Showing papers on "Noise reduction published in 2000"

Fourth-order partial differential equations for noise removal

Abstract: A class of fourth-order partial differential equations (PDEs) are proposed to optimize the trade-off between noise removal and edge preservation. The time evolution of these PDEs seeks to minimize a cost functional which is an increasing function of the absolute value of the Laplacian of the image intensity function. Since the Laplacian of an image at a pixel is zero if the image is planar in its neighborhood, these PDEs attempt to remove noise and preserve edges by approximating an observed image with a piecewise planar image. Piecewise planar images look more natural than step images which anisotropic diffusion (second order PDEs) uses to approximate an observed image. So the proposed PDEs are able to avoid the blocky effects widely seen in images processed by anisotropic diffusion, while achieving the degree of noise removal and edge preservation comparable to anisotropic diffusion. Although both approaches seem to be comparable in removing speckles in the observed images, speckles are more visible in images processed by the proposed PDEs, because piecewise planar images are less likely to mask speckles than step images and anisotropic diffusion tends to generate multiple false edges. Speckles can be easily removed by simple algorithms such as the one presented in this paper.

Noise reduction in 3D perfusion imaging by attenuating the static signal in arterial spin tagging (ASSIST).

Abstract: Phase-encoded multishot SPIRAL approaches were used to acquire true 3D cerebral blood flow images of the human head using arterial spin tagging approaches. Multiple-inversion background suppression techniques, which suppress phase noise due to interacquisition fluctuations in the static magnetic field, reduced the temporal standard deviation of true 3D delta M images acquired using arterial spin tagging approaches by approximately 50%. Background suppressed arterial spin tagging (ASSIST) approaches were used to obtain high-resolution isotropic true 3D cerebral blood flow images, and to obtain true 3D activation images during cognitive (working memory) tasks. Magn Reson Med 44:92-100, 2000. Published 2000 Wiley-Liss, Inc.

Large-vocabulary speech recognition under adverse acoustic environments.

Abstract: We report our recent work on noise-robust large-vocabulary speech recognition. Three key innovations are developed and evaluated in this work: 1) a new model learning paradigm that comprises a noise-insertion process followed by noise reduction; 2) a noise adaptive training algorithm that integrates noise reduction into probabilistic multi-style system training; and 3) a new algorithm (SPLICE) for noise reduction that makes no assumptions about noise stationarity. Evaluation on a large-vocabulary speech recognition task demonstrates significant and consistent error rate reduction using these techniques. The resulting error rate is shown to be lower than that achieved by the matched-noisy condition for both stationary and nonstationary natural, as well as simulated, noises.

Reducing MOSFET 1/f noise and power consumption by switched biasing

Abstract: Switched biasing is proposed as a technique for reducing the 1/f noise in MOSFET's. Conventional techniques, such as chopping or correlated double sampling, reduce the effect of 1/f noise in electronic circuits, whereas the switched biasing technique reduces the 1/f noise itself. Whereas noise reduction techniques generally lead to more power consumption, switched biasing can reduce the power consumption. It exploits an intriguing physical effect: cycling a MOS transistor from strong inversion to accumulation reduces its intrinsic 1/f noise. As the 1/f noise is reduced at its physical roots, high frequency circuits, in which 1/f noise is being upconverted, can also benefit. This is demonstrated by applying switched biasing in a 0.8 /spl mu/m CMOS sawtooth oscillator. By periodically switching off the bias currents, during time intervals that they are not contributing to the circuit operation, a reduction of the 1/f noise induced phase noise by more than 8 dB is achieved, while the power consumption is also reduced by 30%.

Quantile based noise estimation for spectral subtraction and Wiener filtering

Abstract: Elimination of additive noise from a speech signal is a fundamental problem in audio signal processing. In this paper we restrict our considerations to the case where only a single microphone recording of the noisy signal is available. The algorithms which we investigate proceed in two steps. First, the noise power spectrum is estimated. A method based on temporal quantiles in the power spectral domain is proposed and compared with pause detection and recursive averaging. The second step is to eliminate the estimated noise from the observed signal by spectral subtraction or Wiener filtering. The database used in the experiments comprises 6034 utterances of German digits and digit strings by 770 speakers in 10 different cars. Without noise reduction, we obtain an error rate of 11.7%. Quantile based noise estimation and Wiener filtering reduce the error rate to 8.6%. Similar improvements are achieved in an experiment with artificial, non-stationary noise.

Subsonic Jet Noise from Nonaxisymmetric and Tabbed Nozzles

Abstract: Subsonic jet noise from nonaxisymmetric and tabbed nozzles are investigated experimentally and theoretically. It is shown that the noise spectra of these jets are in good agreement with the similarity spectra found empirically earlier by Tam et al. through a detailed analysis of supersonic jet noise data (Tam, C. K. W., Golebiowski, M., and Seiner, J. M., On the Two Components of Turbulent Mixing Noise from Supersonic Jets, AIAA Paper 96-1716, 1996). Furthermore, the radiated noise fields of the jets under study, including elliptic and large-aspect-ratio rectangular jets, are found to be quite axisymmetric and are practically the same as that of a circular jet with the same exit area. These experimental results strongly suggest that nozzle geometry modification into elliptic or rectangular shapes is not an effective method for jet noise suppression. A lobed nozzle, on the other hand, is found to impact significantly the noise field. Noise from large-scale turbulent structures, radiating principally in the downstream direction, is effectively suppressed. Tabs also impact the noise field, primarily by shifting the spectral peak to a higher frequency. A jetlets model is developed to provide a basic understanding of the noise from tabbed jets. The model predicts that the noise spectrum from a jet with N tabs (N ≥ 2) can he obtained from that of the original jet (no tab) by a simple frequency shift. The shifted frequency is obtained by multiplying the original frequency by N 1/2 . This result is in fairly good agreement with experimental data.

Sound-level measurements and calculations of safe noise dosage during EPI at 3 T.

Abstract: This paper describes systematic methods for measuring and controlling sound levels within a magnetic resonance scanner. The methods are illustrated by application to the acoustic noise generated by a 3 T scanner during echoplanar imaging (EPI). Across five measurement sessions, sound pressure levels at the center of the head gradient coil ranged from 122 to 131 dB SPL [123 to 132 dB(A)]. For protection against damaging noise exposure, UK and US industrial guidelines stipulate that the maximum permitted daily noise dosage is equivalent to 90 dB(A) for 8 hours, where noise dosage is a function of the level of an acoustic signal and the length of exposure to it. Without hearing protection, this equivalent level would be exceeded by less than 5 seconds of exposure to the measured levels of scanner acoustic noise. These findings highlight the importance of noise reduction and hearing protection for those exposed to the acoustic noise generated during EPI. J. Magn. Reson. Imaging 2000;12:157–163. © 2000 Wiley-Liss, Inc.

Spatially adaptive image denoising under overcomplete expansion

Abstract: This paper presents a novel wavelet-based image denoising algorithm under overcomplete expansion. In order to optimize the denoising performance, we make a systematic study of both signal and noise characteristics under overcomplete expansion. High-band coefficients are viewed as the mixture of non-edge class and edge class observing different probability models. Based on improved statistical modeling of wavelet coefficients, we derive optimal MMSE estimation strategies to suppress noise for both non-edge and edge coefficients. We have achieved fairly better objective performance than most recently-published wavelet denoising schemes.

ClariNet: a noise analysis tool for deep submicron design

Abstract: Coupled noise analysis has become a critical issue for deep-submicron, high performance design. In this paper, we present, ClariNet, an industrial noise analysis tool, which was developed to efficiently analyze large, high performance processor designs. We present the overall approach and tool flow of ClariNet and discuss three critical large-processor design issues which have received limited discussion in the past. First, we present how the driver gates of a coupled interconnect network are represented with accurate linear models. Second, we show how to speed the analysis of large designs by using noise filters based on reduced interconnect representations and then pruning the nets coupled to a signal net. Third, we show how to incorporate logic and timing correlations into noise analysis to reduce its pessimism. We present the results from several industrial circuits, including a large high performance microprocessor design and a DSP design.

Noise reduction apparatus and method

Abstract: A multi-band spectral subtraction scheme is proposed, comprising a multi-band filter architecture, noise and signal power detection, and gain function for noise reduction. In one embodiment, the gain function for noise reduction consists of a gain scale function and a maximum attenuation function providing a predetermined amount of gain as a function of signal to noise ratio (“SNR”) and noise. In one embodiment, the gain scale function is a three-segment piecewise linear function, and the three piecewise linear sections of the gain scale function include a first section providing maximum expansion up to a first knee point for maximum noise reduction, a second section providing less expansion up to a second knee point for less noise reduction, and a third section providing minimum or no expansion for input signals with high SNR to minimize distortion. According to embodiments of the present invention, the maximum attenuation function can either be a constant or equal to the estimated noise envelope. The disclosed noise reduction techniques can be applied to a variety of speech communication systems, such as hearing aids, public address systems, teleconference systems, voice control systems, or speaker phones. When used in hearing aid applications, the noise reduction gain function according to aspects of the present invention is combined with the hearing loss compensation gain function inherent to hearing aid processing.

Field trials using a digital hearing aid with active noise reduction and dual-microphone directionality.

Abstract: In this study, we measured the effects of a digital hearing aid on speech recognition or reception in noise for two noise reduction concepts: active noise reduction by speech-sensitive processing (SSP) and improved directionality by a dual- or so-called twin-microphone system (TMS). This was conducted in a well-controlled clinical field trial in 16 hearing-aid users, using a single-blind crossover design. The hearing aid fitting was controlled by insertion gain measurements and measurements with loudness scaling. This study combined laboratory experiments with three consecutive field trials of 4 weeks each. We used performance measurements (speech recognition tests in background noise), paired comparisons, and self-report measurements (questionnaires). The speech recognition or reception tests were performed before and after each field trial, the paired comparisons were performed in weeks 4 and 12, and the questionnaires were administered after each field trial. For all subjects, results were obtained for three different settings: no noise reduction, SSP alone, and TMS alone. In the last week, we also performed speech recognition or reception tests in background noise with both noise reduction concepts combined. Three types of results have been reported: "objective" results from the critical signal to noise ratios for speech recognition or reception in different background noises for different settings and "subjective" results: paired comparisons and questionnaires. The subjective scores show the same trend as the objective scores. The effects of TMS were clearly positive, especially for the speech reception threshold tests and for the paired comparisons. The effects of SSP were much smaller but showed significant benefits with respect to aversiveness and speech perception or reception in noise for specific acoustical environments. There was no extra benefit from the combined effect of SSP and TMS relative to TMS alone.

A European Study on Landing Gear Airframe Noise Sources

Abstract: Flyover noise of modern aircraft in their approach configuration can be dominated by airframe noise. Relevant flow noise sources are high lift devices and landing gears. Therefore aeroacoustic tests were performed in the German Dutch Wind Tunnel employing full-scale landing gears of an A340 aircraft. Farfield noise characteristics were determined for both a nose- and a main landing gear, the latter in a 4-wheel and a 6-wheel mock-up configuration, at wind speeds ranging from 50 to 78 m/s. DNW's 3m acoustic mirror and a planar microphone array served for source localization and ranking. Excess tone noise was detected and originated from pin-cavities. Broadband noise levels were confirmed to increase with the 6th power of flow velocity. Noise directivities feature weak level maxima both in the forward and rear arc. Based on results from earlier testing on a smaller A320 gear the effects on noise of gear size and configuration were evaluated. By covering different gear components with streamlined fairings an overall noise reduction on the order of 3 dB was achieved.

Wavelet-based noise removal for biomechanical signals: a comparative study

Abstract: The purpose of this paper is to present wavelet-based noise removal (WBNR) techniques to remove noise from biomechanical acceleration signals obtained from numerical differentiation of displacement data. Manual and semiautomatic methods were used to determine thresholds for both orthogonal and biorthogonal filters. This study also compares the performance of WBNR approaches with four automatic conventional noise removal techniques used in biomechanics. The conclusion of this work is that WBNR techniques are very effective in removing noise from differentiated signals with sharp transients while leaving these transients intact. For biomechanical signals with certain characteristics, WBNR techniques perform better than conventional methods, as indicated by quantitative merit measures.

Image denoising using a local Gaussian scale mixture model in the wavelet domain

Abstract: The statistics of photographic images, when decomposed in a multiscale wavelet basis, exhibit striking non-Gaussian behaviors. The joint densities of clusters of wavelet coefficients are well-described as a Gaussian scale mixture: a jointly Gaussian vector multiplied by a hidden scaling variable. We develop a maximum likelihood solution for estimating the hidden variable from an observation of the cluster of coefficients contaminated by additive Gaussian noise. The estimated hidden variable is then used to estimate the original noise-free coefficients. We demonstrate the power of this model through numerical simulations of image denoising.

Hearing system beamformer

Abstract: The present invention, generally speaking, picks up a voice or other sound signal of interest and creates a higher voice-to-background-noise ratio in the output signal so that a user enjoys higher intelligibility of the voice signal. In particular, beamforming techniques are used to provide optimized signals to the user for further increasing the understanding of speech in noisy environments and for reducing user listening fatigue. In one embodiment, signal-to-noise performance is optimized even if some of the binaural cues are sacrificed. In this embodiment, an optimum mix ratio or weighting ratio is determined in accordance with the ratio of noise power in the binaural signals. Enhancement circuitry is easily implemented in either analog or digital form and is compatible with existing sound processing methods, e.g., noise reduction algorithms and compression/expansion processing. The sound enhancement approach is compatible with, and additive to, any microphone directionality or noise cancelling technology.

Extending Spectral Modeling Synthesis with Transient Modeling Synthesis

Abstract: Sinusoidal modeling has enjoyed a rich history in both speech and music applications, including sound transformations, compression, denoising, and auditory scene analysis. For such applications, the underlying signal model must efficiently capture salient audio features (Goodwin 1998). In this article, we present an accurate, efficient, and flexible three-part model for audio signals consisting of sines, transients, and noise by extending spectral modeling synthesis (SMS) (Serra and Smith 1990) with an explicit flexible transient model called transient-modeling synthesis (TMS). The sinusoidal transformation system (STS) (McAulay and Quatieri 1986) and SMS find the slowly varying sinusoidal components in a signal using spectral-peak-picking algorithms. Subtracting the synthesized sinusoids from the original signal creates a residual consisting of transients and noise (Serra 1989; George and Smith 1992). However, sinusoids do not model this residual well. Although it is possible to model transients and noise by a sum of sinusoidal signals (as with the Fourier transform), it is neither efficient, because transient and noisy signals require many sinusoids for their description, nor meaningful, because transients are short-lived signals, while the sinusoidal model uses sinusoids that are active on a much larger time scale. In the STS system (generally applied to speech), the transient + noise residual is often masked sufficiently to be ignored (McAulay and Quatieri 1986). In music applications, this residual is often important to the integrity of the signal. The SMS system extends the sinusoidal model by explicitly modeling the residual as slowly filtered white noise. Although this technique has been very successful, transients do not fit well into this model, because transients modeled as filtered noise lose sharpness in their attack and tend to sound dull. Because transients are

Enhancement of bowel sounds by wavelet-based filtering

Abstract: This paper evaluates the performance of an automatic method for structural decomposition, noise removal and enhancement of bowel sounds (BS), based on the wavelet transform. The proposed method combines multiresolution analysis with hard thresholding to compose a wavelet transform-based stationary-nonstationary (WTST-NST) filter, for enhanced separation of bowel sounds (BS) from superimposed noise. Quantitative and qualitative analysis of the experimental results, when applying the WTST-NST filter to BS recorded from controls and patients with gastrointestinal dysfunction, prove that the ability of the WTST-NST filter to remove noise and reveal the authentic structure of BS is excellent. By eliminating the need to record a noise reference signal, this method reduces hardware overhead when analysis of BS is the primary aim. The method is independent of subjective human judgement for selection of noise reference templates, is robust to different levels of signal interference, and, due to its simplicity, can easily be used in clinical medicine.

An algorithm for integrated noise reduction and sharpness enhancement

Abstract: Noise reduction and image sharpening are techniques to improve video image quality. However, noise filters tend to blur image detail, while filters for image sharpening tend to increase noise. So, cascading the two filters does not always give the best performance. We present an integrated filter that reduces noise and sharpens details in a noisy video signal depending on local image statistics. This allows both features to be maximally exploited.

Noise reduction method utilizing statistical weighting, apparatus, and program for digital image processing

Abstract: A method of processing a digital image channel to remove noise, includes the steps of: identifying a pixel of interest; identifying at least two sampled local regions of pixels which include the pixel of interest; calculating a noise free pixel estimate for each sampled local region of pixels; calculating a statistical weighting factor for each sampled local region, the calculation of the statistical weighting factor being independent of the calculation of the noise free pixel estimate; and using the noise free pixel estimates and the statistical weighting factors for calculate a noise reduced pixel value.

Pattern recognition training method and apparatus using inserted noise followed by noise reduction

Abstract: A method and apparatus for training and using a pattern recognition model are provided. Under the invention, additive noise that matches noise expected in a test signal is included in a training signal. The noisy training signal is passed through one or more noise reduction techniques to produce pseudo-clean training data. The pseudo-clean training data is used to train the pattern recognition model. When the test signal is received, it is passed through the same noise reduction techniques used on the noisy training signal. This produces pseudo-clean test data, which is applied to the pattern recognition model. Under one embodiment, sets of training data are produced with each set containing a different type of noise.

Radiation detector noise reduction method and radiation detector

Abstract: A radiation detector noise reduction method includes the step of detecting incident radiation with a radiation detecting section, the radiation detecting section having a plurality of pixels arrayed in the form of a matrix, reading out the detection signal from the radiation detecting section through a readouting section, and correcting the readout detection signal with a correction section, wherein the correction step includes the first sub-step of correcting the detection signal on the basis of a first correction value corresponding to noise originating in the radiation detecting section, and the second sub-step of correcting the detection signal on the basis of a second correction value corresponding to noise originating in the readouting section, the second sub-step being executed before the first sub-step

SNR enhancement of intensity noise-limited FOGs

Abstract: The signal-to-noise ratio (SNR) of fiber-optic gyroscopes (FOGs) is directly impacted by the optical power coupled to the interferometric circuit. The optical intensity noise produced by the optical broadband source becomes the major component of the total noise at the output when the optical power reaching the detector is more than a few tens of microwatts. This work shows the implementation of a signal processing scheme capable of promoting reduction of the source intensity noise around the proper frequency of birefringent FOG, where the optimum signal demodulation takes place, yielding a better SNR. The total SNR has been enhanced 6.6 dB by means of a 20.4 dB reduction in the noise component associated only with the source intensity noise.

Noise estimation for blocking artifacts reduction in DCT coded images

Abstract: This paper proposes a postprocessing algorithm that can reduce the blocking artifacts in discrete cosine transform (DCT) coded images. To analyze blocking artifacts as noise components residing across two neighboring blocks, we use 1-D pixel vectors made of pixel rows or columns across two neighboring blocks. We model the blocky noise in each pixel vector as a shape vector weighted by the boundary discontinuity. The boundary discontinuity of each vector is estimated from the difference between the pixel gradient across the block boundary and that of the internal pixels. We make minimum mean squared error (MMSE) estimates of the shape vectors, indexed by the local image activity, based on the noise statistics prior to postprocessing. Once the estimated shape vectors are stored in the decoder, the proposed algorithm eliminates the noise components by simply subtracting from each pixel vector an appropriate shape vector multiplied by the boundary discontinuity. The experimental results show that the proposed algorithm is highly effective in reducing blocking artifacts in both subjective and objective viewpoints, at low computational burden.

Simple Denoising Algorithm Using Wavelet Transform

Abstract: We have presented a new and alternative algorithm for noise reduction using the methods of discrete wavelet transform and numerical differentiation of the data. In our method the threshold for reducing noise comes out automatically. The algorithm has been applied to three model flow systems - Lorenz, Autocatalator, and Rossler systems - all evolving chaotically. The method is seen to work well for a wide range of noise strengths, even as large as 10% of the signal level. We have also applied the method successfully to noisy time series data obtained from the measurement of pressure fluctuations in a fluidized bed, and also to that obtained by conductivity measurement in a liquid surfactant experiment. In all the illustrations we have been able to observe that there is a clean separation in the frequencies covered by the differentiated signal and white noise.

Method and apparatus for noise compensation in imaging systems

Abstract: Methods and apparatus for performing channel dependent and gain dependent smoothing filter (across channels) to compensate for noise in a multislice imaging system are described. The smoothing mainly affects the radial resolution. The filter can be combined with the matrix deconvolution filter typically used in multi-slice scanners. The filter is channel and DAS gain dependent and provides that an image generated from data collected in a multi-slice scan has about the same image quality, e.g., noise reduction, as images generated by other types of scanners.

Noise reduction for broad-band, three-component seismograms using data-adaptive polarization filters

Abstract: Summary We develop a data-adaptive polarization filter that can spectacularly reduce microseismic noise contamination in three-component broad-band seismograms. The filter uses a multitaper spectral analysis method for computing the data spectral density matrix, which is defined as an ensemble average of outer products of the spectrum and its Hermitian adjoint. Under the assumption that strong noise in three-component, broad-band seismograms is additive white noise, and that its spectral density can be determined from seismogram segments without signals, that is, a pre-signal arrival time window, we construct a data-adaptive filter from a spectral density matrix that has been decontaminated of noise. Since the noise corrupting the seismograms is complicated and stochastic, the resulting residual due to the real, non-stationary nature of microseismic noise can leave small-amplitude, quasi-sinusoidal, background oscillations after filtering. These oscillations can be removed by subsequent application of an optimum Wiener filter. Application of the filter to synthetic data with real noise superimposed suppresses the noise by about three orders of magnitude at the expense of less than 5 per cent corruption of the original seismogram in amplitude. Application to several real recordings of teleseismic earthquakes on a three-component broad-band seismic station in Iceland shows that excellent signal-to-noise recovery is possible, rendering such data usable for both arrival time and waveform analysis. This technique may potentially increase by an order of magnitude the volume of usable data collected in seismic experiments in noisy environments, for example, on oceanic islands.

Subband noise reduction methods for speech enhancement

Abstract: Digital noise reduction processing is used in many telecommunications applications to enhance the quality of speech. This investigation focuses on the class of single-channel noise reduction methods employing the technique of short-time spectral modification, a class that includes the popular method of spectral subtraction. The simplicity and relative effectiveness of these subband noise reduction methods has resulted in explosive growth in their use for a variety of speech communications applications. The most commonly used forms of the short-time spectral modification method are discussed, including the Wiener filter, magnitude subtraction, power subtraction, and generalized parametric subtraction. Because of its importance to the subjective performance of any noise reduction method, the subject of real-time signal- and noise-level estimation is also reviewed. A low-complexity noise reduction algorithm is also presented and its implementation is discussed.