Simulation and analysis of optical skin lesion images.

BACKGROUND/PURPOSE: In order to properly analyse the effectiveness of methods for optically differentiating malignant from benign skin lesions, it is necessary to have a set of images for which the ground truth is known. However, aspects of the ground truth of clinical images such as true lesion boundary position are unknown or not known precisely. Therefore, a skin/lesion image simulation with known features including boundary location, skin pattern and lesion colour is needed to enable accurate assessment of feature estimation algorithms for lesion classification. METHODS: In this paper, monochrome and colour skin/lesion images are synthesised with known characteristics including boundary, colour and skin pattern. Skin pattern is simulated with segmented lines with variations in length, orientation and intensity. Skin and lesion textures are modelled by an auto-regressive (AR) process. Monochrome skin lesion images are obtained by combining monochrome skin and lesion textures under the control of a known lesion shape with the addition of skin pattern. Colour skin lesion images are generated by mixing coloured skin and lesion textures. Finally, an inflammation area and image artefacts such as hair and specular reflection are added. RESULTS: The synthesised images provide the image set for evaluating image pre-processing, segmentation and skin pattern analysis. The pre-processing includes hair removal and specular reflection reduction. An AR model interpolation is suggested for hair removal, and multiple illumination processing is developed to decrease specular reflection. A fast snake algorithm is extended to detect the boundaries of skin lesion and inflammation areas. Skin line direction is detected as a feature to measure the disruption of skin pattern caused by lesion. CONCLUSIONS: Simulation of monochrome and colour skin/lesion image has been investigated, which is an alternative way to provide image set with known characteristics to validate image processing algorithms for image pre-processing, lesion/inflammation boundary detection and skin pattern analysis.

Enhancement of lesion classification using divergence, curl and curvature of skin pattern.

BACKGROUND/PURPOSE: The observation that skin pattern tends to be disrupted by malignant but not by benign skin lesions suggests that measurements of skin pattern disruption on simply captured white light optical clinical (WLC) skin images could be a useful contribution to a diagnostic feature set. Previous work which generated a flow field of skin pattern using a measurement of local line direction and variation determined by the minimum eigenvalue and its corresponding eigenvector of the local tensor matrix to measure skin pattern disruption was computationally low cost and encouraging. This paper explores the possibility of extracting new features from the first and second differentiations of this flow field to enhance classification performance. METHODS: The skin pattern was extracted from WLC skin images by high-pass filtering. The skin line direction was estimated using a local image gradient matrix to produce a flow field of skin pattern. Divergence, curl, mean and Gaussian curvatures of this flow field were computed from the first and second differentiations of this flow field. The difference of these measures combined with skin line direction across the lesion image boundary was used as a lesion classifier. RESULTS: A set of images of malignant melanoma and benign naevi were analysed as above and the scatter plot in a two-dimensional dominant feature space using principal component analysis showed excellent separation of benign and malignant lesions. A receiver operating characteristic plot enclosed an area of 0.96. CONCLUSIONS: The experimental results show that the divergence, curl, mean and Gaussian curvatures of the flow field can increase lesion classifier accuracy. Combined with skin line direction they are promising features for distinguishing malignant melanoma from benign lesions and the methods used are computationally efficient which is important if their use is to be considered in clinical practice.

Angle-dependence and reproducibility of dual-beam vector doppler ultrasound in the common carotid arteries of normal volunteers.

Dual-beam vector Doppler has the potential to improve peak systolic blood velocity measurement accuracy by automatically correcting for the beam-flow Doppler angle. Using a modified linear-array system with a split receive aperture, we have assessed the angle-dependence over Doppler angles of 40 degrees -70 degrees and the reproducibility of the dual-beam blood maximum velocity estimate measured in the common carotid arteries (CCA) 1 to 2 cm prior to the bifurcation of 9 presumed-healthy volunteers. The velocity magnitude estimate was reduced by approximately 7.9% as the angle between the transmit beam and the vessel axis was increased from 40 degrees to 70 degrees. With repeat measurements made, on average, approximately 6 weeks apart, the 95% velocity magnitude limits of agreement were as follows: Intraobserver -41.3 to +45.2 cm/s; interobserver -29.6 to +46.8 cm/s. There was an 8.6 cm/s interobserver bias in velocity magnitude. We conclude that the dual-beam vector Doppler system can measure blood velocity within its scan plane with low dependence on angle and with similar reproducibility to that of single-beam systems.

Doppler power spectrum from a Gaussian sample volume.

A closed-form expression for the Doppler power spectrum due solely to the range of blood velocities passing through a Gaussian sample volume placed anywhere in a vessel under conditions of axisymmetric flow, uniform backscatter and negligible intrinsic spectral broadening has been derived. The formulation presented here allows the independent specification of the sample volume position and width, in the three dimensions, and enables simple estimations of spectral shape for pulsed wave Doppler systems. Simpler expressions were derived for the cases of symmetric sample volume projections onto the vessel cross-section and/or sample volumes centred in the vessel. Closed form expressions were derived for mean frequency and spectral width in the case of a symmetric sample volume projection centred in the vessel. The effects of sample volume size and position on the Doppler spectral width and mean frequency are shown for a range of velocity profiles.

Characterization of vortices using pulsed-wave Doppler ultrasound.

The detection and characterization of vortices from a Kaman vortex generator by means of a 20 MHz pulsed-wave Doppler ultrasound system were assessed. Measurements were made at different steady flowrates in a 10 mm internal diameter polyurethane tube, 14 mm distal to a circular cylinder of diameter 2 mm, placed across the tube inlet. The results were compared with those obtained with a two-component laser Doppler anemometer system. There was generally good agreement between the two techniques in the measurement of convective flow velocity, frequency of vortex shedding and the circulation velocity of the vortices. It is concluded that pulsed-wave Doppler ultrasound is a suitable technique for investigating vortical flow structures.

Characterisation of vortex shedding in vascular anastomosis models using pulsed Doppler ultrasound.

Vortex shedding at vascular anastomoses were investigated in vitro using a 20 MHz pulsed-wave Doppler velocimeter. Centreline velocity measurements were made at various axial distances in simplified polyurethane models of proximal and distal end-to-side anastomoses of angles 15, 30, 45, 60 and 80 degrees using pulsatile flow waveforms similar to those in femoropopliteal bypass grafts. The in-phase and quadrature Doppler signals were recorded and the maximum frequency waveform, averaged over 64 cycles, was obtained using short-time Fourier transform. A fourth-order Butterworth low-pass filter was employed to separate the vortex velocity signal from the convective velocity. The vortex signal envelope was calculated using a Hilbert transform method and the vortex amplitude was taken as the maximum of this envelope. The results show that higher vortex amplitude were found in the proximal anastomoses and under resting flow conditions. Although the vortex amplitudes generally increased with angles of anastomosis, they were found to be higher in the 60 degrees than in the 80 degrees proximal anastomosis. The vortex structures were investigated using spectrograms and these show prominent features at 40-50 Hz indicative of the short-duration oscillatory signals during the decelerative phase of systole expected from the passage of vortices. The study indicates that flow disturbances due to vortex shedding may be a common feature in femoropopliteal bypass grafts.

Ultrasonic investigation of blood flow.

Ultrasound and the Doppler effect are used to measure blood velocity non-invasively in order to diagnose blood vessel disease. Intrusive lesions on arterial walls give rise to an alteration of the time-varying blood velocity waveform and local blood flow disturbance which are detected and measured using the envelope and width of the Doppler signal spectrogram respectively. Flow may also be imaged in colour superimposed on a grey-scale anatomical image allowing vessel narrowing and the accompanying flow disturbance to be visualized. Developments in three-dimensional imaging, angle tolerant velocity measurements and increased sensitivity using second harmonic backscatter from encapsulated-bubble contrast media ensure increasing use of this modality.

Finite beam-width ray model for geometric spectral broadening.

The purpose of the study was to compare measured spectral width and maximum frequency with that predicted from ray models of geometric spectral broadening. Zero and finite beam-width models were used. Spectral data were acquired from a string phantom using two commonly-used linear array systems. Beam width and Doppler aperture sizes were measured using a needle hydrophone. The results showed that the experimentally measured data agreed best with the finite beam-width model. The zero beam-width model was in error by up to 50% for calculated spectral width, and up to 10% for maximum frequency. It is concluded that spectral width and maximum frequency are best calculated using the finite beam-width model, and that ultrasound manufacturers could improve the variation in spectral broadening measured at different locations on a single machine by adjusting the aperture size to give a constant subtended angle and beam width.

Spectrum of Doppler ultrasound signals from nonstationary blood flow.

A new formulation for the Doppler signal generation process in pulsatile flow has been developed enabling easier identification and quantification of the mechanisms involved in spectral broadening and the development of a simple estimation formula for the measured rms spectral width. The accuracy of the estimation formula was tested by comparing it with the spectral widths found by using conventional spectral estimation on simulated Doppler signals from pulsatile flow. The influence of acceleration, sample volume size, and time window duration on the Doppler spectral width was investigated for flow with blunt and parabolic velocity profiles passing through Gaussian-shaped sample volumes. Our results show that, for short duration windows, the spectral width is dominated by window broadening and that acceleration has a small effect on the spectral width. For long duration windows, the effect of acceleration must be taken into account. The size of the sample volume affects the spectral width of the Doppler signal in two ways: by intrinsic broadening and by the range of velocities passing through it. These effects act in opposite directions. The simple spectral width estimation formula was shown to have excellent agreement with widths calculated using the model and indicates the potential for correcting not only for window and nonstationarity broadening but also for intrinsic broadening.

Developments in cardiovascular ultrasound. Part 2: Arterial applications.

Many of the changes resulting from arterial disease can be measured, using Doppler ultrasound for measurement of blood velocity and B-scan imaging for measurement of tissue structure and composition. Wall thickness, the degree of arterial narrowing and plaque volume can be measured using B-scan imaging, and 3D ultrasound can be used to improve the accuracy of measurements of plaque volume and for improved visualisation of complex arterial geometries. Measurement of the dynamic properties of the arterial wall permits estimation of wall elasticity and plaque motion. From the Doppler signal, measurements of blood velocity are used to estimate the degree of arterial narrowing and volumetric flow, although measurement errors can be large. Wall shear stress can be estimated by measuring the velocity gradient at the vessel wall. The problems of inadequate spatial resolution and interference from overlying tissue are largely removed when intravascular systems are used, and these have superior capability in the assessment of arterial structure and tissue composition. However, measurement of quantities relating to blood flow is more difficult using the intravascular approach, as the indwelling cather disturbs the blood flow pattern, and currently, assessment of flow and vessel cross-section are not performed at the same site.

The significance of variations in the angular correction factor in in situ gamma spectrometry.

In situ gamma spectrometry is a powerful method of assessing radioactive contamination in soil. The most widely adopted calibration methodology relates the overall sensitivity of the detector system to the product of three calibration factors: (a) the flux at the detector per unit activity in the ground phi/S(A), (b) the detected count-rate per unit flux incident normally at the detector N0/phi and (c) a correction factor to take into account the angular non-uniformity in response of the detector (Nf/N0). The dependence of the latter factor on the activity distribution with soil depth is generally neglected despite the lack of published evidence to support this. By (i) modelling and (ii) use of published experimental profiles, this work examines the range of Nf/N0 values likely to be encountered in the field. It was found that the use of a fixed angular correction factor is justified given that the maximum errors in the derived activity concentration do not exceed 5% and are far outweighed by other uncertainties.

Broadband spectra of seismic survey air-gun emissions, with reference to dolphin auditory thresholds.

Acoustic emissions from a 2120 cubic in air-gun array were recorded through a towed hydrophone assembly during an oil industry 2-D seismic survey off the West Wales Coast of the British Isles. Recorded seismic pulses were sampled, calibrated, and analyzed post-survey to investigate power levels of the pulses in the band 200 Hz-22 kHz at 750-m, 1-km, 2.2-km, and 8-km range from source. At 750-m range from source, seismic pulse power at the 200-Hz end of the spectrum was 140 dB re: 1 microPa2/Hz, and at the 20-kHz end of the spectrum seismic pulse power was 90 dB re: 1 microPa2/Hz. Although the background noise levels of the seismic recordings were far in excess of ambient, due to the proximity of engine, propeller, and flow sources of the ship towing the hydrophone, seismic power dominated the entire recorded bandwidth of 200 Hz-22 kHz at ranges of up to 2 km from the air-gun source. Even at 8-km range seismic power was still clearly in excess of the high background noise levels up to 8 kHz. Acoustic observations of common dolphins during preceding seismic surveys suggest that these animals avoided the immediate vicinity of the air-gun array while firing was in progress, i.e., localized disturbance occurred during seismic surveying. Although a general pattern of localized disturbance is suggested, one specific observation revealed that common dolphins were able to tolerate the seismic pulses at 1-km range from the air-gun array. Given the high broadband seismic pulse power levels across the entire recorded bandwidth, and known auditory thresholds for several dolphin species, we consider such seismic emissions to be clearly audible to dolphins across a bandwidth of tens on kilohertz, and at least out to 8-km range.

Developments in cardiovascular ultrasound. Part 3: Cardiac applications.

Echocardiography is still the principal, non-invasive method of investigation for the evaluation of cardiac disorders. Using Doppler ultrasound, indices such as coronary flow reserve and cardiac output can be determined. The severity of valvular stenosis can be determined by the area of the valve, either directly from 2D echo, from pressure half-time calculations, from continuity equations or from the proximal isovelocity surface area method. Alternatively, the severity of regurgitation can be estimated by colour or pulsed ultrasound detection of the back-projection of the high-velocity jet into the chamber. Myocardial wall abnormalities can be assessed using 2D ultrasound, M-mode or analysis from the radio-frequency-ultrasound signal. Doppler tissue imaging can be used to quantify intra-myocardial wall velocities, and 3D reconstruction of cardiac images can provide visualisation of the complete cardiac anatomy from any orientation. The development of myocardial contrast agents and associated imaging techniques to enhance visualisation of these agents within the myocardium has aided qualitative assessment of myocardial perfusion abnormalities. However, quantitative myocardial perfusion has still to be realised.

Developments in cardiovascular ultrasound: Part 1: Signal processing and instrumentation.

One of the major contributions to the improvement of spectral Doppler and colour flow imaging instruments has been the development of advanced signal-processing techniques made possible by increasing computing power. Model-based or parametric spectral estimators, time-frequency transforms, station-arising algorithms and spectral width correction techniques have been investigated as possible improvements on the FFT-based estimators currently used for real-time spectral estimation of Doppler signals. In colour flow imaging some improvement on velocity estimation accuracy has been achieved by the use of new algorithms but at the expense of increased computational complexity compared with the conventional autocorrelation method. Polynomial filters have been demonstrated to have some advantages over IIR filters for stationary echo cancellation. Several methods of velocity vector estimation to overcome the problem of angle dependence have been studied, including 2D feature tracking, two and three beam approaches and the use of spectral width in addition to mean frequency. 3D data acquisition and display and Doppler power imaging have also been investigated. The use of harmonic imaging, using the second harmonic generated by encapsulated bubble contrast media, seems promising particularly for imaging slow flow. Parallel image data acquisition using non-sequential scanning or broad beam transmission, followed by simultaneous reception along a number of beams, has been studied to speed up 'real-time' imaging.

Nonstationarity broadening reduction in pulsed Doppler spectrum measurements using time-frequency estimators.

The spectral width of Doppler signals is used as measure of lesion-induced flow disturbance. Its estimation accuracy is compromised using the conventional short-term Fourier transform (STFT) since this method implicitly assumes signal stationarity during the signal window while the Doppler signals from arteries are markedly nonstationary. The Wigner-Ville (WVD), Choi-Williams (CWD) and Bessel distributions (BD), specifically designed for nonstationary signals, have been optimized for spectral width estimation accuracy and compared to the STFT under different signal to noise ratios using simulated Doppler signals of known time-frequency characteristics. The optimum parameter values for each method were determined as a Hanning window duration of 10 ms for the STFT, 40 ms for the WVD and CWD and 20 ms for the BD and dimensionless time-frequency smoothing constant values of five in the CWD and two in the BD. Thresholding was used to reduce the effect of cross terms and side lobes in the WVD and BD. With no added noise the WVD gave the lowest estimation error followed by the CWD. At signal-to-noise ratios (SNR's) of 10 dB and 20 dB the CWD and BD had similar errors and were markedly better than the other estimators. Overall the CWD gave the best performance.

Comparison of Doppler signal analysis techniques for velocity waveform, turbulence and vortex measurement: a simulation study.

Simulated time-varying Doppler signals incorporating bandwidth, power variation and vortex simulation have been used to compare a number of signal analysis techniques with a view to optimising the accuracy of convective velocity waveform, spectral broadening and vortex signal estimation. The short-time Fourier transform (STFT), the autoregressive (AR) modified covariance estimator, the time-frequency pseudo-Wigner-Ville and Choi-Williams distributions and a partial stationarising algorithm were investigated for a range of some analysis parameters (such as window duration, AR model order). It was found that all methods could estimate the convective velocity waveform well and all the nonclassical methods were an improvement over the STFT for bandwidth estimation with the stationarising method giving the lowest error. For vortex measurement, using parameters that were optimum for mean frequency and bandwidth estimation, the stationarising, modified covariance, pseudo-Wigner-Ville with a 10-ms window and Choi-Williams methods gave improved performances compared with the STFT.

Cost/benefit criterion for selection of pulsed Doppler ultrasound spectral mean frequency and bandwidth estimators.

A flexible selection criterion for spectral estimators based on the weighted statistical accuracy (benefit) of estimation of decisive spectral parameters under the constraint of low computational complexity (cost) is proposed. This new cost/benefit criterion also selects the model order for parametric spectral estimators-selecting model orders significantly lower than those determined by accepted criteria. The importance of different Doppler signal parameters (e.g., mean frequency and spectral bandwidth) and their accuracy of estimation is incorporated by the use of weighting factors. The use of this method with simulated Doppler signals led to the selection of the modified covariance Alt estimator.

Effect of deviation from plane wave conditions on the Doppler spectrum from an ultrasonic blood flow detector.

Deviation from plane wave conditions within the ultrasound beam of a Doppler blood flow detector leads to a non-linear relationship between the phase angle of the back-scattered signal and the scatterer position. This in turn leads to frequency modulation of the Doppler signal and an increase in the Doppler spectrum width. The relationship between the ultrasound beam and the observed signal spectrum has been investigated by employing a computer-based model of the ultrasound field which enabled the calculation of: 1, pressure (amplitude and phase angle) field distributions from plane disc and focused transducers with unapodized and apodized aperture field distributions; 2, the Doppler signal from a scatterer moving through the field; and 3, the spectrum of this signal. The increase in spectral width resulting from deviations from plane wave conditions was calculated by comparing this spectrum with that of the signal from which frequency modulation had been removed.

Effect of frequency dependent processes on pulsed Doppler sample volume.

The change in the frequency spectrum of ultrasound pulses as a result of frequency dependent attenuation and scatter is known to alter the spectrum of Doppler signals from blood flow parallel to the beam axis. It is shown here that the change in pulse shape accompanying this pulse spectrum change as a result of these processes, together with the velocity dispersion accompanying frequency dependent attenuation and linked to it by the Kramers-Kronig relations, will change the pulsed Doppler sample volume shape and position.