Showing papers on "Noise reduction published in 1991"
TL;DR: The center weighted median (CWM) filter as discussed by the authors is a weighted median filter that gives more weight only to the central value of each window, which can preserve image details while suppressing additive white and/or impulsive-type noise.
Abstract: The center weighted median (CWM) filter, which is a weighted median filter giving more weight only to the central value of each window, is studied. This filter can preserve image details while suppressing additive white and/or impulsive-type noise. The statistical properties of the CWM filter are analyzed. It is shown that the CWM filter can outperform the median filter. Some relationships between CWM and other median-type filters, such as the Winsorizing smoother and the multistage median filter, are derived. In an attempt to improve the performance of CWM filters, an adaptive CWM (ACWM) filter having a space varying central weight is proposed. It is shown that the ACWM filter is an excellent detail preserving smoother that can suppress signal-dependent noise as well as signal-independent noise. >
1,071 citations
TL;DR: A new method of filtering MR images is presented that uses wavelet transforms instead of Fourier transforms, which does not reduce the sharpness of edges but does eliminate any small structures that are similar in size to the noise eliminated.
Abstract: A new method of filtering MR images is presented that uses wavelet transforms instead of Fourier transforms. The new filtering method does not reduce the sharpness of edges. However, the new method does eliminate any small structures that are similar in size to the noise eliminated. There are many possible extensions of the filter.
292 citations
TL;DR: This work presents an optimal solution to the shadowing problem in the sense of least-mean-squares, which also provides an effective and convenient numerical method for noise reduction for data generated by a dynamical system.
219 citations
Patent•
24 Sep 1991TL;DR: In this article, a steerable pyramid architecture is used for image enhancement for the first time, with the steering being provided by the above orientation tuned filters, which is a set of orientation-tuned filters of a specialized design to permit steering, with analysis and synthesis filters also having a self-inverting characteristic.
Abstract: A technique is provided to remove noise from images and to enhance their visual appearance through the utilization of a technique which converts an image into a set of coefficients in a multi-scale image decomposition process, followed by modification of each coefficient based on its value and the value of coefficients of related orientation, position, or scale, which is in turn followed by a reconstruction or synthesis process to generate the enhanced image. Also contributing to the improved enhancement is a set of orientation tuned filters of a specialized design to permit steering, with the analysis and synthesis filters also having a self-inverting characteristic. Additionally, steerable pyramid architecture is used for image enhancement for the first time, with the steering being provided by the above orientation tuned filters. The utilization of related coefficients permits coefficient modification with multipliers derived through a statistical or neural-network analysis of coefficients derived through the utilization of clean and degraded images, with the modifiers corresponding to vectors which result in translating the degraded image coefficients into clean image coefficients, in essence by cancelling those portions of a coefficient due to noise. Further improvements include an overlay of classical coring on single coefficients. Thus, the subject technique provides improved image enhancement through the use of a multi-band or scale-oriented analysis and synthesis transform having improved coefficient modification, good orientation tuning, improved bandpass characteristics, and good spatial localization.
145 citations
TL;DR: A modification of the error-diffusion algorithm is shown that introduces an input-dependent threshold into the process and establishes a fast and efficient way to increase or decrease edge enhancement in the algorithm.
Abstract: Error diffusion is a popular algorithm for the binarization of continuous-tone images. A modification of the error-diffusion algorithm is shown that introduces an input-dependent threshold into the process. This modification establishes a fast and efficient way to increase or decrease edge enhancement in the algorithm.
137 citations
01 Aug 1991
TL;DR: In this article, aeroacoustic related problems are evaluated, and approaches to their solutions are suggested without extensive tables, nomographs, and derivations, focusing on underlying physical concepts.
Abstract: Methodology recommended to evaluate aeroacoustic related problems is provided, and approaches to their solutions are suggested without extensive tables, nomographs, and derivations. Orientation is toward flight vehicles and emphasis is on underlying physical concepts. Theoretical, experimental, and applied aspects are covered, including the main formulations and comparisons of theory and experiment. The topics covered include: propeller and propfan noise, rotor noise, turbomachinery noise, jet noise classical theory and experiments, noise from turbulent shear flows, jet noise generated by large-scale coherent motion, airframe noise, propulsive lift noise, combustion and core noise, and sonic booms.
129 citations
TL;DR: In this paper, a locally linear model is used to obtain a trajectory consistent with the dynamics as well as with the measured data, which leads to a significant improvement of correlation dimension estimates.
127 citations
TL;DR: A large increase in the reliability of shipboard or stationary underwater acoustic telemetry systems is achievable by using spatially distributed receivers with aperture sizes from 0.35 to 20 m, and a Viterbi error-correction decoder operating on each receiver is shown to perform reliability in a variety of non-Gaussian noise and jamming environments.
Abstract: A large increase in the reliability of shipboard or stationary underwater acoustic telemetry systems is achievable by using spatially distributed receivers with aperture sizes from 0.35 to 20 m. Output from each receiver is assigned a quality measure based on the estimated error rate, and the data, weighted by the quality measure, are combined and decoded. The quality measure is derived from a Viterbi error-correction decoder operating on each receiver and is shown to perform reliability in a variety of non-Gaussian noise and jamming environments and reduce to the traditional optimal diversity system in a Gaussian environment. The dynamics of the quality estimator allow operation in the presence of high-power impulsive interference by exploiting the signal and noise differential travel times to individual sensors. The spatial coherence structure of the shallow water acoustic channel shows relatively low signal coherence at separations as short as 0.35 m. Increasing receiver spacing beyond 5 m offers additional benefits in the presence of impulsive noise and larger-scale inhomogeneities in the acoustic field. A number of data transmission experiments were carried out to demonstrate system performance in realistic underwater environments. >
82 citations
TL;DR: This work addresses the problem of identifying the characterizing parameter of the PSF, which corresponds to motion or out-of-focus blur, from blurred and noisy images, by adding a preprocessing stage for noise reduction and applying an adaptive, quefrency-varying, comb-like window in the cepstral domain to enhance the blur parameter identification.
78 citations
TL;DR: The authors present a modified direct-digital synthesizer which uses noise shaping to reduce the effects of phase-accumulator truncation on the output spectrum, making the synthesizer suitable for high-performance signal processing.
Abstract: The authors present a modified direct-digital synthesizer which uses noise shaping to reduce the effects of phase-accumulator truncation on the output spectrum. The discrete spectral disturbances associated with this truncation error are strongly reduced with the proposed method, making the synthesizer suitable for high-performance signal processing. The proposed architecture uses first-order noise shaping. Higher order noise shaping can also be used if required. The modified circuit offers considerable improvement in the synthesizer performance when the generated frequency is low with respect to the clock frequency. >
TL;DR: The proposed homogeneous region growing mean filter technique effectively smoothes ultrasonic speckle and completely suppresses isolated impulsive noise over the entire texture in addition to preserving the edge information.
Patent•
12 Jul 1991TL;DR: In this paper, a quantizing error reducer for an audio signal is constructed to feed a quantising error back to the input side of a quantizer through a noise filter, wherein the coefficient of the noise filter is set on the basis of information relating to the equiloudness curve, thereby making it possible to reduce the noise in the hearing sense.
Abstract: There is provided a quantizing error reducer for an audio signal, which is constructed to feed a quantizing error back to the input side of a quantizer through a noise filter, wherein the coefficient of the noise filter is set on the basis of information relating to the equiloudness curve, thereby making it possible to reduce the noise in the hearing sense.
TL;DR: The algorithm is based upon a probabilistic analysis of local behavior in the phase space in which the deterministic source is embedded and creates a kind of time domain “filter” for the noisy data from areference orbit of the system, and corrects the observed noisy data in an incremental fashion to adjust it to match this filter.
Abstract: We present a probabilistic approach to the problem of additive source separation characterized by wide band power spectra when one of the sources is deterministic and low-dimensional (regular or chaotic). The algorithm is based upon a probabilistic analysis of local behavior in the phase space in which the deterministic source is embedded. It creates a kind of time domain “filter” for the noisy data from areference orbit of the system, and it then corrects the observed noisy data in an incremental fashion to adjust it to match this filter in the sense of having as close as possible the same invariant distribution in the phase space. After describing the general method, we demonstrate its use on three familiar chaotic systems: first, two examples of chaotic maps (Henon and Ikeda) in two dimensions and then a three-dimensional flow (the Lorenz system). The data from each of these is perturbed by uniformly distributed noise or by a spike signal, and we show that the method gives excellent results in recovering the original clean signal even when noise levels are quite large. Our present method appears to work even when the signal-to-noise ratio is quite small. The method works in cases where classical linear methods fail because both “signal,” and “noise,” are broadband spectrally.
14 Apr 1991
TL;DR: The problem of speech recognition in the presence of interfering nonstationary noise is addressed and a method for noise reduction in the cepstral domain based on a multilayer network is proposed and tested on a large database of isolated words contaminated with non stationary F-16 jet noise.
Abstract: The problem of speech recognition in the presence of interfering nonstationary noise is addressed. A method for noise reduction in the cepstral domain based on a multilayer network is proposed and tested on a large database of isolated words contaminated with nonstationary F-16 jet noise. The speech recognition system consists of an auditory preprocessing module, the cepstral noise reduction multilayer network, and a neural network classifier. The noise reduction network performs a nonlinear autoassociative mapping in the cepstral domain between a set of noisy cepstral coefficients and a set of noise-free cepstral coefficients. The average recognition rate on a test database was improved up to 65% when the noise reduction network was added to the speech recognition system. >
TL;DR: An all-fiber gyroscope that uses a 1.06-microm multicore Nd-doped fiber source pumped by a laser-diode array is demonstrated and intensity noise subtraction is used to reduce the excess noise that arises from a source bandwidth limitation.
Abstract: An all-fiber gyroscope that uses a 1.06-microm multicore Nd-doped fiber source pumped by a laser-diode array is demonstrated. Intensity noise subtraction is used to reduce the excess noise that arises from a source bandwidth limitation. The random walk coefficient is improved by a factor of 3, to 2.5 x 10(-4) deg radicalh (1-km fiber, 16-cm radius).
TL;DR: In this article, the physical characteristics of global optimum active noise control on the surface and in the far field of extended radiators were examined, and the authors showed that with a suitable number of sources, a structure radiating like a monopole to the far-field will radiate as a dipole after activation of the active-noise control sources.
Abstract: This paper examines the physical characteristics of global optimum active noise control on the surface and in the far field of extended radiators. Global optimum active noise control expresses the total acoustic power radiated by a collection of noise sources and active noise control sources with a quadratic equation in terms of source strengths or, for extended structures and control sources, surface velocities. Quadratic optimization theory is then used to obtain the secondary source strengths or surface velocities that yield the minimum total radiated power. The nature of the quadratic expression guarantees that the solution yields the global minimum, not simply a local minimum. The physical mechanism for optimum active noise control is the increased complexity of the combined sources. That is, with a suitable number of sources, a structure radiating like a monopole to the far field will radiate as a dipole after activation of the active noise control sources. While global optimum control is guaranteed...
Patent•
16 Sep 1991TL;DR: In this article, the fit parameters for a noise reduction method were computed using a slide sum technique, where a digital image is first decomposed into two types of regions, smooth regions and edge regions, and a plane fit parameter is computed at every point.
Abstract: The present invention efficiently and quickly computes the fit parameters for a noise reduction method wherein a digital image is first decomposed into two types of regions; smooth regions and edge regions. Within the digitized image a plane fit parameter is computed at every point. The computation of the fit parameters incorporates convolutions carried out in an efficient manner by the use of a slide sum technique.
TL;DR: In this paper, simple vector concepts can be used in the determination of noise parameters from measured data, leading to a simplification in the least-square fitting algorithm, complete determination of the admittance loci that produce ill conditioning, and simple criteria for the selection of source admittances that minimize the sensitivity of the noise parameters to experimental error.
Abstract: Simple vector concepts can be used in the determination of noise parameters from measured data. The use of such concepts leads to a simplification in the least-square fitting algorithm, complete determination of the admittance loci that produce ill conditioning, and simple criteria for the selection of source admittances that minimize the sensitivity of the noise parameters to experimental error. The sensitivity of the noise parameters to small perturbations in the reflection coefficients is compared for a group of source admittances presented in previous work. The results show that a great reduction in the error of the noise parameters can be achieved by properly selecting the source admittances. >
TL;DR: In this article, the authors proposed an approach to active noise control (ANC) applicable to three-dimensional radiators, which is based on minimizing the total radiated power by adding secondary acoustic sources on the primary noise source.
Abstract: This paper presents the theoretical development of an approach to active noise control (ANC) applicable to three-dimensional radiators. The active noise control technique, termed ANC Optimization Analysis, is based on minimizing the total radiated power by adding secondary acoustic sources on the primary noise source. ANC Optimization Analysis determines the optimum magnitude and phase at which to drive the secondary control sources in order to achieve the best possible reduction in the total radiated power from the noise source/control source combination. For example, ANC Optimization Analysis predicts a 20 dB reduction in the total power radiated from a sphere of radius at a dimensionless wavenumber ka of 0.125, for a single control source representing 2.5 percent of the total area of the sphere. ANC Optimization Analysis is based on a boundary element formulation of the Helmholtz Integral Equation, and thus, the optimization analysis applies to a single frequency, while multiple frequencies can be treated through repeated analyses.
TL;DR: In this paper, the mode competition theory was applied to reduce the mode hopping noise in the injection current of a semiconductor injection laser, and it was confirmed that the coupling effect among the longitudinal modes is released with the high-frequency (HF) current superposition because of enhanced vibration of the injected electron.
Abstract: Semiconductor injection lasers sometimes reveal excess noise called the mode hopping noise, which is associated with the mode hopping phenomena among longitudinal modes. The mode hopping phenomena are caused by coupling effects among lasing modes. To reduce this kind of noise superposition of high-frequency (HF) current on the injection current is frequently used. This reduction method is theoretically analyzed based on the mode competition theory and is compared with experimental measurements. It is confirmed that the coupling effect among the longitudinal modes is released with the HF superposition because of enhanced vibration of the injected electron, resulting in reduction of the mode hopping noise. >
01 Aug 1991
TL;DR: In this article, the authors provide a physical and theoretical model to explain the measured jet noise suppression characteristics of different concepts, and provide a set of guidelines for evolving jet nozzles.
Abstract: The objectives of this chapter are to review and summarize the jet noise suppression technology, to provide a physical and theoretical model to explain the measured jet noise suppression characteristics of different concepts, and to provide a set of guidelines for evolving jet noise suppression designs. The underlying principle for all jet noise suppression devices is to enhance rapid mixing (i.e., diffusion) of the jet plume by geometric and aerothermodynamic means. In the case of supersonic jets, the shock-cell broadband noise reduction is effectively accomplished by the elimination or mitigation of the shock-cell structure. So far, the diffusion concepts have predominantly concentrated on jet momentum and energy (kinetic and thermal) diffusion, in that order, and have yielded better noise reduction than the simple conical nozzles. A critical technology issue that needs resolution is the effect of flight on the noise suppression potential of mechanical suppressor nozzles. A more thorough investigation of this mechanism is necessary for the successful development and design of an acceptable noise suppression device for future high-speed civil transports.
TL;DR: The technique has been adopted to process magnetic resonance images in order to reduce noise and artifacts, thereby improving image quality and implementing this technique to phantom and clinical images.
TL;DR: The empirical results indicate that a statistical method applied to a single measurement gives an image of about the same quality as an average of three single measurements.
TL;DR: An optical signal-processing technique for additive noise reduction that uses the noisy signal and a Gaussian reference beam to produce an adaptive Wiener filter is presented.
Abstract: We present an optical signal-processing technique for additive noise reduction that uses the noisy signal and a Gaussian reference beam to produce an adaptive Wiener filter. We experimentally demonstrate an improvement from 1 to 8 in the signal-to-noise ratio by using nonlinear gain in two-beam coupling in barium titanate to transmit 50% of the signal and 6% of the noise.
14 Apr 1991
TL;DR: A spectral sharpening process suitable for speech enhancement and noise reduction is presented and a particular characteristic of the approach is its virtually delayless performance and its structure which easily lends itself to integration.
Abstract: A spectral sharpening process suitable for speech enhancement and noise reduction is presented. The speech enhancement feature is an efficient precompensation technique for the reduced frequency selectivity encountered among many hearing impaired people. The noise reduction scheme has wider applications. A particular characteristic of the approach is its virtually delayless performance and its structure which easily lends itself to integration. >
Patent•
31 Oct 1991TL;DR: In this article, a circuit arrangement includes a motion compensated frame recursive noise reduction circuit and a motion-compensated Field Recursive Noise Reduction (FRSR) circuit, and the outputs of these two noise reduction circuits are then combined.
Abstract: In order to obtain the benefits of benefits of field and frame recursive noise reduction while reducing the detriments thereof, a circuit arrangement includes a motion compensated frame recursive noise reduction circuit and a motion compensated field recursive noise reduction circuit. The outputs of these two noise reduction circuits are then combined. The motion compensation vectors for the field and frame recursive noise reduction circuits are determined such that in the case of little or no detected motion, the system is biased toward frame recursive noise reduction to achieve the best resolution on still and slow moving scenes. While in the case of increased motion, the circuit reverts more to field recursive noise reduction to reduce smearing.
TL;DR: In this article, an approach that exploits space-varying objective functions, which results in locally varying filtering schemes for the restoration problem, is introduced, where the ultimate objective is to invert the linear and nonlinear transformations introduced by the image-formation system and remove as much long- as well as short- and/or medium-tailed noise.
Abstract: The restoration of images in the presence of mixed noise is considered. An approach that exploits space-varying objective functions, which results in locally varying filtering schemes for the restoration problem, is introduced. The analysis is performed for both linear and nonlinear image-formation systems. It is found that the approach results in a degradation of the detailed structure existing in the original image. Hence, the problem of simultaneous noise removal and edge preservation in image restoration is addressed. The ultimate objective is to invert the linear and nonlinear transformations introduced by the image-formation system and remove as much long- as well as short- and/or medium-tailed noise, while simultaneously preserving the detailed structure of the image. The properties of the proposed algorithms are demonstrated through restoration examples. >
14 Apr 1991
TL;DR: Techniques for improving the performance of CELP (code excited linear prediction) type speech coders while maintaining reasonable computational complexity are explored and a harmonic noise weighting function which enhances the perceptual quality of the processed speech is introduced.
Abstract: Techniques for improving the performance of CELP (code excited linear prediction) type speech coders while maintaining reasonable computational complexity are explored. A harmonic noise weighting function which enhances the perceptual quality of the processed speech is introduced. The combination of harmonic noise weighting and subsample resolution pitch significantly improves the coder performance for voiced speech. A 6.9 kb/s VSELP speech coder which incorporates subsample resolution pitch and harmonic noise weighting is described. Complexity reduction techniques are discussed which allow the coder to be implemented using a single fixed point digital signal processor. >