No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Analysis of low-order autoregressive models for ultrasonic grain signal characterization
Abstract
When testing materials nondestructively with ultrasound, the grain scattering signal provides information that may be correlated to regional microstructure variation. Second and third-order autoregressive (AR) models are used to evaluate the spectral shift in grain signals by utilizing features such as resonating frequency, maximum energy frequency, or AR coefficients. Then, Euclidean distance, based on these features, is applied to classify grain scattering characteristics. Using both computer simulated data and experimental results, the probability of correct classification is found to be about 75% for the second-order AR model and 88% for the third-order AR model, when the conditions are such that the expected shift between the center frequency of echoes is less than 4%. This implies that, by increasing the order of the AR model, the frequency information extracted from the random signal is increased, which can result in obtaining a better classification.< >
... The order of the model (p, q) controls the error associated with the AR signal approximation [26]. Small orders ignore the main and long-term statistical properties of the original signal while larger ones may lead to overfitting effects [25], [26]. Therefore, selecting the order of the model is a key problem and there are several methods to do it [25]- [28]. ...
... Small orders ignore the main and long-term statistical properties of the original signal while larger ones may lead to overfitting effects [25], [26]. Therefore, selecting the order of the model is a key problem and there are several methods to do it [25]- [28]. Here, the firstorder model was adopted because it was confirmed by [43] that in this scope it leads to the minimum error probability. ...
... US textural features extracted from the first-order AR model coefficients and the multiscale Haar WT analysis (level 1) are particularly relevant. This result is in accordance with the results of [21], [23]- [25]. WT [21], [23], [24] and AR coefficients [25] based features have high discriminative power in the assessment of CLD stages. ...
Chronic liver disease is a progressive disease, most of the time asymptomatic, and potentially fatal. In this chapter, an automatic procedure to stage the disease is proposed based on ultrasound (US) liver images, clinical and laboratorial data.
A new hierarchical classification and feature selection (FS) approach, inspired in the current diagnosis procedure used in the clinical practice, here called Clinical-Based Classifier (CBC), is described. The classification procedure follows the well-established strategy of liver disease differential diagnosis. The decisions are taken with different classifiers by using different features optimized to the particular task for which they were designed. It is shown that the CBC method outperforms the traditional one against all (OAA) method because it take into account the natural evolution of the hepatic disease. Different specific features are used to detect and classify different stages of the liver disease as it happens in the classical diagnosis performed by the medical doctors.
The proposed method uses multi-modal features, extracted from US images, laboratorial and clinical data, that are known to be more appropriated according to the disease stage we want to detect. Therefore, a battery of classifiers and features are optimized and used in a hierarchical approach in order to increase the accuracy of the classifier.
For the normal class we achieved 100% accuracy, for the chronic hepatitis 69. 2%, for compensated cirrhosis 81. 48%, and for decompensated cirrhosis 91.7%.
... Among others, those procedures have given raise to the development of heuristic inverse problems, based on empirical models (trained over empirical data, in contrast to phenomenological approaches relying on underlying physical processes). The absence of a direct link between the physical process and the empirical model allows one to investigate the feasibility of using blind signal models (e.g. the modeling of ultrasonic signals using an autoregressive model [12]). Most of the heuristic techniques devoted to material characterization aim at mapping the signal space (e.g. ...
... Autoregressive analysis has also been used as an alternative approximation to enhance ultrasonic signals. Wang et al. [12] used second and third-order autoregressive (AR) models to evaluate the spectral shift in grain signals by utilizing features such as resonating frequency, maximum energy frequency or AR-coefficients. On the other hand, Izquierdo et al. [20] presented a method that considers the time-varying spectral content of the received echoes, based on a time-varying autoregressive model of the structural noise. ...
... In this second experiment, the specimen was located at the focal distance (d f = 30 mm) of the focused transducers, and scanned over a two-dimensional plane parallel to the transducer areas (C-scan mode) in an immersion tank with degassed water at room temperature equipped with three-dimensional motion controllers. The response signals were 12 Courtesy from the Institute of Polymers and Composites, TU Hamburg-Harburg, Germany Initially, the response signal was measured at an undamaged location (far from the impacted area) for calibration. Then, the measurement procedure was repeated over an area of 40 × 20 mm 2 with a step of 1 mm around the impacted area, providing a data set of 860 measurements (plus one taken in water only, after removing the specimen). ...
... Disease processes in several different organs have been shown to be accompanied by changes in ultrasonic scattering properties [11]. In this sense, the pulse echo data from different grain types contain distinguishable statistical regularities [10]. The microstructure of the backscattered echoes is complex, depending on the overall characteristics of the ultrasound scanner in terms of bandwidth and beamwidth, on the scattering properties of the propagation path and on the attenuation caused by absorption, scattering, and beam spreading (diffraction effect) [10]. ...
... In this sense, the pulse echo data from different grain types contain distinguishable statistical regularities [10]. The microstructure of the backscattered echoes is complex, depending on the overall characteristics of the ultrasound scanner in terms of bandwidth and beamwidth, on the scattering properties of the propagation path and on the attenuation caused by absorption, scattering, and beam spreading (diffraction effect) [10]. ...
... The first goal was to establish the optimal AR order for classification purposes and then used the AR coefficients has features. Like in [10], the results displayed in Table 1, confirm that the low-order AR model can characterize the classes in study. The error probability (PE) increases with the order of the model up to the 4 th . ...
Liver steatosis is mainly a textural abnormality of the hepatic parenchyma due to fat accumulation on the hepatic vesicles. Today, the assessment is subjectively performed by visual inspection. Here a classifier based on features extracted from ultra-sound (US) images is described for the automatic diagnos-tic of this phatology. The proposed algorithm estimates the original ultrasound radio-frequency (RF) envelope signal from which the noiseless anatomic information and the tex-tural information encoded in the speckle noise is extracted. The features characterizing the textural information are the coefficients of the first order autoregressive model that de-scribes the speckle field. A binary Bayesian classifier was implemented and the Bayes factor was calculated. The classification has re-vealed an overall accuracy of 100%. The Bayes factor could be helpful in the graphical display of the quantitative results for diagnosis purposes.
... The order of the model (p, q) controls the error associated with the AR signal approximation [26]. Small orders ignore the main and long-term statistical properties of the original signal while larger ones may lead to overfitting effects [25], [26]. Therefore, selecting the order of the model is a key problem and there are several methods to do it [25]- [28]. ...
... Small orders ignore the main and long-term statistical properties of the original signal while larger ones may lead to overfitting effects [25], [26]. Therefore, selecting the order of the model is a key problem and there are several methods to do it [25]- [28]. Here, the firstorder model was adopted because it was confirmed by [43] that in this scope it leads to the minimum error probability. ...
... US textural features extracted from the first-order AR model coefficients and the multiscale Haar WT analysis (level 1) are particularly relevant. This result is in accordance with the results of [21], [23]- [25]. WT [21], [23], [24] and AR coefficients [25] based features have high discriminative power in the assessment of CLD stages. ...
Chronic Liver Disease (CLD) is most of the time an asymptomatic, progressive and ultimately potentially fatal disease. In this work, an automatic hierarchical procedure to stage CLD using ultrasound images, laboratory tests and clinical records is described. The first stage of the proposed method, called Clinical Based Classifier (CBC), discriminates healthy from pathologic conditions. When non healthy conditions are detected the method refines the results in three exclusive pathologies in a hierarchical basis: i) chronic hepatitis, ii) compensated cirrhosis and iii) decompensated cirrhosis. The features used as well as the classifiers (Bayes, Parzen, Support Vector Machine and k- Nearest Neighbor) are optimally selected for each stage. A large multi modal feature database was specifically built for this work containing 30 chronic hepatitis cases, 34 compensated cirrhosis cases and 36 decompensated cirrhosis cases, all validated after histopathologic analysis by liver biopsy. CBC classification scheme outperformed the non-hierachical one against all scheme, achieving an overall accuracy of 98.67% for the normal detector, 87.45% for the chronic hepatitis detector and 95.71% for the cirrhosis detector.
... In some cases [7, 8, 10, 11], under the linear dependence hypothesis and assuming a Gaussian envelope pulse, explicit relations between the attenuation slope and the representative frequency can be obtained. In other cases [3, 12, 13], there are experiments directly showing correlations between the centroid frequency and properties of the material such as the grain size. In this paper we present some new contributions to the problem of estimation of attenuation by means of time– frequency analysis of the backscattering noise. ...
... {A n } and {τ n } are also independent. The superposition model of equation (1) is used by many authors working with different materials and tissues2345678910111213 due to its simplicity and to the possibility of introducing elements of signal theory into the development of algorithms for nondestructive examination. It by no means tries to be an exact description of the physical mechanisms that are produced in the complex ultrasonic propagation, nor an exact representation of the material microstructure. ...
... Note that α z (ω) in (4) may include not only the attenuation due to scattering but also that due to absorption or other effects like beam divergence. Depending on D/λ, the different components may have more or less significance, as well as different frequency and grain size dependences, but the exponential model is valid for the Rayleigh, stochastic and diffusion regions [3, 12, 13]. Our analytical work does not impose any constraints or particular frequency dependence, except for the exponential decay of the amplitude and its general dependence on frequency (equation (4)). ...
We consider in this paper the general problem of ultrasonic characterization of materials by means of analysing the dependence of attenuation on the frequency and depth of the backscattering noise. Some theoretical analysis is included to define the procedure and to gain insights into the suitability of the approach. From the depth-and frequency-dependent attenuation diagrams we may derive material signatures to be used for classification-oriented problems or derive parameters to be correlated with material properties. A particular case is considered: the characterization of cement pastes. For this case we propose the use of attenuation profiles as material signatures, and we show that the area of the profile exhibits good correlation with the porosity measured by destructive methods.
... Under certain assumptions related to the wavelength of the ultrasonic wave and the scattering size, we can model scattering material following di erent authors [2,5,7,12,14,16], as shown in Fig. 2. The model is composed of an homogeneous non-dispersive media, and randomly distributed punctual scatterers. In general, this model is not intended to be a rigorous description of the material microstructure, although in some cases it could be a good approximation. ...
... Eqs. (9), (16) and (21) open many di erent possibilities for material characterization from knowledge of the HOS of the backscattered signal. In particular, the moments m i; A i = 1; 2; 3 may be extracted from knowledge of the cumulants up to order three. ...
... This, in turn, may be done in many di erent ways depending on the particular strategy for using Eqs. (9), (16) and (21) and on the particular selected values for in (16) or for 1 , 2 in (21). We propose to ÿrst estimatem 1; A from Eq. (9), then estimatem 2;A from (16) and ÿnallym 3;A from (21). ...
We have derived equations relating to the cumulants of the backscattered signal to material and transducer parameters. Then, we proposed a practical method to estimate the material grain moments from estimates of the cumulants at some particular values. The proposed model and techniques are verified on some phantoms having different scatterer density and grain sizes.
... 1. Continuous monitoring the process of the cementpaste thickening, 2. Monitoring the cement quality, 3. Continuous monitoring the process of the cementslurries thickening, 4. Defining the time of cement-slurries readiness for a treatment, and the time of cement-slurries binding, 5. Monitoring the quality of the natural and artificial concrete aggregate, 6. Monitoring the quality of other concrete admixtures. ...
... If the scattered distribution s is assumed to be Gaussian distributed with mean µ and standard deviation δ, then its characteristic function and the statistical scatterer power spectrum can be obtained [6]. This power spectrum exhibits nearly periodic peaks. ...
When constructing High Voltage Generators, we are faced with a problem of choice: high power, high voltage or high frequency. It is known that we can either do it with a compromise or we could choose appropriate preferential properties. For voltages by a couple of hundred Volts and currents of a couple of Amperes we can use an analogue high-voltage amplifier, which could be combined with a corresponding function generator and used as a source for testing; its shortcoming is its limitation regarding frequencies. A generator for efficient spectral analysis should be able to change frequencies and amplitudes within wide range. We can use high-voltage impulse-source with adoptive amplitude AHVG (Adopting High Voltage Generator), which is an attempt at unification within an HV impulse generator of the necessary properties for a spectral analysis. The specially adopting circuits for adjusting output signal amplitude enable the security and reliability of the impulse source, which also prevent unnecessary losses in the output amplifier. Adopting circuits enable the output amplifier always to be in its optimal working condition, which has an important influence on the characteristics and operational reliability of the output power stage.
... The material is modeled using a continuous autoregressive (AR) model [4], [5] with parameters connected to physical properties, related to the thicknesses of the material layers and the reflection coefficients given by the layer boundaries. ...
... All subsequent analysis is then done on the model parameters rather than on the entire waveform, which enables costeffective storage and fast processing. The material is modeled using a continuous autoregressive (AR) model [4], [5] with parameters connected to physical properties, related to the thicknesses of the material layers and the reflection coefficients given by the layer boundaries. In this paper we derive a general model structure for reflections from layered media. ...
In materials consisting of several thin layers, multiple reflections within the structure give rise to received ultrasonic signals composed of overlapping echoes. In this paper we present a parametric model that can be used to decompose such signals into the individual reflections. We derive a maximum likelihood estimator for the the model parameters, which are then used in a generalized likelihood ratio test (GLRT) to detect flaws in multi-layered structures. We show with simulations how the presence of a thin bonding layer in a three-layer structure can be detected. The probability of detection is shown to be ap96%, for a signal-to-noise ratio (SNR) of 15 dB and a probability of false alarm of 5 %
... A tentative alternative is to characterize the material by using the many superimposed echoes scattered by the inner microstructure of the material. This generates the so called grain noise (GN) [2], [3], [5], [6], [10]- [12], [15]. The only constraint to obtain GN is to use an adequate excitation frequency so that [ ] 3 . ...
... Linear dependence is a realistic hypothesis in tissue analysis, but it is not valid in general for all materials. Instead, in [10], [11], [12], [15], the representative frequency variations are directly correlated with material properties, in a purely experimental manner, with no special constraints about the attenuation dependence on frequency. In [5], the maximum energy frequency profiles are considered for flaw detection. ...
In this paper, we propose a technique for material characterization by using centroid frequency profiles (CFP) of ultrasound echo signals. These echo signals are composed by grain noise due to the superposition of many small echoes from the inner microstructure plus observation noise. A CFP indicates the centroid frequency dependence on depth, corresponding to power spectrum density assessments at different depths. We show in the paper the relation between the mean and variance of the CFP and the grain-to-observation-noise-ratio (GOR) at every depth. The GOR depends on the material ultrasound attenuation, so that CFP may be used for material characterization. Although we consider here the estimation of cement paste porosity, the proposed technique may have general applicability. Cement paste is the main component of mortar and concrete. Therefore, cement porosity is an important problem because the vulnerability (and thence the durability) of these construction materials to external agents depends heavily on it. Experiments have been made to show the correlation between cement paste porosity and a penetration parameter obtained from the CFP.
... If the attenuation may be considered to be linearly dependent on frequency , the center frequency estimates are used for estimating the attenuation coefficient1234. In other cases the center frequency estimates are correlated with properties of material567, or for flaw detection [8]. There are different alternatives for computing the center frequency. ...
... Once we have the PSM, the centroid frequency may be computed and then considered to be a center frequency estimate but some bias is unavoidable due to integration band. The problem of bias is not present if the center frequency is estimated by means of the maximum [4,7,8] of the PSM, because no integration band is required. Additionally the background noise problem is less important, as the maximum is located at the highest signal to noise ratio band of the spectrum. ...
In this paper we propose a new technique for estimating the center frequency of the ultrasound pulse from records of backscattering noise. We start by considering that the conventional maximum frequency method can be seen as a filtering (differentiator) of the pulse spectrum magnitude followed by a searching for the zero-crossing value. The new approach replaces the differentiator by a Hilbert transformer. We show in the paper that the proposed method has less variance than the maximum frequency method. In particular, we analyse the performance assuming that the real cepstrum method is used for extracting pulse spectrum magnitude. We give an upper bound for the variance reduction when practical criteria are applied for fitting the cepstrum cut-off frequency. The analytical work is verified by real and simulated data.
... All these algorithms help in making more precise diagnostics in medical systems or faster and more reliable detections in industrial applications. It is very frequently assumed in the derivation of the algorithms that the A-scan can be segmented into stationary parts345678. It is also common to assume that the ultrasonic system can be modeled as a linear system. ...
... In the following, we analyze the nonlinearity detection capability of the different proposed metrics on the two types of nonlinearities described in section 3. The ultrasonic pulse h(n) was modeled with a second-order AR [3]. Although some authors propose higher order AR models [13] or even ARMA models [4], the simulations carried out did not give significantly different results (detection percentages) than those obtained in the second-order AR model. ...
In this paper, we propose and analyze by means of simulations the use of surrogate data algorithms for blind detection of nonlinearities in multiple-echo ultrasonic signals. We assume a blind scheme so that no information about the input (emitted ultrasonic pulse) can be used. The metrics and equations that model some nonlinear situations are carefully reviewed. Also, closed form equations of the third-order metrics from a simplified second-order Volterra kernel are derived. Computer simulations show that the surrogate data technique is a potentially powerful tool for blind detection of nonlinearities in multiple-echo ultrasonic signals if adequate metrics are chosen. They also reveal interesting trade-offs among parameters that model ultrasonic systems and detection percentages.
... However. to measure the signal attenuation is difficult as the attenuation shows variations of local grain size across the entire propagation path. Therefore, to estimate the grain size, the scattering signal is characterized by utilizing the statistical variation in the scattered energy as a function of depth [3]. In addition, the characterization of the backscattered signal is since the signal intensity is the non-explicit function of average grain size, random distribution of grains, and ultrasonic frequency [4]. ...
... However, it is difficult to measure the signal attenuation which shows local grain size variations along the entire propagation path. Thus, the scattering signal is characterized to estimate the grain size distribution using the statistical variation in the scattered energy as a function of depth [3] [4]. The intensity of the backscattered signal is the non-explicit function of average grain size, ultrasonic frequency, and random distribution of grains. ...
... La fréquence de résonanceétant déduite de la phase de ces pôles, il en résulte [WSJ91], ...
The tire is an essential element for the handling, comfort and safety of a vehicle. However, an under-inflation increases the risk of burst, causes rapid tire wear and increases fuel consumption. It is therefore important to develop tire pressure monitoring systems (TPMS). A first approach, reffered to as "direct" method, consisting on using pressure sensors is expensive and unreliable (possible sensors faults). The new generation of TPMS promotes "indirect" methods, without pressure sensors. Monitoring is performed from the physical quantities related to the pressure. The tire pressure drop results in increasing the wheel angular velocity, shifting the vehicle vibratory modes, reducing the wheel effective radius and increasing its rolling resistance. The approach based on a comparative analysis of the wheels angular velocities does not meet the requirements regarding the minimum pressure drop that must be detected and the number of detectable deflated tires. Adding a spectral analysis of each wheel angular velocity signal enhances the robustness of such an approach. However, it shows significant convergence time. The first aim of the thesis work is therefore to optimize such algorithms and reduce the computational time. An alternative approach requiring low maintenance costs is the implementation of observers for quantities related to the pressure. The second objective of this thesis is to propose observers able to estimate both the effective radius and the rolling resistance of the wheels without using additional sensors. The work in this thesis deals with methodological and experimental aspects through the implementation of the methods and their application to real data.
... Raw RF signal preserves the important information within the signal [1] with direct and precise interpretation for medical diagnosis. In ultrasonic non-destructive evaluation applications, the backscattered RF signal has information about the geometric shape, size and orientation of the scatterers within the propagation path [2,3]. ...
Ultrasonic systems are widely used in imaging applications for non-destructive evaluation, quality assurance and medical diagnosis. These applications require large volumes of data to be processed, stored and/or transmitted in real-time. Therefore it is essential to compress the acquired ultrasonic radio frequency (RF) signal without inadvertently degrading desirable signal features. In this paper, two algorithms for ultrasonic signal compression are analysed based on: sub-band elimination using discrete wavelet transform; and decimation/interpolation using time-shift property of Fourier transform. Both algorithms offer high signal reconstruction quality with a peak signal-to-noise ratio (PSNR) between 36 to 39 dB for minimum 80% compression. The computational loads and signal reconstruction quality are examined in order to determine the best compression method in terms of the choice of DWT kernel, sub-band decomposition architecture and computational efficiency. Furthermore, for compressing a large amount of volumetric information, three-dimensional (3D) compression algorithms are designed by utilising the temporal and spatial correlation properties of the ultrasonic RF signals. The performance analysis indicates that the 3D compression algorithm presented in this paper offers an overall 3D compression ratio of 95% with a minimum PSNR of 27 dB.
... The material is modeled using a continuous autoregressive (AR) model [4], [5] with parameters connected to physical properties, related to the thicknesses of the material layers and the reflection coefficients given by the layer boundaries. In this paper we derive a general model structure for reflections from layered media. ...
In materials consisting of several thin layers, multi-ple reflections within the structure give rise to received ultrasonic signals composed of overlapping echoes. In this paper we present a parametric model that can be used to decompose such signals into the individual reflections. We derive a Maximum Likelihood Estimator for the the model parameters, which are then used in a Generalized Likelihood Ratio Test (GLRT) to detect flaws in multi-layered structures. We show with simulations how the presence of a thin bonding layer in a three-layer structure can be detected. The probability of detection is shown to be ≈ 96%, for a signal-to-noise ratio (SNR) of 15 dB and a probability of false alarm of 5%.
... The AR(3) choice is convenient also as far as the PSD centroid evaluation is concerned, which can be approximated by the maximum amplitude frequency. given by the following analytical formula (Wang et al. 1991): ...
The paper addresses the problem of topological maps production for tissue characterization, based on spectral parameters extracted from RF backscattered ultrasonic signals. The spectral parameter dealt with is the power spectral density centroid, since it is an efficient indicator of the tissue microstructure characteristics. The analysis is performed using a recursive least squares scheme with a variable forgetting factor, based on low-order autoregressive models. The method has a small computational burden and good tracking properties, suitable for ocular pathologies differentiation. It was tested on simulated signals, on a text-object and finally applied to signals scattered by in-vitro eye specimens.
... A-Scans can be modeled as shown in figure 1 [3,4,5,6,7,8], where central frequency of the bandpass filter f 0 (t) depends of time to resemble the selective attenuation of some materials and the flaw echo can be artificially introduced at any time. ...
Echo detection on time-varying signals is a typical problem of signal processing. A not so typical application of this problem is the detection of foreign bodies in the alimentary industry. In this work we are going to present some results of a project whose objective was to develop an ultrasonic au- tomatic system for detection of foreign bodies. The algorithm presented merge some ideas of time fre- quency representation (TFR) and morphological image pro- cessing to get an easy to implement and highly customizable algorithm that could be applied to many different products and situations.
... In [18], normalized LMS is explored for ultrasonic backscattered signal. In [19][20], ultrasonic NDE signals and medical ultrasound images are deconvoluted using adaptive filter. In [21], a simulation study of adaptive filter on ultrasonic NDE signals has been conducted. ...
Adaptive filter has been widely used in different applications for interference cancellation, predication, inverse modeling and identifications. In this paper, Field Programmable Gate Array (FPGA)-based adaptive noise cancellation is studied for adaptive filtering in ultrasonic non-destructive evaluation. Simulation and experimental results showed that backscattered noise from microstructures inside material can be efficiently reduced by adaptive filter. Additionally, four different architectures of filter realization on FPGA are discussed and compared. This type of study could have a broad range of applications such as target detection, object localization and pattern recognition.
... En otros casos, [9] y [10], hay experiencia en los que se muestran correlaciones entre la frecuencia centroide y las propiedades como el tamaño de grano del material. En estas aplicaciones se trabaja con la medida de la atenuación en la envolvente. ...
Ultrasonic backscattering noise appears in a large number of nondestructive testing applications, in the general area of tissue or materials characterization. We consider in this contributions the time-frequency analysis of backscatterin g noise with aim of obtaining depth profiles of some parameters related to the attenuation. The proposed analysis may have general applicability in the characterization of materials or tissues having depth dependent properties. For example, it may be useful for measuring the penetration of repairing substances in deteriorated building elements. Also it may be of interest to obtain signatures of the material for classification purposes.
... In order to increase the probability of defect detection and to decrease the probability of false alarm, various signal processing techniques have been utilised for noise reduction and enhancement of detected echoes. Signal averaging, autoand cross-correlation [18,19], matched filtering, adaptive filtering, non-coherent detector [20], autoregressive analysis [21,22], deconvolution, order statistics, frequency spectrum analysis [23], cepstral analysis, split spectrum processing (SSP), WT and model-based optimisation algorithms [6] have all been used to analyse ultrasonic signals. In all these techniques, the signal is analysed either in the time domain, or in the frequency domain or in the time -frequency domain. ...
Various signal processing techniques have been used for the enhancement of defect detection and defect characterisation. Cross-correlation, filtering, autoregressive analysis, deconvolution, neural network, wavelet transform and sparse signal representations have all been applied in attempts to analyse ultrasonic signals. In ultrasonic nondestructive evaluation (NDE) applications, a large number of materials have multilayered structures. NDE of multilayered structures leads to some specific problems, such as penetration, echo overlap, high attenuation and low signal-to-noise ratio. The signals recorded from a multilayered structure are a class of very special signals comprised of limited echoes. Such signals can be assumed to have a sparse representation in a proper signal dictionary. Recently, a number of digital signal processing techniques have been developed by exploiting the sparse constraint.This paper presents a review of research to date, showing the up-to-date developments of signal processing techniques made in ultrasonic NDE. A few typical ultrasonic signal processing techniques used for NDE of multilayered structures are elaborated. The practical applications and limitations of different signal processing methods in ultrasonic NDE of multilayered structures are analysed.
... This is the classical method used in level II courses for inspectors [1] ; nevertheless, it is relatively unreliable and it is very dependent of inspector's experience and skills. Since the development of computers after 1980, and the implementation of a great deal of signal processing software, the use of the Fourier Transform Representation to evaluate different material properties and characteristics were extensively reviewed [2][3][4][5] ; nonetheless, the frequency representation was not appropriate enough to achieve the signal classification task. Nowadays computers and software allows high velocity and a lot of operations for signal processing, it is possible to change the domain representation, transforming, filtering, decomposing, denoissing, compression and many others unsuspected operations. ...
The problem of flaw identification in fusion welded joints has been addressed using different methods and principles; nevertheless, we think this is a still open field, not solved completely yet. The deal amount of data that is possible to record with the new available technology today, makes automatic flaw classification an important engineering task. In the present work, we used two approximations to show other possibilities to tackle this unsolved problem. The first one is based on the properties of the covariance matrix, calculated from the level 2 approximation wavelet coefficients and obtained from a B-scan image with a defect. In the second procedure, the kurtosis and skewness of the continuous wavelet coefficients of five scales (taken from several ultrasonic signals) are used to create a feature space to train a Fisher Linear Classifier to discriminate common defects in welding joints. The behavior of the classifier was tested to differentiate discontinuities and the preliminary results are presented.
... Among these methods, the ones based on autoregressive modeling (AR) has become more and more popular over the last few years, mainly because of their superior frequency resolution, the reduced variance of the estimation and their computation efficiency. Several applications of AR techniques for ultrasound signal processing can be noted: application to Doppler techniques [4], [21], grain characterization [26] and mean scatterer spacing measurement [27]. In this study, a fast center frequency estimator based on AR spectral modeling is proposed for attenuation estimation. ...
A second-order autoregressive (AR2) model, whose parameters are
estimated with the Burg algorithm, is used to estimate the
center-frequency along echo signals and its evolution versus depth. Data
simulation of independent A-lines reflected by a homogeneous medium of
scatterers are generated by a computer model with attenuation values
ranging from 1 to 5 dB/cm MHz and an ultrasonic frequency of 5 MHz. The
performance of the estimator is evaluated for duration of time windows
ranging from 5.12 to 0.32 μs and different spectral sampling. The
comparison is made with the classical Fourier spectrogram technique
(FFT). It is found that the AR model provides a better estimation of
attenuation than the Fourier technique: the relative error of
attenuation is below 5% for windows between 0.64 to 2.56 μs, while
the one obtained with the Fourier technique lies between 10 and 70% for
the same window sizes. These results offer promises for determining
attenuation in biological medium which are highly attenuating either
because of their structure, like bone, or because high frequencies are
used
... Nevertheless, the echo pattern observed in the US liver images in different chronic liver disease stages is difficult to distinguish [2], as shown inFig. 1. In [3] it is referred that the pulse echo data from different grain types contain distinguishable statistical regularities. In addition, the study in [4] proposes a quantitative tissue characterization to increase the usefulness of US for evaluating the diffuse liver disease. ...
In this work the identification and diagnosis of various stages of chronic liver disease is addressed. The classification results of a support vector machine, a decision tree and a k-nearest neighbor classifier are compared. Ultrasound image intensity and textural features are jointly used with clinical and laboratorial data in the staging process. The classifiers training is performed by using a population of 97 patients at six different stages of chronic liver disease and a leave-one-out cross-validation strategy. The best results are obtained using the support vector machine with a radial-basis kernel, with 73.20% of overall accuracy. The good performance of the method is a promising indicator that it can be used, in a non invasive way, to provide reliable information about the chronic liver disease staging.
We analyze contemporary trends in the development of ultrasonic methods aimed at the evaluation of the degree of damage according to the “nonclassical” manifestation of structurally induced nonlinear effects, “structural” noise, and forward scattered signals of diffusion ultrasound and wave interferometry.
Methods for assessing scattered damage to materials are considered. Ultrasonic methods based on recording backscattered ultrasonic signals are among the most practically feasible methods for assessing scattered damage to materials at the mesolevel (the size of the probing wavelength). We describe methods for determining scattered damage in the volume of a material based on scanning the object surface with a normal double-crystal piezoelectric transducer, recording and statistically processing the backscattered signal in the form of an A-scan, constructing the spatial distribution of backscattering cross section in the form of a B-scan or tomographic image, and assessing damage in the material volume based on the relative change in the backscattering cross section or the “disorder” of its spatial image.
A method for processing backscattered signals in the ultrasonic testing of thick-walled articles is proposed. The method is based on normalizing the intensity of recorded signal to the energy per one cycle from the moment of signal emission to the first reflected bottom signal. The method allows one to take account of the level of probing signal, the effect of the acoustic contact, the reception-path gain, and the damping of the propagating signal as well as produce the profile of the cross section of backscattering of ultrasound by the material along the signal propagation path. The backscattering cross-section profile provides a basis for identifying scattered damage in the bulk of the material. Results of experimental testing of the method are provided. A technique for determining the minimum gain required for the “correct” reception of backscattered signals has been developed.
An algorithm has been developed for spectral analysis of backscatteredradiofrequency signals. It has been tested on a tissue model, a gelsuspension of calibrated latex spheres (between 20 and 100 μm in diameter),and on malignant intraocular tumours. The results so far appear promising regarding tissue differential diagnosis.
The interest in evaluating the histological features of uveal lesions by non-invasive techniques raises mainly from two issues.
Starting with a theoretical model of a dispersive material illuminated by an ultrasonic pulse [1] we developed a classifier based on the use of third order cumulants. We will compare the results with those obtained with a similar classifier but using only second order statistics. The underlying application is the classification of materials by non-destructive ultrasonic testing.
Great progress1 has been made recently in high frequency ultrasound imaging. In particular, several works have shown the interest of echographic exploration for frequencies ranging from 20 to 100 MHz, in dermatology and ophthalmology. These results are extending the field cf application of tissue characterization. The measurement of acoustic attenuation tissues has received much interest in the field of ultrasound tissue characterization. Indeed, several clinical applications have shown a correlation between attenuation values and pathological states.
The key to detecting damage to civil engineering structures is to find an effective damage indicator. The damage indicator should promptly reveal the location of the damage and accurately identify the state of the structure. We propose to use the distance measures of low-order AR models as a novel damage indicator. The AR model has been applied to parameterize dynamical responses, typically the acceleration response. The premise of this approach is that the distance between the models, fitting the dynamical responses from damaged and undamaged structures, may be correlated with the information about the damage, including its location and severity. Distance measures have been widely used in speech recognition. However, they have rarely been applied to civil engineering structures. This research attempts to improve on the distance measures that have been studied so far. The effect of varying the data length, number of parameters, and other factors was carefully studied.
Human bone has the remarkable capacity of remodeling and self repair through a complex regenerative healing process resulting in the gradual restoration of its mechanical properties and load bearing capacity [1]. Fracture healing goes through three distinct stages: reactive phase which includes inflammatory phase and granulation tissue formation, reparation phase which includes callus formation and lamellar bone deposition, and remodeling phase in which woven bone is replaced by mature bone. Although surgical intervention and fracture immobilization facilitate healing, fracture healing is known to be a physiological process. The connective tissue membrane covering the bone determines the healing process of a fractured bone. This connective tissue is the primary source of bone cells responsible for generating new bone during growth and repair. The length of the healing process depends on the extent of the fracture. It takes up to three weeks for the majority of upper bodily fractures and it takes up to four weeks for lower body fractures.
Ultrasound has appealing properties as a means to monitor quality of industrially manufactured food items. Changes in ultrasonic signal properties are associated with changes in raw material or process parameters. This paper introduces a concept based on an ultrasound transmission measurement which has shown ability to monitor the temperature of minced beef meat during simulated automated roasting. Commercially available beef samples were roasted, and the change in internal temperature of the beef was correlated to changes in the ultrasonic transmission signal. r2values of 0.94 were obtained.
The application of a technique called Split-Spectrum Processing, which was developed to achieve frequency diversity in one-dimensional (A-scan) ultrasonic application, to radar was examined. The major highlights and accomplishments of the project are: (1) adaptive techniques for determination of optional processing parameters; (2) robust techniques for target detection; (3) non-parametric target detection techniques; (4) rank determination for order statistic filters; and (5) order statistic characterization and signal processing for nonparametric robust and adaptive detection systems.
In radar, sonar, medical ultrasound, and ultrasonic nondestructive evaluation, environmental noise makes target detection challenging. Therefore, clutter rejection and noise cancellation are necessary for the system to correctly identify targets. In this study, an adaptive filtering algorithm is used to reject clutter and detect small targets in noisy ultrasonic backscattered signals. Simulation and experimental results show that adaptive filter can efficiently reduce the clutter and improve the detection capability.
Objective: To improve the measurement resolution of the fetal heart rate (FHR) beat-to-beat variation, which is an important prognostic indicator of fetal well-being. Method: The goal was reached using an autoregressive (AR) approach, instead of conventional algorithms based on the Fast Fourier Transform (FFT), for the analysis of the umbilical artery Doppler signals. Results: The resolution improvement obtainable with the proposed technique in the FHR beat-to-beat variation was proved both with simulation and with experimental results. The method was implemented in a personal computer (PC) based equipment which works in a quasi-real time mode. Conclusions: The proposed technique can be usefully applied to the FHR beat-to-beat variation measurements; the developed prototype is at present used for clinical practice at the cClinica Ostetrica e Ginecologica' of the University of Firenze.
In the ultrasonic nondestructive evaluation (NDE) of materials, spectral analysis of backscattered echoes is a useful tool for flaw detection, frequency-shift estimation, and dispersive echo characterization. In order to evaluate the local information, spectral analysis must be applied to short data segments and must offer high-frequency resolution. In this paper three high-resolution model-based spectral estimation techniques, i.e., the autoregressive (AR) method using the Burg algorithm, Prony's method for exponential signal representation, and the multiple signal classification (MUSIC) method, have been studied for ultrasonic NDE applications. These algorithms have been applied to both simulated data and experimental measurements for frequency estimation and flaw detection. The maximum energy frequency estimates using these methods show significant sensitivity to changes in the frequency of ultrasonic echoes. The AR method shows a more robust performance for frequency estimation than the Prony or MUSIC methods. (C) 1996 Acoustical Society of America.
In ultrasonic NDE applications, it is challenging to analyze ultrasonic backscattered signal in presence of high scattering noise. The ultrasound trace embodies the shape, size and orientation of reflectors and reveals the physical property of propagation path. The signal loss due to the effect of attenuation and scattering makes ultrasonic backscattered signal nonstationary. Therefore, a proper signal analysis technique is needed. As an effective tool of nonstationary signal analysis, time- frequency (TF) signal representations have been used in many applications, including ultrasound NDE. In this paper, other time-frequency analysis techniques have been reviewed. Especially, a model-based TF analysis and parameter estimation technique is discussed. Simulation and experimental results have shown its effectiveness in signal decomposition. Additionally two different models, i.e., chirplet model and Gabor model are compared in terms of energy compact in time-frequency representation of ultrasonic NDE signals.
In ultrasonic nondestructive evaluation, in order to successfully
detect flaw echoes corrupted by scattered random echoes, a robust and
efficient method is required. In this paper, a method utilizing
split-spectrum processing (SSP) combined with an adaptive-network-based
fuzzy inference system (ANFIS) has been developed and applied to
ultrasonic signals to perform the signal classification task. SSP can
display signal diversity and is therefore able to provide the signal
feature vectors for signal classification. ANFIS maps signal feature
vectors to outputs according to an adaptive learning process and fuzzy
If-Then rules. The combination of SSP and ANFIS can perform both
ultrasonic flaw detection and signal classification. The SSP-ANFIS
method has been tested using both simulated and experimental ultrasonic
signals, and the results show that SSP-ANFIS has good sensitivity in
detecting ultrasonic flaw echoes in the presence of strong clutter when
the signal-to-noise ratio is about zero dB
In ultrasonic non-destructive evaluation, the tasks of flaw detection and characterization in polycrystalline materials are inhibited by grain noise. An estimate of the average power spectrum of the noise can be useful in assessing the probability of flaw detection and in suppressing the noise in order to enhance flaw detection and characterization. In this paper, a model-based approach is presented for estimating the average power spectrum associated with backscattered grain noise. The approach allows grain noise measurements made at one measurement system configuration to be used as a basis for estimating the noise power spectrum for different measurement system configurations. The modelling approach determines a noise power spectrum estimate by combining an estimate of the material's longitudinal-wave backscatter coefficient with the distributed scatterer measurement system response function for the measurement system configuration of interest. Power spectrum estimates are presented for normal incidence testing, which demonstrates the capability of the approach to handle variations in transducer type (planar versus focussed) water path and depth of penetration into the material.
This letter describes a novel algorithm that is based on autoregressive decomposition and pole tracking used to recognize two patterns of speech data: normal voice and disphonic voice caused by nodules. The presented method relates the poles and the peaks of the signal spectrum which represent the periodic components of the voice. The results show that the perturbation contained in the signal is clearly depicted by pole’s positions. Their variability is related to jitter and shimmer. The pole dispersion for pathological voices is about 20% higher than for normal voices, therefore, the proposed approach is a more trustworthy measure than the classical ones.
In this paper, it is presented an insight of the effects of the application of autoregressive (AR) decomposition and pole tracking to voice signals. The AR model is used to decompose the signals in a set of poles which has a correspondence to the peaks of the signals power spectral density function (PSD). The aim of this work is to show the differences in the behavior of these poles for voice signals collected from two groups of people, one with healthy glottal tract and another with nodule pathology in vocal folds.
The differential diagnosis among the diseases that may simulate retinoblastoma is supported by several techniques; however, none of them can give an undeniable answer. In the authors' opinion the tissue characterization by power spectrum analysis of the radiofrequency ultrasound data may play an important role in the backscattered signal spectrum while being sensitive to the spectral shift trend. In order to evaluate the patterns of regression of quiescence of retinoblastoma foci after conservative treatment the authors try to create a model of tissue characterization that provides information not available in conventional A & B scan ultrasonography about cell type, vascularization and necrosis.
This article presents a new approach to the problem of obtaining topological maps for tissue characterization, based on spectral parameters extracted from radio frequency (RF) backscattered ultrasonic signals. The spectral parameter we deal with is the power spectral density centroid, since it is an efficient indicator of the tissue microstructure characteristics as far as the particle dimensions are concerned. The spectral analysis of RF ultrasonic echoes is performed using a recursive least-squares scheme with a variable forgetting factor, based on low-order autoregressive models. The proposed technique is particularly tailored to the differentiation of ocular pathologies; moreover, it is capable of tracking the spatial high-varying signal characteristics. The proposed approach was tested on simulated signals and on a gel suspension of calibrated latex spheres; finally, it was applied to signals scattered by in vitro eye specimens, giving satisfactory results in terms of frequency resolution and computational efficiency. The reduced computational burden allows an on-line implementation of the procedure. Topological spectral maps, combined with the conventional B-mode display, may offer a complete and integrated diagnostic tool, able to locally characterize the investigated tissue region in terms of amplitude and frequency shift of the corresponding echoes.
In a companion paper [T. A. Bigelow and W. D. O'Brien Jr., J. Acoust. Soc. Am. 116, 578 (2004)], theory, supported by simulations, showed that accurate scatterer size estimates could be obtained using highly focused sources provided that the derived generalized attenuation-compensation function was used and the velocity potential field near the focus could be approximated as a three-dimensional Gaussian. Herein, the theory is further evaluated via experimental studies. A calibration technique is developed to find the necessary equivalent Gaussian dimensions for a focused source using reflections obtained from a rigid plane scanned through the focus. Then, the theoretical analysis of focused sources is validated experimentally using three spherically focused ultrasound transducers to estimate the radius of glass beads imbedded in tissue mimicking phantoms. Both the impact of focusing (f/1, f/2, and f/4) and the effect of scatterer type (comparing glass bead results to simulation results that used scatterers with Gaussian impedance distributions) were tested. The simulated differences agree with the measured differences to within 2.5% provided that the comparison is made between the same scatterer type and sources with the same equivalent Gaussian dimensions. The improvement provided by the generalized attenuation-compensation function is greatly influenced by the type of scatterer whose size is being estimated and decreases as the wavelength dependence of the Gaussian depth of focus is reduced.
In this paper, we present hardware realization of a reconfigurable ultrasonic flaw detection system (RUFD) for real-time applications. It can be utilized in diverse environments and applications due to the run-time reconfiguration capability based on the parameters of the ultrasonic transceiver and the target material. Furthermore, this paper presents FPGA implementation results and discusses the steps for optimizing and enhancing the system in terms of the device resources usage and ultimately instrument production cost.
In ultrasonic grain size characterization, grain signals are the only tangible results that can be obtained directly from grain scattering in the specimen. Grain scattering results in an upward shift in the expected frequency of broadband, returned ultrasonic echoes, while the attenuation effect influences the frequency shift in a downward direction. Both the upward and downward shifts are closely related to the grain structures of the materials. In this paper, the second order linear predictive method is used to characterize the spectral shift by utilizing of features such as resonating frequency, system poles and linear predictive coefficients. The feasibility of applying pattern recognition techniques based on these features are discussed and supported by simulated computer and experimental results.
Methods of time series analysis based on the G. E. Box- G. M. Jenkins method were used to analyze random variations and inherent variability in geotechnical test data. The paper presents a summary of the time series models and their application to the analysis of data from unconfined compression tests, field vane tests, and cone penetration tests. The integrated-moving-average-autoregressive model was found to be a versatile tool and served this purpose well. The model permits the determination of the autocavariance function and the random testing error that best fit the test data.
Insonification of the microstructure of materials results in a backscattered signal consisting of multiple interfering echoes with random amplitudes and phases. Information pertaining to grain scattering cross section and grain size distribution is an inherent property of the backscattered signal. A statistical model of grain signals is developed that describes the spatial and time averaged data, and their relationship to signal attenuation caused by scattering and absorption. Both spatial and temporal averaging permits the estimation of the attenuation coefficient, which has been found to be position dependent. Furthermore, it has been shown experimentally and theoretically that the performance of spatial and time averaging is governed by correlation properties of the grain signal.
Grain size characterization using ultrasonicbackscattered signals is an important problem in nondestructive testing of materials. In this paper, a heuristic model which relates the statistical characteristics of the measured signal to the mean ultrasonicwavelet and attenuation coefficient in different regions of the sample is investigated. The losses in the backscattered signal are examined using temporal averaging, correlation, and probability distribution functions of the segmented data. Furthermore, homomorphic processing is used in a novel application to estimate the mean ultrasonicwavelet (as it propagates through the sample) and the frequency‐dependent attenuation. In the work presented, heat‐treated stainless steel samples with various grain sizes are examined. The processed experimental results support the feasibility of the grain size evaluation techniques presented here using the backscattered grain signal.
The center frequency of a narrowband, discrete-time random process, such as a reflected ultrasound signal, is estimated from the parameter values of a reduced, second-order autoregressive (AR) model. This approach is proposed as a fast estimator that performs better than the zero-crossing count estimate for determining the center-frequency location. The parameter values are obtained through a linear prediction analysis on the correlated random process, which in this case is identical to the maximum entropy method for spectral estimation. The frequency of the maximum of the second-order model spectrum is determined from these parameters and is used as the center-frequency estimate. This estimate can be computed very efficiently, requiring only the estimates of the first three terms of the process autocorrelation function. The bias and variance properties of this estimator are determined for a random process having a Gaussian-shaped spectrum and compared to those of the ideal FM frequency discriminator, zero-crossing count estimator and a correlation estimator. It is found that the variance values for the reduced-order AR model center-frequency estimator lie between those for the ideal FM frequency discriminator and the zero-crossing count estimator.
This paper gives an exposition of linear prediction in the analysis of discrete signals. The signal is modeled as a linear combination of its past values and present and past values of a hypothetical input to a system whose output is the given signal. In the frequency domain, this is equivalent to modeling the signal spectrum by a pole-zero spectrum. The major part of the paper is devoted to all-pole models. The model parameters are obtained by a least squares analysis in the time domain. Two methods result, depending on whether the signal is assumed to be stationary or nonstationary. The same results are then derived in the frequency domain. The resulting spectral matching formulation allows for the modeling of selected portions of a spectrum, for arbitrary spectral shaping in the frequency domain, and for the modeling of continuous as well as discrete spectra. This also leads to a discussion of the advantages and disadvantages of the least squares error criterion. A spectral interpretation is given t
Thesis (Ph. D.)--Illinois Institute of Technology, 1987. Includes bibliography.
Spectral analysis of ultrasonic reflections from biological tissues can be used to determine basic tissue parameters for use in differential diagnosis. This paper describes the use of the technique under circumstances encountered in several types of clinical examinations. The applications are illustrated with results obtained from laboratory measurements with a system now being employed in a clinical evaluation programme. The test objects studied simulate tissues with planar boundaries, tissues with heterogeneous interior structure, and tissues causing acoustic 'shadowing' of posterior regions.
The theoretical relationship between center frequency downshift and the spectral bandwidth was investigated for pulses with a sinc(x) spectrum propagating through lossy media. Power law and exponential models for frequency dependence of attenuation were used. Six target materials encompassing a range of attenuation parameters were used to verify the theoretical model. The frequency downshift data from these materials was used to calculate their respective attenuation parameters. It was shown theoretically and verified experimentally that for small frequency downshifts, the sinc(x) model yields the same material parameters as the Gaussian model. The choice of the model for the attenuation of the material was found to be inconsequential.
Short-time Fourier analysis is well suited for processing tissue echographic signals which are nonstationary. We have investigated the use of short-time Fourier analysis to provide an estimation of the echographic spectral composition as a function of time. It will be shown that the time dependence of the spectral centroid of this representation allows one to deduce easily the frequency-dependent attenuation. A simple correction of the noninvariant filtering effect due to diffraction is used to unbias the attenuation slope estimation. This new signal processing technique was first tested on simulated echographic data from a 1-D tissue model. Experimental results obtained from echo signals on a tissue-like phantom and on in vivo liver tissue show the influence of diffraction and attenuation respectively.
A model for the grain signal is presented, which includes the effect of frequency-dependent scattering and attenuation. This model predicts that the expected frequency increases with scattering and decreases with attenuation. Homomorphic processing was used for spectral smoothing, and the selection of parameters for optimal performance was examined. Experimental results are presented that show both the upward shift in the expected frequency with grain boundary scattering and the downward shift with attenuation. Furthermore, it is shown that the expected frequency shift can be correlated with the grain size of the material. It is important to point out that the quantitative relationship between the average grain size and the expected frequency shift (either upward or downward) is dependent on the type of material, the quality of grain boundaries, and the characteristics of the measuring instruments.< >
A review of recent developments in radar signal processing in the presence of clutter is presented, with particular reference to an air-traffic environment. Two different signal processing issues are considered. The first issue relates to the adaptive suppression of clutter (of unknown statistics) and consequent enhancement of the echo produced by a moving target (e.g., aircraft). Here, several adaptive techniques are presented to respond to changes in environmental conditions. The second issue relates to the problem of identifying the sources of clutter with the aim of vectoring aircraft around hazardous areas. In both cases, results based on real-life radar data are presented to support the theory.
Ultrasound signals reflected from the human liver contain significant information about its condition. Several techniques have been developed to process such signals. In this paper, two approaches are described for estimating the slope of the acoustic attenuation coefficient, denoted by β, which provides important cues about the condition of diffuse liver disease. The first is a nonparametric approach which determines the β value from the slope of log periodogram differences. The second is a parametric approach which requires a Gaussian-shaped spectrum and determines the β value from s shift in the spectral center-frequency. A zero-crossing count procedure allows the parametric approach to be implemented with minimal hardware and in real-time. Simulated and actual signals illustrate the two approaches.
A summary of many of the new techniques developed in the last two decades for spectrum analysis of discrete time series is presented in this tutorial. An examination of the underlying time series model assumed by each technique serves as the common basis for understanding the differences among the various spectrum analysis approaches. Techniques discussed include the classical periodogram, classical Blackman-Tukey, autoregressive (maximum entropy), moving average, autotegressive-moving average, maximum likelihood, Prony, and Pisarenko methods. A summary table in the text provides a concise overview for all methods, including key references and appropriate equations for computation of each spectral estimate.
Quantitative grain size evaluation Tao Wang was born in Sijiazhuang, Hebei He received the B.S. degree in electrical engineering from the University of Science and Technology of China
- J Saniie
- N M Bigutay
J. Saniie and N. M. Bigutay, " Quantitative grain size evaluation
Tao Wang was born in Sijiazhuang, Hebei,
China on November 4, 1965. He received the
B.S. degree in electrical engineering from the
University of Science and Technology of
China, Hefei, Anhui, China, in 1982, and the
M.S. and Ph.D. degrees, both in electrical engineering, from the Illinois Institute of Technology, in 1984 and 1987, respectively.
From 1988 to 1989, he was a postdoctoral
research fellow with the p g artment of Electrical and Computer Engirtegring at Illinois Inusing ultrasonic backscattered echoes, " J Acoust Soc Amer,
vol 80, pp 1816-1824, Dec 1986
Application of linear predictive analysis in ultrasonics grain signal characterization Review of Progress in Quantitative Nondestructive Jafar Saniie (S'80-M'Sl) was born in Iran on Evaluation Ultrasonic signal processing and pattern recognition Maryland, College Park
- J Saniie
- Wang
- N Jin
- Bilgutay
J Saniie, T Wang, X Jin, and N M Bilgutay, " Application
of linear predictive analysis in ultrasonics grain signal characterization, " Review of Progress in Quantitative Nondestructive
Jafar Saniie (S'80-M'Sl) was born in Iran on
Evaluation, vol 8, D 0. Thompson and D. E Chimenti, Eds
March 21, 1952 He received the B S degree
New York Plenum, 1989
in electrical engineering from the Univeraity of
T Wang, " Ultrasonic signal processing and pattern recognition
Maryland, College Park, 1974 In 1977 he rein evaluating the microstructure of materials, " Ph D dissertaceived the M S degree in biomedical engition, Illinois Institute of Technology, Chicago, 1987
neering from Case Western Reserve
he was a postdoctoral research fellow with the p g artment of Electrical and Computer Engirtegring at Illinois Inusing ultrasonic backscattered echoes
- J Saniie
- N M Bigutay
J. Saniie and N. M. Bigutay, "Quantitative grain size evaluation
Tao Wang was born in Sijiazhuang, Hebei,
China on November 4, 1965. He received the
B.S. degree in electrical engineering from the
University of Science and Technology of
China, Hefei, Anhui, China, in 1982, and the
M.S. and Ph.D. degrees, both in electrical engineering, from the Illinois Institute of Technology, in 1984 and 1987, respectively.
From 1988 to 1989, he was a postdoctoral
research fellow with the p g artment of Electrical and Computer Engirtegring at Illinois Inusing ultrasonic backscattered echoes," J Acoust Soc Amer,
vol 80, pp 1816-1824, Dec 1986
Linear Prediction ofspeech Berlin
- J Markel
- A H Gray
J D Markel and A. H. Gray, Jr, Linear Prediction ofspeech
Berlin. Springer-Verlag, 1982
Ultrasonic signal processing and pattern recognition Maryland, College Park, 1974 In 1977 he rein evaluating the microstructure of materials
Evaluation, vol 8, D 0. Thompson and D. E Chimenti, Eds
March 21, 1952 He received the B S degree
New York Plenum, 1989
in electrical engineering from the Univeraity of
T Wang, "Ultrasonic signal processing and pattern recognition
Maryland, College Park, 1974 In 1977 he rein evaluating the microstructure of materials," Ph D dissertaceived the M S degree in biomedical engition, Illinois Institute of Technology, Chicago, 1987
neering from Case Western Reserve Univer-
Ultrasonic signal prosity, Cleveland, OH, and, in 1981, the Ph D cessing for in vivo attenuation measurement
- M Fink
- F Hottier
- J F Cardoso
M Fink, F Hottier and J F. Cardoso, "Ultrasonic signal prosity, Cleveland, OH, and, in 1981, the Ph D
cessing for in vivo attenuation measurement," Ultrason Imagdegree in electrical engineering from Purdue
ing, vol. 5, pp 117-135, 1983
Since StitUte of Technology Since October i989, hehas been an Ultrasonic Systems Scientist with Physical Acoustics Corporation His research interests are digital signal processing, ultrasonic imaging, statistical pattern recognition and communication systems
- P A Narayana
- J Ophir
P. A. Narayana and J Ophir, "Spectral shifts of ultrasonic prop-In 1981 he joined the Applied Physics Labagation' A study of theoretical and experimental methods," Uloratory, University of Helsinki, Finland, to
trason Imanznn. vol 5. DD 22-29. 1983
conduct research in ohotothermal and uhotoacoustic Imaging. Since
StitUte of Technology Since October i989, hehas been an Ultrasonic
Systems Scientist with Physical Acoustics Corporation His research
interests are digital signal processing, ultrasonic imaging, statistical
pattern recognition and communication systems.
vol. 84, pp. 400-407, July 1988
vol. 63, pp 561-580, 1975
Grain size estimation by the intercept method
- J E Hilliand
J. E. Hilliand, "Grain size estimation by the intercept method,"
Northwestern University, Dept. Materials Science and Materials
Research Center (Int. Rep.), Evanston, IL, Nov. 1963.