of 636K identities. Training with the new data, Arcface surpasses state-of-the-art performance by a notable margin to reach 95.62% TPR at 1e-5 FPR on the IJB-C benchmark.
60 citations
14 Jun 2020
TL;DR: This work proposes a new loss correction approach, named as Meta Loss Correction (MLC), to directly learn T from data via the meta-learning framework, which is model-agnostic and learns T fromData rather than heuristically approximates it using prior knowledge.
Abstract: Label noise may significantly degrade the performance of Deep Neural Networks (DNNs). To train noise-robust DNNs, Loss correction (LC) approaches have been introduced. LC approaches assume the noisy labels are corrupted from clean (ground-truth) labels by an unknown noise transition matrix T. The backbone DNNs and T can be trained separately, where T is approximated with prior knowledge. For example, T is constructed by stacking the maximum or mean predic- tions of the samples from each class. In this work, we pro- pose a new loss correction approach, named as Meta Loss Correction (MLC), to directly learn T from data via the meta-learning framework. The MLC is model-agnostic and learns T from data rather than heuristically approximates it using prior knowledge. Extensive evaluations are conducted on computer vision (MNIST, CIFAR-10, CIFAR-100, Cloth- ing1M) and natural language processing (Twitter) datasets. The experimental results show that MLC achieves very com- petitive performance against state-of-the-art approaches.
TL;DR: Noise2Inverse is proposed, a deep CNN-based denoising method for linear image reconstruction algorithms that does not require any additional clean or noisy data and demonstrates an improvement in peak signal-to-noise ratio and structural similarity index compared to state-of-the-art image Denoising methods, and conventional reconstruction methods, such as Total-Variation Minimization.
Abstract: Recovering a high-quality image from noisy indirect measurements is an important problem with many applications. For such inverse problems, supervised deep convolutional neural network (CNN)-based denoising methods have shown strong results, but the success of these supervised methods critically depends on the availability of a high-quality training dataset of similar measurements. For image denoising, methods are available that enable training without a separate training dataset by assuming that the noise in two different pixels is uncorrelated. However, this assumption does not hold for inverse problems, resulting in artifacts in the denoised images produced by existing methods. Here, we propose Noise2Inverse, a deep CNN-based denoising method for linear image reconstruction algorithms that does not require any additional clean or noisy data. Training a CNN-based denoiser is enabled by exploiting the noise model to compute multiple statistically independent reconstructions. We develop a theoretical framework which shows that such training indeed obtains a denoising CNN, assuming the measured noise is element-wise independent, and zero-mean. On simulated CT datasets, Noise2Inverse demonstrates an improvement in peak signal-to-noise ratio and structural similarity index compared to state-of-the-art image denoising methods, and conventional reconstruction methods, such as Total-Variation Minimization. We also demonstrate that the method is able to significantly reduce noise in challenging real-world experimental datasets.
TL;DR: An optimal event-triggered fault detection observer (FDO) design criterion is proposed such that the generated residual is sensitive to system faults while robust against exogenous disturbance and measurement noise.
Abstract: This article is concerned with the event-triggered fault detection problem for discrete-time systems subject to unknown-but-bounded (UBB) process disturbance and measurement noise via zonotope-based residual evaluation. To save communication resources, a novel discrete-time dynamic event-triggered mechanism is proposed. An optimal event-triggered $l_1 /H_{\infty }$ fault detection observer (FDO) design criterion is proposed such that the generated residual is sensitive to system faults while robust against exogenous disturbance and measurement noise. On the basis of the designed FDO, a zonotope-based dynamic threshold for residual evaluation is well constructed by considering the impacts of disturbance, noise, and event-triggered communication. Finally, a vehicle lateral dynamic system is adopted to illustrate the effectiveness of the proposed zonotope-based dynamic event-triggered fault detection mechanism.
TL;DR: This article considers the problem of measurement noise rejection in a linear output-feedback control system and proposes a novel noise estimator (NE)-based robust control solution, which takes into account not only the rejection of high-frequency stochastic noises but also the compensation for low-frequency measurement errors, such as bias and drift.
Abstract: This article considers the problem of measurement noise rejection in a linear output-feedback control system. Specifically, we take into account not only the rejection of high-frequency stochastic noises but also the compensation for low-frequency measurement errors, such as bias and drift, which cannot be well-handled by the classic frequency-domain filters or Kalman filters. A novel noise estimator (NE)-based robust control solution is proposed. The NE is designed in the frequency domain by exploiting the system model and control structure information and is embedded into the controller instead of being an independent functional module in the closed-loop system. The adverse effects of model uncertainties on the performance of the NE-based solution are investigated, and an improved solution is proposed by incorporating a simple low-pass filter as the prefilter of NE. This solution is applied to the angle tracking problem of a 2-DOF experimental helicopter platform equipped with a low-cost and low-accuracy microelectromechanical system (MEMS) inertial measurement unit (IMU) (MEMS IMU) for angular position/rate measurements. Both numerical simulation and experimental comparisons with other existing approaches demonstrate: 1) constant bias and time-varying drift in the IMU measurements are systematically addressed by the solution; 2) it is accessible to improve the steady-state tracking accuracy by tuning the parameter of NE to extend its bandwidth; and 3) when model uncertainties limit the feasible bandwidth of NE, the improved solution is able to largely maintain its noise rejection performance.
TL;DR: A Boolean Bayesian filter is designed that can be utilized to provide the minimum MSE state estimate for the STVBNs and a recursive matrix-based algorithm is obtained to calculate the one-step prediction and estimation of the forward–backward state probability distribution vectors.
Abstract: In this article, a general theoretical framework is developed for the state estimation problem of stochastic time-varying Boolean networks (STVBNs). The STVBN consists of a system model describing the evolution of the Boolean states and a model relating the noisy measurements to the Boolean states. Both the process noise and the measurement noise are characterized by sequences of mutually independent Bernoulli distributed stochastic variables taking values of 1 or 0, which imply that the state/measurement variables may be flipped with certain probabilities. First, an algebraic representation of the STVBNs is derived based on the semitensor product. Then, based on Bayes’ theorem, a recursive matrix-based algorithm is obtained to calculate the one-step prediction and estimation of the forward–backward state probability distribution vectors. Owing to the Boolean nature of the state variables, the Boolean Bayesian filter is designed that can be utilized to provide the minimum MSE state estimate for the STVBNs. The fixed-interval smoothing filter is also obtained by resorting to the forward–backward technique. Finally, a simulation experiment is carried out for the context estimation problem of the $p53$ - $MDM2$ negative-feedback gene regulatory network.
TL;DR: This brief proposes an improved variable kernel width MCC algorithm, which is derived by minimizing the squared deviation at each iteration, and designs a reset mechanism for the proposed algorithm to improve its tracking capability when the estimated vector encounters a sudden change.
Abstract: The maximum correntropy criterion (MCC) algorithm has attracted much attention due to its capability of combating impulsive noise. However, its performance depends on choice of the kernel width, which is a hard issue. Several variable kernel width schemes based on various error functions have been proposed to address this problem. Nevertheless, these methods may not provide an optimal kernel width because they do not contain any knowledge of the background noise that actually has influence on the optimization of the kernel width. This brief proposes an improved variable kernel width MCC algorithm, which is derived by minimizing the squared deviation at each iteration. We also design a reset mechanism for the proposed algorithm to improve its tracking capability when the estimated vector encounters a sudden change. Simulations for system identification and echo cancellation scenarios show that the proposed scheme outperforms other variable kernel width algorithms.
TL;DR: This paper proposes a high power-performance-area efficient background noise aware keyword-spotting (KWS) processor based on an optimized binarized weight network (BWN) processor with adaptively configured to use dual computing modes for both high recognition accuracy under high background noise and ultra-low power consumption under low background noise.
Abstract: This paper proposes a high power-performance-area efficient background noise aware keyword-spotting (KWS) processor based on an optimized binarized weight network (BWN). To reduce the power consumption while maintaining the system recognition accuracy for different background noise, the KWS processor with a SNR prediction module can be adaptively configured to use dual computing modes (standard computing mode and approximate computing mode) for both high recognition accuracy under high background noise and ultra-low power consumption under low background noise. The mel-scale frequency cepstral coefficients (MFCC) module is optimized with approximate computing technologies, which can reduce the power consumption by up to $3.1\times $ and $5.7\times $ for high/low background noise, respectively. Based on the evaluation of the architecture design space exploration, an ultra-low power BWN accelerator with low voltage, area and leakage power and using precision self-adaptive approximate computing units was proposed. Evaluated under 22nm process technology, this work can support up to 10 keywords real time recognition with power consumption of $15.1~\mu \text{W}$ for high background noise and $10.8~\mu \text{W}$ for low background noise. Compared to the state-of-the-art KWS architectures, our work can achieve ultra-low power consumption (about $1.7\times $ reduced), while maintaining high system capability and adaptability.
TL;DR: This study addresses joint speech separation and dereverberation, which aims to separate target speech from background noise, interfering speech and room reverberation, and proposes a novel multimodal network that exploits both audio and visual signals.
Abstract: Background noise, interfering speech and room reverberation frequently distort target speech in real listening environments. In this study, we address joint speech separation and dereverberation, which aims to separate target speech from background noise, interfering speech and room reverberation. In order to tackle this fundamentally difficult problem, we propose a novel multimodal network that exploits both audio and visual signals. The proposed network architecture adopts a two-stage strategy, where a separation module is employed to attenuate background noise and interfering speech in the first stage and a dereverberation module to suppress room reverberation in the second stage. The two modules are first trained separately, and then integrated for joint training, which is based on a new multi-objective loss function. Our experimental results show that the proposed multimodal network yields consistently better objective intelligibility and perceptual quality than several one-stage and two-stage baselines. We find that our network achieves a 21.10% improvement in ESTOI and a 0.79 improvement in PESQ over the unprocessed mixtures. Moreover, our network architecture does not require the knowledge of the number of speakers.
TL;DR: This method uses traditionally singular value transform [singular value decomposition (SVD)] to reconstruct narrowband interference and remove it and the PD signal is obtained by time-domain denoising.
Abstract: Online partial discharge (PD) monitoring is an important means to detect insulation deterioration. However, it is difficult to extract the PD signal due to various interferences in the field. Noisy PD signal is used to judge the status of insulation, which would affect the conclusion; therefore, denoising PD signal is a major task in online PD monitoring. Common methods for PD denoising include the empirical mode decomposition (EMD) and wavelet transform; however, the denoising results are highly dependent on the modal aliasing, the selection of mother wavelets, and decomposition levels. This article proposes a method to solve these problems. This method uses traditionally singular value transform [singular value decomposition (SVD)] to reconstruct narrowband interference and remove it. Next, the empirical wavelet transform (EWT) is carried out for the PD signal that has residual white noise. Then, the noisy signal is decomposed into several modes corresponding to each spectrum segment. The $3~\sigma $ principle is used to denoise the modes with large kurtosis, and the modes are combined into a reference signal. The start-end positions of PD signal are then obtained from the reference signal. Finally, the PD signal is obtained by time-domain denoising. The results from both simulated and actual field detection signals show the excellent performance of this method.
TL;DR: The influence of a stationary white noise on FSF-based frequency estimation of the power system is investigated and the variance expression of the frequency estimator is derived theoretically and compared to its unbiased Cramer–Rao lower bound (CRLB).
Abstract: The frequency shifting and filtering (FSF) algorithm, a variant of DFT, has the merit of high efficiency for frequency analysis thanks to its simple implementation in the time domain. However, the inevitable white noise injected by various factors leads to inaccurate frequency estimation in practical measurement. This article investigates the influence of a stationary white noise on FSF-based frequency estimation of the power system. The variance expression of the frequency estimator is derived theoretically and compared to its unbiased Cramer–Rao lower bound (CRLB). The obtained results are validated by several computer simulations.
TL;DR: A slide window variational adaptive Kalman filter is presented in this brief based on adaptive learning of inaccurate state and measurement noise covariance matrices, which is composed of the forward Kalman filtering, the backward Kalman smoothing, and the online estimates of noise covariant matrices.
Abstract: A slide window variational adaptive Kalman filter is presented in this brief based on adaptive learning of inaccurate state and measurement noise covariance matrices, which is composed of the forward Kalman filtering, the backward Kalman smoothing, and the online estimates of noise covariance matrices. By imposing an approximation on the smoothing posterior distribution of slide window state vectors, the posterior distributions of noise covariance matrices can be analytically updated as inverse Wishart distributions by exploiting the variational Bayesian method, which avoids the fixed-point iterations and achieves good computational efficiency. Simulation comparisons demonstrate that the proposed method has better filtering accuracy and consistency than the existing cutting-edge method.
TL;DR: To approach real desert seismic data, a variety of seismic wavelets are used to simulate different types of seismic events, and then these synthetic seismic events and real desert low-frequency noise to construct training set are used.
Abstract: Lots of low-frequency noise including random noise and surface waves seriously reduces the quality of desert seismic data. However, the suppression for desert low-frequency noise faces three main problems: nonstationary and non-Gaussian of random noise; strong energy of low-frequency noise; a more serious frequency-band overlap between effective signals and low-frequency noise. Robust principal component analysis (RPCA) is a classical low-rank matrix (LM) recovery method which is very suitable for processing nonlinear noise. It can decompose noisy data to the optimal LM and sparse matrix (SM), which include most effective signals and noise, respectively. Therefore, the RPCA is introduced to suppress desert low-frequency noise. However, due to the low signal-to-noise ratio (SNR) and serious frequency-band overlap, much low-frequency noise still remains in the LM of desert seismic data after the decomposition of RPCA. Meanwhile, some nonnegligible effective signals are decomposed into the SM of desert seismic data. To solve this problem, the convolutional neural network (CNN) is introduced to extract effective signals from SM and LM. By constructing suitable training sets to guide the CNN’s training, the CNN denoising models after training are used to predict the effective signals from these two matrices, respectively. In this article, to approach real desert seismic data, we use a variety of seismic wavelets to simulate different types of seismic events, and then use these synthetic seismic events and real desert low-frequency noise to construct training set. In experiments, our method can raise the SNR of synthetic noisy data from −8.69 to 9.63 dB.
TL;DR: This article develops a conditional VAE (CVAE) where the audio speech generative process is conditioned on visual information of the lip region, and it improves the speech enhancement performance compared with the audio-only VAE model.
Abstract: Variational auto-encoders (VAEs) are deep generative latent variable models that can be used for learning the distribution of complex data. VAEs have been successfully used to learn a probabilistic prior over speech signals, which is then used to perform speech enhancement. One advantage of this generative approach is that it does not require pairs of clean and noisy speech signals at training. In this article, we propose audio-visual variants of VAEs for single-channel and speaker-independent speech enhancement. We develop a conditional VAE (CVAE) where the audio speech generative process is conditioned on visual information of the lip region. At test time, the audio-visual speech generative model is combined with a noise model based on nonnegative matrix factorization, and speech enhancement relies on a Monte Carlo expectation-maximization algorithm. Experiments are conducted with the recently published NTCD-TIMIT dataset as well as the GRID corpus. The results confirm that the proposed audio-visual CVAE effectively fuses audio and visual information, and it improves the speech enhancement performance compared with the audio-only VAE model, especially when the speech signal is highly corrupted by noise. We also show that the proposed unsupervised audio-visual speech enhancement approach outperforms a state-of-the-art supervised deep learning method.
15 Jul 2020
TL;DR: In this article, a low-power dynamic comparator for ultralow power applications is presented, which uses cross-coupled devices to prevent the comparator internal nodes from fully discharging to ground in contrast to the conventional architecture.
Abstract: This letter presents a low-power dynamic comparator for ultralow power applications. The prototype is designed in a 65-nm CMOS process with a supply voltage of 1 V and is compared against the widely used double tail latch comparator in terms of power consumption and input referred rms noise. The addition of cross-coupled devices to the input differential pair prevents the comparator internal nodes from fully discharging to ground in contrast to the conventional architecture. This reduces the power consumption while achieving similar noise levels. Measurements demonstrate that the proposed comparator achieves an input referred rms noise voltage of 220 $\mu \text{V}$ against 210 $\mu \text{V}$ for the conventional comparator with a 30% reduction in power. The proposed circuit consumes 0.19-pJ energy per comparison.
TL;DR: The experimental results show that the proposed COM-W method is a closed-form and computationally efficient method that has an excellent positioning performance under both narrow and square geometric arrangements, specifically in the central cross-shaped area and in the vicinity of the nodes.
Abstract: A novel and practical combined weighted (COM-W) method for target localization based on time difference of arrival (TDOA) is proposed in this paper. For 2-D situations with more than three nodes, the main idea of the proposed method is to obtain multiple preliminary localization results corresponding to the different three-node combinations first, and then use the weighted averaging technique based on the Cramer–Rao lower bound (CRLB) to obtain the final estimation. Considering target localization with a set of fixed-positioning nodes employing TDOA measurements as an example, the performance of the proposed method is evaluated by both simulations and acoustic indoor positioning experiments. The experimental results show that the COM-W method is a closed-form and computationally efficient method. It has an excellent positioning performance under both narrow and square geometric arrangements, specifically in the central cross-shaped area and in the vicinity of the nodes. Compared to the three classic weighted linear least squares (WLS)-based positioning methods, the proposed COM-W method is less sensitive to the measurement noise and the NLOS environment. The proposed method has an important application value for the realization of a TDOA-based indoor target positioning system.
TL;DR: The discrete-time version of Levant's arbitrary order robust exact differentiator is extended by taking into account also nonlinear higher order terms, which preserves the asymptotic accuracies with respect to sampling and noise known from the continuous-time algorithm.
Abstract: The discrete-time version of Levant's arbitrary order robust exact differentiator, which is a forward Euler discretized version of the continuous-time algorithm enhanced by linear higher order terms, is extended by taking into account also nonlinear higher order terms. The resulting differentiator preserves the asymptotic accuracies with respect to sampling and noise known from the continuous-time algorithm. It is demonstrated in a simulation example and by differentiating a measured signal that the nonlinear higher order terms allow reducing the high-frequency switching amplitude whenever the $(n+1)$ th derivative of the signal to be differentiated vanishes, leading to an improvement in the precision.
TL;DR: A novel parallel-data-free speech enhancement method, in which the cycle-consistent generative adversarial network (CycleGAN) and multi-objective learning are employed, which is effective to improve speech quality and intelligibility when the networks are trained under the parallel data.
Abstract: Recently, deep neural networks (DNNs) have become the mainstream strategy for speech enhancement task because it can achieve the higher speech quality and intelligibility than the traditional methods. However, these DNN-based methods always need a large number of parallel corpus consisting of clean speech and noise to produce noisy data for the training of the DNN in order to improve the generalization of the network. As a result, this implies that many noisy speech signals that are collected in real environment cannot be used to train the DNN because of the lack of corresponding clean speech and noise. Additionally, as we know, noise varies with the time and scenario, so we cannot obtain parallel speech and noise due to infinite noise data and some limited speech data. Thus, the network training with unparallel speech and noise data is essential for the generalization of the network. To address this problem, we propose a novel parallel-data-free speech enhancement method, in which the cycle-consistent generative adversarial network (CycleGAN) and multi-objective learning are employed. Our method is also able to make best use of the benefits of multi-objective learning. On the training stage, we utilize two different encoders to encode the features of clean speech and noisy speech, respectively. Then, two forward generators are immediately used to predict the ideal time-frequency (T-F) mask and log-power spectrum (LPS) of clean speech. Two inverse generators are applied to map the magnitude spectrum (MS) and LPS of noisy speech, respectively. In addition, four discriminators are used to distinguish the real speech features from the generated features. Two encoders, four generators and four discriminators are simultaneously trained by using adversarial, identity-mapping, latent similarity and cycle-consistent loss. On the test stage, we directly utilize the forward generators and encoders to acquire the enhanced speech. The experimental results indicate that the proposed approach is able to achieve the better speech enhancement performance than the reference methods. Moreover, the proposed method is also effective to improve speech quality and intelligibility when the networks are trained under the parallel data.
TL;DR: Results show that the proposed filter significantly improves edges over exiting literature, and Peak Signal to Noise Ratio was improved by 1.2 dB in de-noising of medical images corrupted by medium to high noise densities.
Abstract: Image corruption is a common phenomenon which occurs due to electromagnetic interference, and electric signal instabilities in a system. In this letter, a novel multi procedure Min-Max Average Pooling based Filter is proposed for removal of salt, and pepper noise that betide during transmission. The first procedure functions as a pre-processing step that activates for images with low noise corruption. In latter procedure, the noisy image is divided into two instances, and passed through multiple layers of max, and min pooling which allow restoration of intensity transitions in an image. The final procedure recombines the parallel processed images from the previous procedures, and performs average pooling to remove all residual noise. Experimental results were obtained using MATLAB software, and show that the proposed filter significantly improves edges over exiting literature. Moreover, Peak Signal to Noise Ratio was improved by 1.2 dB in de-noising of medical images corrupted by medium to high noise densities.
TL;DR: An extended state observer is employed to estimate the accurate state information of each DG, which is significantly influenced by measurement noise, and the distributed controllers based on a fast terminal sliding mode surface and an adaptive super-twisting algorithm are designed to track the voltage reference and to fasten the convergence rate against disturbances and uncertainties caused by parameter perturbation.
Abstract: This paper proposes an extended-state-observer-based distributed robust secondary voltage and frequency control for an autonomous microgrid (MG) with inverter-based distributed generators (DGs) considering the uncertainties from models and measurement noise. The MG is considered as a multi-agent system where each DG is defined as an agent and its controller only requires its own information and the information of its neighbors, but each DG obtains noisy measurements of the states of itself and its neighbors easily due to stochastic noise. Therefore, in this paper, an extended state observer is employed to estimate the accurate state information of each DG, which is significantly influenced by measurement noise. Furthermore, the distributed controllers based on a fast terminal sliding mode surface and an adaptive super-twisting algorithm are designed to track the voltage reference and to fasten the convergence rate against disturbances and uncertainties caused by parameter perturbation. Moreover, the distributed frequency controllers are also designed to restore the frequency and to guarantee the accurate active power sharing without power information of DGs. Finally, the effectiveness of the propose control strategy is illustrated by the simulation of an autonomous MG in MATLAB/Simulink.
TL;DR: In this paper, the authors study the noisy version of the group testing problem, where the outcome of each standard noiseless group test is subject to independent noise, corresponding to passing the result through a binary channel.
Abstract: The group testing problem consists of determining a small set of defective items from a larger set of items based on a number of possibly-noisy tests, and is relevant in applications such as medical testing, communication protocols, pattern matching, and more. We study the noisy version of this problem, where the outcome of each standard noiseless group test is subject to independent noise, corresponding to passing the noiseless result through a binary channel. We introduce a class of algorithms that we refer to as Near-Definite Defectives (NDD), and study bounds on the required number of tests for asymptotically vanishing error probability under Bernoulli random test designs. In addition, we study algorithm-independent converse results, giving lower bounds on the required number of tests under Bernoulli test designs. Under reverse Z-channel noise, the achievable rates and converse results match in a broad range of sparsity regimes, and under Z-channel noise, the two match in a narrower range of dense/low-noise regimes. We observe that although these two channels have the same Shannon capacity when viewed as a communication channel, they can behave quite differently when it comes to group testing. Finally, we extend our analysis of these noise models to a general binary noise model (including symmetric noise), and show improvements over known existing bounds in broad scaling regimes.