Search

The ultrasound phantoms used to educate medical students should not only closely mimic the ultrasound characteristics of human soft tissues but also be inexpensive and easy to manufacture. I have been studying handmade ultrasound phantoms and proposed an ultrasound phantom comprising calcium alginate hydrogel that met these requirements but caused a speckle pattern similar to that observed in ultrasound images of liver. In this study, I show that adding ethanol to the precursors used to fabricate the phantom reduces the speckle pattern. The ultrasound propagation velocity and attenuation coefficient of the phantom were 1561 ± 8 m/s and 0.54 ± 0.18 dB/cm/MHz, respectively (mean ± standard deviation), which are within the ranges of those in human soft tissues (1530-1600 m/s and 0.3-1.0 dB/cm/MHz, respectively). This phantom is easy to fabricate without special equipment, is inexpensive, and is suitable for elementary training on ultrasound diagnosis, operation of ultrasound-guided needles, and blind catheter insertion.

After 100 years of theoretical treatment of speckle patterns from coherent illumination, there remain some open questions about the nature of ultrasound speckle from soft vascularized tissues. A recent hypothesis is that the fractal branching vasculature is responsible for the dominant echo pattern from organs such as the liver. In that case an analysis of cylindrical scattering structures arranged across a power law distribution of sizes is warranted. Using a simple model of echo strength and basic transformation rules from probability, we derive the first order statistics of speckle considering the amplitude, the intensity, and the natural log of amplitude. The results are given by long tailed distributions that have been studied in the statistics literature for other fields. Examples are given from simulations and animal studies, and the theoretical fit to these preliminary data support the overall framework as a plausible model for characterizing ultrasound speckle statistics., 26 pages, 8 figures

The short-lag spatial coherence (SLSC) imaging has been demonstrated to be advantageous over the traditional B-mode ultrasound imaging. With focused scanning beams, the SLSC imaging has an excellent performance in clutter reduction and lesion detection, especially in the low signal-to-noise ratio (SNR) scenarios. The synthetic aperture (SA) imaging is an appropriate mode for the SLSC imaging as the dynamic transmit focusing could keep a good focusing quality at any depth. However, the SLSC image may still suffer a bad resolution performance when a low lag value is used in the coherence summation to ensure the contrast enhancement. In this paper, an adaptive synthetic transmit (Tx) aperture focusing strategy is proposed for the SLSC imaging with the SA mode. Based on the achievements of adaptive beamforming, a minimum variance beamformer is applied in the Tx aperture to realize adaptive focusing. Spatial coherence is then measured in the receive aperture to form the SLSC image. Simulation and experimental studies were conducted to evaluate the proposed method. Experiments showed that the proposed method not only improved the poor resolution of the original SLSC image but also enhanced the speckle performance, which led to increased contrast-to-noise ratio and speckle SNR values.

Obtaining accurate ultrasonically estimated displacements along both axial (parallel to the acoustic beam) and lateral (perpendicular to the beam) directions is an important task for various clinical elastography applications (e.g., modulus reconstruction and temperature imaging). In this study, a partial differential equation (PDE)–based regularization algorithm was proposed to enhance motion tracking accuracy. More specifically, the proposed PDE-based algorithm, utilizing two-dimensional (2D) displacement estimates from a conventional elastography system, attempted to iteratively reduce noise contained in the original displacement estimates by mathematical regularization. In this study, tissue incompressibility was the physical constraint used by the above-mentioned mathematical regularization. This proposed algorithm was tested using computer-simulated data, a tissue-mimicking phantom, and in vivo breast lesion data. Computer simulation results demonstrated that the method significantly improved the accuracy of lateral tracking (e.g., a factor of 17 at 0.5% compression). From in vivo breast lesion data investigated, we have found that, as compared with the conventional method, higher quality axial and lateral strain images (e.g., at least 78% improvements among the estimated contrast-to-noise ratios of lateral strain images) were obtained. Our initial results demonstrated that this conceptually and computationally simple method could be useful for improving the image quality of ultrasound elastography with current clinical equipment as a post-processing tool.

To achieve the fast and accurate segmentation of ultrasound image, a novel edge detection method for speckle noised ultrasound images was proposed, which was based on the traditional Canny and a novel multiplicative gradient operator. The proposed technique combines a new multiplicative gradient operator of non-Newtonian type with the traditional Canny operator to generate the initial edge map, which is subsequently optimized by the following edge tracing step. To verify the proposed method, we compared it with several other edge detection methods that had good robustness to noise, with experiments on the simulated and in vivo medical ultrasound image. Experimental results showed that the proposed algorithm has higher speed for real-time processing, and the edge detection accuracy could be 75% or more. Thus, the proposed method is very suitable for fast and accurate edge detection of medical ultrasound images.

Ultrasonography is often preferred over the other medical imaging modalities due to its noninvasive nature, cost-effectiveness, and portability. However, the resolution of the ultrasound image greatly depends upon the presence of speckle noise. Speckle noise generally tends to reduce the image resolution and contrast, thereby reducing the diagnostic resolution of this imaging modality. In this paper, we propose a speckle suppressing anisotropic diffusion (SSAD) filter, to remove the speckle noise from B-Mode Ultrasound images. The performance of the SSAD filter is compared with the existing diffusion filters. The evaluation is based on their application to images simulated by Field II (developed by Jensen et al.). The algorithms were also tested for clinical ultrasound images of polycystic ovaries obtained from HDI 5000 Ultrasound Scanner. Performance evaluation was done by both numerical and functional parameters. The proposed filter yields better results in terms of greatest structural similarity index map (SSIM) of 0.95 and accuracy of 99.5.

In ultrasound-based elastography methods, the estimation of shear wave velocity typically involves the tracking of speckle motion due to an applied force. The errors in the estimates of tissue displacement, and thus shear wave velocity, are generally attributed to electronic noise and decorrelation due to physical processes. We present our preliminary findings on another source of error, namely, speckle-induced bias in phase estimation. We find that methods that involve tracking in a single location, as opposed to multiple locations, are less sensitive to this source of error since the measurement is differential in nature and cancels out speckle-induced phase errors.

Ultrasound is one of the most important medical imaging modalities for its real-time and portable imaging advantages. However, the contrast resolution and important details are degraded by the speckle in ultrasound images. Many speckle filtering methods have been developed, but they are suffered from several limitations, difficult to reach a balance between speckle reduction and edge preservation. In this paper, an adaptation of the nonlocal total variation (NLTV) filter is proposed for speckle reduction in ultrasound images. The speckle is modeled via a signal-dependent noise distribution for the log-compressed ultrasound images. Instead of the Euclidian distance, the statistical Pearson distance is introduced in this study for the similarity calculation between image patches via the Bayesian framework. And the Split-Bregman fast algorithm is used to solve the adapted NLTV despeckling functional. Experimental results on synthetic and clinical ultrasound images and comparisons with some classical and recent algorithms are used to demonstrate its improvements in both speckle noise reduction and tissue boundary preservation for ultrasound images.

When spatial compounding is applied to targets with significant acoustic velocity inhomogeneities, the correlation between speckle patterns of the images to be averaged decreases, thereby increasing the speckle reduction nominally obtained. Phase correction applied to these targets improves the coherence of the wavefield and restores image spatial frequencies. Combining these two modes can be used to effectively increase the contrast-to-noise ratio (CNR) of imaging targets and improve the general image quality of these targets over spatial compounding alone. This paper presents a clinical evaluation of combined spatial compounding and adaptive imaging in breast tissue and compares this combined technique to conventional imaging and to adaptive imaging and spatial compounding operating independently. Experiments were performed on a 1.75-D, 8 × 96 array attached to a commercially-available scanner. Cysts, microcalcifications and other breast structures were targeted in order to assess the impact of the combined mode on CNR, target width, target brightness and target peak-to-background ratio (PBR). In general, phase correction improved cyst CNR by 7.7%, decreased target width by 18.7%, increased target brightness by 30.1% and increased PBR by 17.9%. Compounding alone, using three overlapping 9.71 mm subapertures, increased cyst CNR by 24.6%, but increased target width by 25.4% and decreased PBR by 13.2%. Combining both modes, however, increased cyst CNR by 32.6%, inappreciably increased target width by 1.1% and marginally decreased PBR by 2.8%. The increase in target brightness with this combined mode was 20.0%

An efficient speckle tracking algorithm is proposed for motion estimation in ultrasonic imaging. Speckle tracking involves a matching process and a searching process. The matching process of the proposed algorithm is based on a Block Sum Pyramid algorithm that significantly reduces the computational complexity while maintaining the same accuracy as the conventional sum of absolute difference approach. The searching process, on the other hand, is based on a multilevel search strategy rather than the full-search strategy used by most conventional tracking methods. Both simulated speckle images and clinical breast images were used to test the performance of the proposed algorithm. The results show that the computation efficiency is improved by up to a factor of five over the conventional approach. The improved efficiency enables real-time or near-real-time implementation of motion estimation in ultrasonic imaging, which is particularly beneficial in areas such as blood flow estimation, elasticity imaging, speckle image registration, and strain compounding.

Various sources of variability, such as speckle noise, depth dependence and inhomogeneous intervening tissue, are involved in B-mode images, even when using the same ultrasonic equipment with fixed settings. The behavior of these sources of variability was investigated by texture analysis of images obtained from simulations and from a tissue-mimicking phantom, a normal adult liver and a pediatric renal (Wilms') tumor. First-order statistics (MEAN and SNR) and second-order statistics from the co-occurrence matrix (ENT and COR) were calculated. In a phantom, the SNR and ENT show a clear depth dependence. In biological tissue, the variability is mainly caused by the speckle noise and inhomogeneous intervening tissue. In addition, almost the entire range of the COR feature is present in images of liver and tumor. Furthermore, all the features calculated in windows of 1 cm2 exhibit an overlap among the different media. With the second-order features, it is possible to discriminate 85% reliable (average) between the normal, adult, liver and the pediatric renal tumor above a window size of 9 cm2. The SNR can not discriminate between these tissues. The maximum resolution of 9 cm2 reveals a serious limitation of parametric imaging. Finally, the features reproduce well in the case of follow-up of an abdominal tumor during chemotherapy.

Effects of spatial compounding on image resolution and speckle noise are studied. Using computer simulation, it is shown that spatial compounding using averaged reconstruction can significantly improve lateral resolution while slightly deteriorate axial resolution. The amount of net resolution improvement depends mainly on the compound angle, but is insensitive to the number of component images used in compounding. While the fact that spatial compounding can effectively reduce speckle noise is well known, the analysis in this paper indicates that to maximize speckle reduction, the component echo amplitudes must meet two conditions: to be mutually independent and to have the same mean power. These findings provide useful guidelines for the analysis and optimization of the performance of an ultrasound scanning system that has been specially developed for imaging residual limbs.

In coronary arteries, knowing the relative stiffness of atherosclerotic lesions can help physicians select the most appropriate therapeutic modality. Because soft material supports larger strains than hard, measurements of this quantity can distinguish tissue of differing stiffness. In a previous paper, we described techniques for computing displacements and strains in coronary arteries using an integrated angioplasty and imaging catheter. Here, we demonstrate that hard and soft materials in a tissue-mimicking phantom can be differentiated with this device. Because tissue motion cannot be distinguished from catheter motion a priori, we perform all computations in the coordinate system centered at the balloon's geometric center. This reference frame depends only on balloon shape and is independent of catheter motion. A specialized correlation-based, phase-sensitive speckle tracking algorithm has been developed to compute strain. Maximum phantom displacement was about 25 μm, and the maximum radial, normal strain was about 1.5 percent.

Spatial compounding has long been explored to reduce coherent speckle noise in medical ultrasound. By laterally translating a one-dimensional array, partially correlated measurements made at different look directions can be obtained and incoherently averaged. The lateral resolution, however, is limited by the sub-array length used for each independent measurement. To reduce speckle contrast without compromising lateral resolution, a new spatial compounding technique using two-dimensional, anisotropic arrays is proposed. This technique obtains partially correlated measurements by steering the image plane elevationally with small inclinations. Incoherent averaging is then performed by adding image magnitudes. Therefore, contrast resolution is improved only at the price of a slightly wider elevational beam. Note that although anisotropic arrays have limited steering capability in elevation, grating lobes are not considered influential since only small inclinations are needed between measurements. Simulations have been performed to show both the change in spatial resolution and the improvement in contrast resolution. Results indicated minimal increase in the correlation length both laterally and axially. It was also shown that detectability can be significantly enhanced by increasing the number of measurements or increasing the differential inclination between measurements. This technique is therefore effective for reducing speckle noise while maintaining in-plane spatial resolution. Furthermore, it demonstrates a new application of two-dimensional anisotropic arrays in spite of their limited elevational steering capability.

The goal of myocardial tissue characterization is to augment information provided by two- and three-dimensional echocardiographic imaging, Doppler blood flow and speckle- or Doppler-derived tissue motion. Tissue characterization based on the systematic variation of backscattered ultrasound during the cardiac cycle (‘cyclic variation’) appears to be effective in characterizing both focal and diffuse myocardial pathologies. Unfortunately, comparison of results from different laboratories is difficult because of a lack of consistency among the several reported methods of analyzing the cyclic variation data. The goals of the present work are to present an improved method of analysis and to demonstrate that apparent disagreements are attributable primarily to the distinct approaches employed by different investigators. The improved automated method for determining the magnitude of cyclic variation utilizes binomial smoothing and an average deviation method and was validated using data acquired from 23 patients. This method illustrates a systematic means for resolving differences between laboratories. This resolution facilitates future comparisons between the cyclic variation of myocardial backscatter and measurements derived, for example, from strain-related approaches.

An investigation into statistical properties of ultrasonic image texture from three-dimensional clusters of anisotropic scatterers is carried out. The structural properties of the clusters are modeled after those of soft biological tissues, such as skeletal muscle tissues, both in their healthy condition and at the early stage of degenerative diseases. The average axial autocorrelation function of the intensity of the image texture is used to characterize and monitor changes of the geometrical properties of the tissue components. A distinct local increase of the autocorrelation is observed within a range of small time shifts, and it is explained in terms of the structure of the time-domain backscattered signal from each individual scatterer. It is shown that such an increase is sensitive to structural variations of the cluster similar to those occurring at the early stage of several muscular diseases.

The application of fractal analysis to parametric imaging of B-mode echograms and to differentation of echographic speckle textures was investigated. Echograms were obtained from realistic simulations and from a clinical study on diffuse liver disease. The simulations comprised tissue models with randomly positioned scatterers in a 3-D volume in which the number density was varied over a range from 0.5 to 25 mm–3. The clinical echograms comprised both normals and patients with liver cirrhosis. Three methods of estimating the fractal dimension were investigated, two in the spatial image domain and one in the spatial frequency domain. The results of these methods are compared and the applicability and the limitations of texture differentiation using fractal analysis is discussed. The main conclusion is that fractal analysis offers no obvious advantage over statistical analysis of the texture of echographic images. Its use for parametric imaging is further limited by the need to use relatively large windows for local estimation of the fractal dimension.

A technique has been developed for measuring the two in-plane components of blood velocity using ultrasound. The principle is based on analyzing a time sequence of successive images of the speckle pattern, and forming spatial projections of the resulting three-dimensional data set. The mathematics may be understood in terms of sections through a three-dimensional Fourier transform of the original data. The technique has been tested with data from a computer-controlled scanner using a tissue-mimicking phantom, and with data from a commercial ultrasound scanner using images of flowing blood in a jugular vein. In both cases, velocity vectors with many different orientations were measured, demonstrating the feasibility of the method. A less restrictive real-time implementation is suggested.

An objective measure (Lesion Signal-to-Noise Ratio) quantifying the detectability of lesions in echographic images was employed. This measure was used to determine the performance of digital speckle reduction filters, which were applied to computer simulated ultrasound B-mode images. One linear (mean filter) and two nonlinear filters (median and L2-mean filters) have been investigated. A comparison was made between fixed and adaptive versions of these filters. The influence of the size of the filter window on the Lesion Signal-to-Noise Ratio was systematically investigated. Also, the effect of the shape of the filter window is illustrated. The difference in performance of the linear and nonlinear filters was found to be small. Adaptive filters did not perform significantly better than fixed filters. The maximum improvement of lesion detectability was in the order of 40 percent. The choice of a correct window size was critical. For all types of filters, an optimum window size appeared to be present in the curves relating the Lesion Signal-to-Noise Ratio to this size.

As speckle on ultrasonic B-scan may reveal information relative to tissue structure, the present work attempts to discriminate the various prostatic tissues (normal tissue, benign prostatic hypertrophy and cancer) by means of texture analysis. We select regions of interest by their homogeneous appearance. A pre-processing stage is required to obtain stationary samples. The method used measures the second-order statistics, namely co-occurrence matrices. Fairly good tissue signatures have been obtained with parameters derived from these matrices. Of 37 images processed, 78 percent of the samples were classified with success, which is a high score considering that the images cannot be discriminated visually. However, while such results are obtained when wide regions of interest are investigated (64 × 64 pixels), they are not as good with smaller sample sizes, i.e., when the pathological area is very small.

This paper describes an efficient computer method for generating three-dimensional (3-D) ultrasonic speckle patterns for sector scans of an annular array with dynamic focusing in both transmit and receive. Assuming random scatterers in an attenuating medium, the system synthesizes the waveform for each scan line using echo data received at all annular elements of the transducer when short pulses are transmitted by one annulus after another. The amount of echo data needed to synthesize one waveform is tremendously reduced by reducing the 3-D distributions of scatterers to line distributions and by representing scatterers on a small line segment by a single equivalent scatterer taking into account the two-way travel time differences. Furthermore, by a judicious 3-D arrangement of the scatterer lines and scan lines, it is possible to synthesize the waveform for a new scan line with negligible computation overhead. The waveforms of the scan lines are detected to obtain a B-mode image or speckle patterns. The patterns obtained on several differently-oriented image planes showed a good statistical agreement with experimentally obtained patterns.

Speckle tracking has been used for motion estimation in ultrasound imaging. Unlike conventional Doppler techniques, which are angle-dependent, speckle tracking can be utilized to estimate velocity vectors. However, the accuracy of speckle-tracking methods is limited by speckle decorrelation, which is related to the displacement between two consecutive images, and, hence, combining high-frame-rate imaging and speckle tracking could potentially increase the accuracy of motion estimation. However, the lack of transmit focusing may also affect the tracking results and the high computational requirement may be problematic. This study therefore assessed the performance of high-frame-rate speckle tracking and compared it with conventional focusing. The effects of the signal-to-noise ratio (SNR), bulk motion, and velocity gradients were investigated in both experiments and simulations. The results show that high-frame-rate speckle tracking can achieve high accuracy if the SNR is sufficiently high. In addition, its computational complexity is acceptable because smaller search windows can be used due to the displacements between frames generally being smaller during high-frame-rate imaging. Speckle decor-relation resulting from velocity gradients within a sample volume is also not as significant during high-frame-rate imaging.

A unified approach to the design of nonlinear filters for speckle suppression in ultrasound B-mode images is presented. The detection of the (lesion) signal is formulated as a binary hypothesis-testing problem. The structure of the optimal decision rules is derived both in the case where the lesion signal is assumed either a constant or random variable. In the case of a constant signal, the maximum likelihood (ML) estimator and the optimal L-estimator are derived. In the case of a random lesion signal, the maximum a posteriori probability estimator of the lesion signal has also been found. Experimental results verify the superiority of the proposed ML-estimator and the L-estimator over the straightforward choice of an arithmetic mean for speckle filtering in simulated tissue mimicking phantom ultrasound B-mode images.

Based on physical models, it is indicated that the received pressure field in ultrasound B-mode images can be described by a convolution between a tissue reflection signal and the emitted pressure field. This result is used in a description of current image formation and in formulating a new processing scheme. The suggested estimator can take into account the dispersive attenuation, the temporal and spatial variation of the pulse, and the change in reflection strength and signal-to-noise ratio. Details of the algorithm and the estimation of parameters to be used are given. The performance is indicated by two examples. One is for a synthetic signal and the other is for data measured from a tissue mimicking phantom. The last example shows a finer speckle pattern, giving an increased resolution.

Conventional ultrasonic B-mode images qualitatively describe tissue structures but are unsuitable for quantitative analyses of scatterer properties. We have recently developed an ultrasonic parametric imaging technique based on the Nakagami statistical distribution that is able to quantify scatterer concentrations. The aim of the present study is to further explore both the behavior of a Nakagami image in characterizing different scatterer structures at different signal-to-noise ratios (SNRs) and the feasibility of Nakagami imaging using a general commercial ultrasound scanner for tissue examinations. Simulations, experiments on a tissue-mimicking phantom and in vitro measurements on a muscle tissue before and after microwave treatment were carried out. The SNR and contrast-to-noise ratio (CNR) were estimated to quantify image performance. The results demonstrate that a Nakagami image can differentiate different scatterer concentrations for single, hypoechoic and hyperechoic targets. Also, a Nakagami image, when combined with an ultrasound scanner, can complement the B-scan to characterize tissue and to identify the region of interest with a larger CNR. However, the noise effect can degrade the performance of a Nakagami image. When the signal SNR decreased to 15 dB in simulations and to 8 dB in experiments, the CNR of the hyperechoic Nakagami image decreased by 4% and 27%, respectively, and that of the hypoechoic one decreased by 42% and 80%, respectively. These results indicate that a Nakagami image behaves well in identifying regions with high scatterer concentrations but does not perform well when both the scatterer concentration and SNR are low.

Breast lesions that may be cancerous generally possess a higher elastic modulus than the surrounding tissue. Thus, they are sensed to move differently from the surrounding tissue during finger-based palpation. Since the palpation is subjective in nature, a method using ultrasound for detection of the differential motion associated with hard lesions is proposed. It is intended that regions of detected ‘lump’ be highlighted on top of a conventional B-mode image so that the user can positively associate palpated lumps with specific regions on the anatomical B-mode image. We detect the differential motion associated with a lesion by detecting a local abnormality in a map of image-to-image correlations as the transducer is swept in an elevational direction. Experiments were performed with three different tissue mimicking phantoms. The results indicate that the location of the mobile lesion can be successfully estimated whereas immobile lesions are not very detectable. This approach is simple and requires only a short period of time for each scan. It has practically no theoretical incremental cost.

A new imaging method, microwave-induced thermal imaging (MITI), was developed to differentiate tissue based on thermal and dielectric properties. Image contrast depends on temporal strain in tissue, which was determined by one-dimensional speckle tracking using a phase-sensitive, correlation-based technique. The underlying mechanisms were analyzed and experimental results on biologic tissue agreed well with theoretical predictions. Because of its strong contrast between water-bearing and lipid-bearing tissue, the technique may enhance existing intravascular ultrasound (IVUS) imaging systems to identify vulnerable arterial plaque.

A method for adaptive clutter rejection via blind source separation (BSS) using principal and independent component analyses is presented in application to blood velocity measurement in the carotid artery. In particular, the filtering method's efficacy for eliminating clutter and preserving lateral blood flow signal components is presented. The performance of IIR filters is compromised by shorth data ensembles (10 to 20 temporal samples) as implemented for color-flow and high frame-rate imaging due to initialization requirements. Further, the ultrasonic imaging system's transfer function maps axial wall and lateral blood motion to overlapping spectra. As such, frequency domain-based approaches to wall filtering are ineffective for distinguishing wall from blood motion signals. Rather than operating in the frequency domain, BSS performs clutter rejection by decomposing the input data ensemble into N constitutive source signals in time, where N is the ensemble length. Source signal energy coupled with respective signal depth and time course profiles reveal which source signals correspond to blood, noise and clutter components. Clutter components may then be removed without disruption of lateral blood flow information needed for two-dimensional blood velocity measurement. A simplistic data simulation is employed to offer an intuitive understanding of BSS methods for signal separation. The adaptive BSS filter is further demonstrated using a Field II simulation of blood flow through the carotid artery including tissue motion. BSS clutter filter performance is compared to the performance of FIR, IIR and polynomial regression clutter filters. Finally, the filter is employed for clinical application using a Siemens Elegra scanner, carotid artery data with lateral blood flow collected from healthy volunteers, and Speckle Tracking; velocity magnitude and angle profiles are shown. Once again, the BSS clutter filter is contrasted to FIR, IIR and polynomial regression clutter filters using clinical examples. Velocities computed with Speckle Tracking after BSS wall filtering are highest in the center of the artery and diminish to low velocities near the vessel walls, with velocity magnitudes consistent with physiological expectations. These results demonstrate that the BSS adaptive filter sufficiently suppresses wall motion signal for clinical lateral blood velocity measurement using data ensembles suitable for color-flow and high frame-rate imaging.

3D positioning is essential for quantitative volume analysis in 3D ultrasound. Particularly for freehand scanning without an external positioning device, such information must be estimated by analyzing the original 2D images prior to 3D reconstruction. Previous work on freehand 3D positioning has focused on elevational displacement estimation using speckle decorrelation for linear scans. However, the effects of other types of motion have been ignored. Given that all types of motion potentially introduce speckle decorrelation, the accuracy of the elevational displacement estimation is likely to be diminished by the presence of motions of other types. In the present study, simulations were performed to probe the effects of various motions on elevational displacement estimation. In particular, the effects of rotational motions on image correlation are investigated in detail. It is found that these motions significantly affect the estimation results and thus should be taken into account when reconstructing the 3D image. In addition, speckle variations also affect the estimation even when only elevational motion is present. Finally, full motion analysis in freehand scanning may not be possible by only using speckle correlation analysis unless speckle variations can be reduced and the correlation distribution under complex motion can be obtained.

Speckle noise is known to be signal-dependent in ultrasound imaging. Hence, separating noise from signal becomes a difficult task. This paper describes a wavelet-based method for reducing speckle noise. We derive from the model of the displayed ultrasound image the optimal wavelet-domain filter, in the least mean-square sense. Simulations on synthetic data have been carried out in order to assess the performance of the proposed filter with regards to the classical wavelet shrinkage scheme, while phantom and tissue images have been used for testing it on real data. The results show that the filter effectively reduces the speckle noise while preserving resolvable details.

This paper reports on a scanning technique, denoted multi-angle compound imaging (MACI), using spatial compounding. The MACI method also contains elements of frequency compounding, as the transmit frequency is lowered for the highest beam angles in order to reduce grating lobes. Compared to conventional B-mode imaging MACI offers better defined tissue boundaries and lower variance of the speckle pattern, resulting in an image with reduced random variations. Design and implementation of a compound imaging system is described, images of rubber tubes and porcine aorta are shown and effects on visualization are discussed. The speckle reduction is analyzed numerically and the results are found to be in excellent agreement with existing theory. An investigation of detectability of low-contrast lesions shows significant improvements compared to conventional imaging. Finally, possibilities for improving diagnosis of atherosclerotic diseases using MACI are discussed.

The potential of intravascular ultrasound imaging for characterizing regional arterial elasticity was examined in an experimental tissue-equivalent vessel model. Differential intrawall displacement measurement, the first step in regional elasticity determination, was investigated using a crosscorrelation tracking algorithm. Calibration studies showed that tracking accuracy varied significantly with tracking direction (axial versus lateral) and position in the field of the transducer. Midfield geometric error in the axial direction for a nominal displacement of 100 μm was 5.5 μm whereas the corresponding error in the lateral direction was 31.7 μm. Displacement was tracked in serial intravascular images of vessel phantoms acquired during stepwise pressurization experiments from 0–250 mmHg. Two-dimensional grey scale maps of axial, lateral and net intrawall displacement components over the full pressurization range were generated. Displacement profiles demonstrated successful detection of differential radial displacement and good correlation with theoretical profiles (root mean square difference 3%). The corresponding experimental strain profiles were significantly noisier (root mean square difference 76%) due to small fluctuations in the displacement data. This work demonstrates that, with further refinement, regional strain mapping in vessel walls with intravascular ultrasound imaging is feasible. Mechanical characterization of arteries may provide a new tool to aid in assessing and treating atherosclerotic lesions.

Speckle appears in all conventional ultrasound images and is caused by the use of a phase-sensitive transducer. Speckle is an undesirable property as it can mask small but perhaps diagnostically significant image features. In this paper a homomorphic, hybrid nonlinear processing method, based on cancellation of scattering interference, is developed and examined. Experiments with synthetic and real ultrasound imagery show that the proposed method improves the contrast-to-noise ratio in both lesion and cyst areas and preserves edge clarity.

The SNR (signal-to-noise ratio) of the echo envelope image is a monotonically-increasing function of scatterer number density. Various SNRs, like amplitude SNR and intensity SNR, can be used to quantify the scatterer density. The problem of using a SNR based on higher order moments like the intensity SNR is that they require large sample sizes to obtain estimates with high confidence (the variance of the estimate becomes large for higher moments). In this paper, we consider SNRs based on fractional order moments (moments of order less than 1), and obtain mathematical analyses of their properties using the K distribution, which has been shown to be a good model for the density function of backscatter echo envelope signal. Statistics of SNRs based on fractional moment are derived and appear to be more robust and useful than the amplitude and intensity SNRs previously studied. The SNRs based on fractional order moments have greater dynamic range and the sample size requirements are smaller than those for integral order moment SNRs, like amplitude SNR or intensity SNR. Thus, SNRs based on fractional order moments could be used to better quantify the variations in scatterer density which can be used for tissue classification problems.

We propose a novel method for obtaining the maximum a posteriori (MAP) probabilistic segmentation of speckle-laden ultrasound images. Our technique is multiple-resolution based, and relies on the conversion of speckle images with Rayleigh statistics to subsampled images with Gaussian statistics. This conversion reduces computation time, as well as allowing accurate parameter estimation for a probabilistic segmentation algorithm. Results appear to provide improvements over previous techniques in terms of low-contrast detail and accuracy.

A novel signal postprocessing method has been developed for application in lesion detection. A signal distribution, directly related to the 2-dimensional envelope distribution comprising a B-mode image, is convolved with a block shaped kernel of which both base area (i.e., window size) and height are modulated by the local signal level. A theoretical analysis shows that the output includes information regarding the local mean brightness in the original image. This information can be displayed as a gray scale image. The predicted behavior agrees with experimental observations. On the basis of twenty simulated lesion images, it is demonstrated that the present algorithm is superior to the conventional L2-mean filter in lesion detection problems, whereby the performance is quantified by the area AROC under the Receiver Operating Characteristic. An improvement in the median value of AROC of the order of 0.1 was observed. Two parameters are used for adjustment. The total number of summations involved in the processing of a frame is comparable to the number required by the arithmetic mean filter, allowing real-time implementation.

The concept of level-dependent spatial summation is applied to the 2-dimensional amplitude distribution comprising an ultrasonic brightness image. The distribution is convolved with a level-dependent kernel. The size (i.e., extent) of the kernel is inversely proportional to the local signal level. The result is a 2-dimensional distribution of Mach bands, accentuating edges and providing useful information for the reduction of speckle. In addition, a small but powerful extension of the processing scheme is presented that supports global edge enhancement and contrast enhancement. The level of improvement has been quantified in terms of an increase in the lesion signal-to-noise ratio on the order of 60 percent. Only two system-dependent parameters have to be supplied for adjustment. The algorithm is rather insensitive to the exact choice of the parameters, making it quite robust. Since only a limited number of simple operations is required for the processing of a frame, the algorithm is expected to be suitable for real time application.

New techniques to estimate local blood and tissue velocities have been developed by several groups, including our own. The performance of these techniques is ultimately limited by the characteristics of ultrasonic imaging systems that determine the second-order statistics of speckle. These statistical parameters vary widely depending on the dimension of analysis in the image plane (lateral or axial) and on the echo input signal (radio-frequency or detected data). We use experiments and theory to examine these factors and describe their impact on the performance of our correlation-based technique for angle-independent tracking of blood or tissue motion. The results indicate that the second-order statistics determine the performance of our correlation-based algorithm and can be used to predict the performance of other angle-independent flow detection techniques.

Signal-to-noise-ratio (SNR) images were constructed by a sliding window technique. Theoretical derivations of the first order statistical properties of SNR images are given for the case of intensity coding of the original images for low number densities and for the limit case of fully developed speckle. The original images were obtained by realistic simulations of backscattering by homogeneous media, while using a range of number densities of the scatterers (isotropic scattering). The SNR imaging method is illustrated by results obtained from simulated scattering media containing a lesion that is contrasting in reflectivity level or in number density of the scatterers. A considerable improvement in lesion detection is obtained if the primary difference from the background speckle is due to a different number density of the scatterers within the lesion.

In this paper, we propose a new post formation adaptive image filtering technique, to be called the homogeneous region growing mean filter, in order to reduce speckle noise with edge preservation in medical ultrasonic images. First, to find a proper seed region, an initially assumed seed region, which is larger than the average speckle size around a given filtering point, is successively contracted according to a certain local homogeneity criterion. Once the seed region is determined, the next step is to grow the homogeneous region successively based on some measures of local homogeneity and similarity of the neighboring region. The output of the proposed filter for each filtering point is obtained from the arithmetic mean of the grown locally homogeneous region. Several simulation results are presented to illustrate the performance of the proposed filter. They show that the proposed technique effectively smoothes ultrasonic speckle and completely suppresses isolated impulsive noise over the entire texture in addition to preserving the edge information.

We describe a new method for quantitative imaging of strain and elastic modulus distributions in soft tissues. The method is based on external tissue compression, with subsequent computation of the strain profile along the transducer axis, which is derived from cross-correlation analysis of pre- and post-compression A-line pairs. The strain profile can then be converted to an elastic modulus profile by measuring the stresses applied by the compressing device and applying certain corrections for the nonuniform stress field. We report initial results of several phantom and excised animal tissue experiments which demonstrate the ability of this technique to quantitatively image strain and elastic modulus distributions with good resolution, sensitivity and with diminished speckle. We discuss several potential clinical uses of this technique.

Ultrasound speckle is a consequence of the stochastic nature of the reflectivity of scattering media (e.g., biological tissue) and of the coherent nature of piezoelectric transducers. This speckle noise can be reduced by the use of incoherent processing techniques (e.g., spatial compounding, incoherent summation, random phase and phase insensitive transducers.). We present a unified framework that explains the limitations of incoherent processing in terms of the information grain theory. This theory predicts the gains in SNR as well as the losses in directivity. We also present the random phase transducer approach to incoherent processing. This approach allows continuous control of the coherence from total incoherence to total coherence. We present applications to speckle reduction, detection of specular reflectors, attenuation estimation and ultrasound imaging. We show that totally incoherent transducers completely remove diffraction effects. They might be used in attenuation estimation, in which case, correction for diffraction is no longer required, in order to obtain unbiased estimates. Partially coherent transducers might also be useful in imaging to reduce speckle noise.

Ultrasonic waveform data from a tissue-mimicking phantom containing low contrast targets was digitized, stored and processed prior to creating and displaying ultrasonic images. Speckle reduction was performed by digital filtering of the waveform data with appropriately spaced and weighted digital filters, prior to both coherent and incoherent image averaging. The resultant images showed increased signal-to-noise ratios, consistent with theory. Better definition of low contrast target boundaries was noted in the processed, compared to unprocessed, images. Incoherent and coherent processing were investigated, and appeared to be equivalent.

Ultrasonic B-scan images of the pancreatic parenchyma in normal and cystic fibrosis patients were numerically analyzed. Images of both the maximum and minimum echoes from the tissue were generated by a recently-developed digital ultrasound system. Complete segregation of the two groups was achieved by averaging the echo amplitude over a selected region in the image. The dependence of the numerical values on the B-scan imaging mode, transducer properties, dynamic range compression curve, and operator scanning technique is discussed in depth.

A simple model for predicting the degree of speckle reduction in A-mode images via the correlation coefficient was tested experimentally. The model appeared to account correctly for the influence of image system bandwidth on frequency separations required for nearly independent sampling in compound frequency imaging. Moreover, the notion of compound bandwidth imaging was tested; the correlation between samples at different bandwidths decreased as the ratio of bandwidths decreased.

Factors inherent to the ultrasound imaging system can influence quantitative ultrasound image texture and may produce changes that mask or mimic those due to alterations in tissue structure. In this study, we assessed the effect of the method of image data acquisition and analysis on the variations in quantitative texture measures that occurred solely due to the position of a region of interest (ROI) within the field of view. When ROI's, which varied in range and azimuth within the image of a tissue equivalent phantom (of uniform composition), were assessed by conventional analysis of scan-converted (rectangular coordinate) data, over 50 percent of all texture measures showed significant differences. Pseudo polar analysis reduced the number of texture measures showing regional variation by 80 percent (p < 0.01) and true polar coordinate data analysis reduced the number by 74 percent (p < 0.01). True polar analysis completely abolished differences in texture measures between ROI's separated in azimuth. Acquisition and analysis of tissue texture data using polar coordinates should allow a more definitive identification of abnormal tissue.

The potential of using the phase derivative (PD) of the radio-frequency echograms for producing 2-dimensional grey scale images was further investigated. The PD images were produced by five different algorithms, which according to the results described in the companion paper yield PD images dominated by the amplitude (envelope, ENV), mixed PD-AM images and pure PD images. These images are termed according to their algorithms: ZCS, zero crossing counter with squelch; ASS, analytic signal with squelch; ASW, analytic signal with Wiener kernel; UNP, unwrapped phase; and SAS, smoothed analytic signal. The rf data were obtained from simulations and from experiments with a tissue mimicking phantom. PD images were analysed by calculation of the first and second order grey level statistics: mean level, signal-to-noise ratio (SNR) and the full width at half maximum (FWHM) of the autocovariance functions (ACVF). These parameters were systematically investigated for a range of depths with respect to the transducer and a range of scatterer densities of the insonated medium. The UNP and SAS images do not suffer much from the diffraction effect but do not display much information about the scatterer density either. The ASW and ASS images qualitatively display beam diffraction effects similar to those of the AM images, with the exception of the mean value which is at the minimum in the focus, where the AM yields a sharp maximum at that depth. The mean and the SNR of the ASW and ASS images increase with increasing scatterer density and saturate at a density of 5000 cm-3. The mean value of the envelope, however, displays a square root dependency over the whole range. The axial and lateral FWHM of the ACVF of the UNP and SAS methods are not significantly dependent on the scatterer density and decrease with increasing density for the ASW and ASS images, as was observed in the envelope images. It may be concluded that ASW and ASS methods produce PD grey scale images which are equally well suited for the diagnosis of diffuse diseases of parenchymal tissues as conventional AM images. The smaller “speckle” size of the ASW images might be advantageous, for the detection of focal lesions, but the lesion contrast is found to be much lower than for the ENV.

The statistics of speckle patterns in ultrasound images have potential for tissue characterization. In “fully developed speckle” from many random scatterers, the amplitude is widely recognized as possessing a Rayleigh distribution. This study examines how scattering populations and signal processing can produce non-Rayleigh distributions. The first order speckle statistics are shown to depend on random scatterer density and the amplitude and spacing of added periodic scatterers. Envelope detection, amplifier compression, and signal bandwidth are also shown to cause distinct changes in the Signal distribution.