IEC 61102:1991, Measurement and characterisation of ultrasonic fields using hydrophones in the frequency range 0,5 MHz to 15 MHz IEC 61685:2001, Ultrasonics – Flow measurement systems
Test methods
To conduct the test procedures, essential items include tissue-mimicking test objects with targets positioned at precise locations, a tissue-mimicking test object featuring a 3D structure of exact dimensions, and a tank filled with degassed working liquid.
The specifications of these devices are given in the annexes.
Instruments
The equipment outlined in this section is designed for testing ultrasonic scanners in clinical settings, ensuring that data collection and analysis are both objective and reproducible.
For accurate spatial measurements in ultrasound imaging, it is essential to use digitally encoded images from modern ultrasound devices, which provide objective and reproducible data Hospital-based users with digital measurement expertise can effectively utilize these digital images for spatial measurements To assess Point Spread Function (PSF) and Line Spread Function (LSF), a characteristic curve of linear echo amplitude must be established, or a sparse representation created using calibrated reflectors In systems where radiofrequency (rf) scan-line data is available, these measurements offer greater precision due to the importance of linear signal amplitude It is crucial to document the source and level of rf data used For devices lacking digital imaging capabilities, a frame grabber can digitize ultrasound images, requiring a minimum spatial resolution of 512 × 512 pixels and at least 256 grey shades Additionally, suitable image analysis software is necessary for performing measurements on the digitized ultrasound images The digitizer must maintain linearity with spatial uncertainty of less than 1% over 75% of the image dimension, signal level linearity within 3% of the full range, and signal level stability of less than 5% over a year.
The digital imaging software must enable users to position the cursor anywhere on the screen to retrieve the pixel address, specifically the row and column coordinates This functionality is essential for calibrating digital images to actual distances captured in ultrasound images After calibration, digitized ultrasound images can undergo advanced software analysis that surpasses the capabilities of direct ultrasound display Additionally, the software should facilitate the reading of the grey value at any specified pixel address.
To calibrate the digital image pixel distance relative to the ultrasound imaging system, first, scan an image of a test object with a known working liquid, ensuring to note the magnification level for consistency in measurements Next, use electronic calipers to measure the actual distances between two wires or filaments positioned at approximately 75% of the screen size, confirming that the measured distance aligns with the actual distance This measurement should be conducted for both vertical and horizontal pairs of wires; any deviations necessitate scanner adjustments If adjustments are not feasible, use the actual distances for further calculations Then, digitize the scanned image and utilize imaging software to measure pixel distances between wire pairs by obtaining their pixel addresses and calculating the differences Repeat this process for various locations and orientations Finally, calculate the pixel distances for different wire positions and directions, dividing each by the actual distance in millimeters The average of these ratios will provide the pixels per millimeter calibration for the digitizer, which can be applied to compute relative distances in future digitized images for the specific scanner and magnification used.
Tissue-mimicking test objects shall contain structures that allow the following types of measurement to be made: a) linear; b) curvilinear; c) circumferential; d) area; e) volume; f) image distortion; g) M-mode calibration
Examples of tissue-mimicking test objects are given in Annex B.
Test settings
Due to the numerous combinations of scanner settings and transducers, it is not feasible to test every possible configuration Consequently, tests are conducted for each ultrasonic transducer using two specific settings: one that delivers a complete image and another that maximizes the resolution of the test objects To attain optimal resolution across all visible targets, the focusing of the ultrasonic beam should be extended over a wide range.
The test object, containing an array of filaments as in Figure A.1, is used for the procedures described below
6.3.2 Display settings (focus, brilliance, contrast)
To achieve optimal image quality, first, set the focus to sharp and lower the brilliance and contrast controls Gradually increase the brilliance until the echo-free zone reaches the faintest visible shade of grey Next, adjust the contrast to maximize the range of grey shades in the image Finally, verify the sharpness of the focus, and if necessary, repeat the entire process for further refinement.
In the sensitivity settings of ultrasonic transducers, the nominal frequency is established, and suppression or reject controls are fine-tuned to display the faintest signals The output power and gain are adjusted to ensure that target filaments appear as the smallest visible points on the screen Additionally, time-gain compensation (TGC) controls are set to maintain uniform brightness across the image, with the TGC slope ideally close to zero when scanning in a working liquid.
A final optimisation of the image may be carried out by a small change in the suppression level, gain or output power
When using a scanner with automatic time-gain compensation (ATGC), it is essential to conduct tests in this operational mode The test object should be imaged with ATGC activated, while optimizing the image through any available manual controls, such as overall gain or output power.
Digital acquisition of ultrasound images enables objective measurements and facilitates the storage of images for future comparison A key benefit of digital recording is its resistance to degradation, unlike traditional photographic or video recording systems.
Test parameters
Techniques are described in this standard for the following types of measurement:
To prevent pulse distortion from non-linear propagation, the transmitted intensity must remain low (refer to IEC 61102) It is essential to compile a comprehensive list of factors affecting scanner operation, including transducer type, frequency, sensitivity control settings, focusing, and image processing options This information should be documented in detail to enable precise repetition of the test by another operator in the future and must accompany the measurement results.
6.4.2 Measurement accuracy (linear, curvilinear, circumferential, area)
To evaluate the measurement system's accuracy of a scanner, the filaments in the test object are imaged with optimal sensitivity for sharp echoes If the test object is sealed, a coupling agent is necessary A digitized ultrasound image is captured of the filament targets located at the transducer assembly's typical working range Additionally, factors influencing resolution, such as scan converter image processing and focusing options, are recorded This procedure is then repeated for the other ultrasonic transducers in the scanner.
Measurements are conducted in straight lines on the screen, covering approximately 75% of the displayed range Image analysis software generates a linear brightness profile for each dimension, with distances measured from peak to peak of the brightness profiles of wires or filaments If the measurements are noisy, the peak value is replaced by the midpoint between the –3 dB points, and this adjustment is documented Measurements are taken along at least one vertical and one horizontal line, as shown in Figures A.1 and A.2, and, when feasible, along near-vertical directions within the field of view The average percentage error is recorded for each length in every direction, and the process is repeated for all available display scales.
To assess the precision of measurements for curved lines and cross-sectional areas, closed figures occupying approximately 0.75 of the field-of-view are centrally traced on the display The polygon-shaped regions are measured for circumference and area, with percentage errors calculated accordingly Additionally, two smaller figures, covering areas of 0.1 and 0.25 of the field-of-view, are positioned at the top and bottom of the display for further measurements This evaluation process is conducted across the various scales available on the display.
To ensure thorough testing, it is essential to assess sources of variability, applicable to both short-term and long-term measurement and analysis procedures Short-term testing, conducted on the same day, should be performed multiple times from setup to final analysis In manual testing, an operator must repeat the tests within a brief timeframe, followed by repetitions of the tests by several different operators.
6.4.3 Display and recording of image distortion
To analyze the two-dimensional array of filaments depicted in Figure A.2, ensure that their echoes are uniformly visible across the entire field-of-view Identify and select wires or filaments positioned both horizontally and vertically from the center Measure the distance from the center of each selected wire or filament to a reference wire or filament situated near the center of the field-of-view Finally, calculate and present the percentage errors in a tabulated format.
Observe the directly viewed image of the array of filaments to check that any distortion (i.e failing orthogonality) of dimensions in the display is less than 3 %
An M-mode facility exists on most real-time scanners A partial assessment of its performance can be carried out using the test objects described in Annex A
6.4.4.2 Spatial measurements (scrolling A-scan line)
Conducting an M-mode scan with the ultrasound beam aimed at wires or filaments in a resolution test object allows for the assessment of measurement errors within the system, similar to the previously described B-mode technique.
Distortions of the display or record are checked and recorded using the array of target filaments in the test object as is done for a B-mode image
To verify the accuracy of the time-axis calibration of the M-mode trace, one can inject bursts of ultrasound into the ultrasonic transducer using an external pulse generator This process involves sending 1 ms bursts at precise intervals, such as every 200 ms.
Measurement checks should be carried out for the digitized image of the M-mode trace on the display screen The errors in measurement should be less than 3 %
For accurate evaluation of tissue-thickness M-mode measurements, a tissue-like phantom that can be compressed and relaxed at specific positions is essential This phantom should enable comparison between the compressed and relaxed thicknesses and the M-mode readings, while also allowing for varying rates of compression and relaxation A deformable sponge phantom may be particularly effective for these measurements Additionally, it is crucial to test the synthetic M-mode at various depths, necessitating the ability to reposition the transducer for different target-to-transducer distances Furthermore, the testing should be conducted for each M-mode sweep speed to ensure comprehensive assessment.
7 Methods for calibrating 3D-measurement systems
General
Three-dimensional (3D) imaging systems can be categorized into those used solely for visualization and those that also offer measurement capabilities It is crucial to analyze the volume reconstruction methods and their related challenges to assess the accuracy of the reconstructed images This discussion focuses specifically on measuring the dimensional accuracy of the reconstruction Additionally, the resolution of 3D systems will be addressed in document IEC 61391-2, which covers system resolution and sensitivity.
Types of 3D-reconstruction methods
True 3D imaging necessitates the assembly of data into a three-dimensional voxel matrix, typically derived from a series of ultrasonic scan planes that encompass the target volume The accuracy of this 3D imaging system relies on the spatial density of data points, which is influenced by the number of ultrasonic scan lines, pulse length, and the configuration of the ultrasonic scan planes in the depth dimension Proper scanning techniques are crucial, as the fidelity of the reconstruction hinges on the preservation of physical distances It is essential that the spacing between successive ultrasonic scan planes remains constant and is less than the elevational resolution of the transducer; otherwise, interpolation between adjacent planes may introduce dimensional errors in the reconstructed volumes.
After constructing a 3D-matrix, data can be extracted from any dimension within the volume For instance, by collecting multiple xy-ultrasonic scan-plane images to create a 3D volume, the 3D-matrix can be sliced perpendicular to the y-axis in the xz plane, generating C-scan slices Additionally, the 3D reconstruction allows for rotation in space, enabling viewing and measurement from previously inaccessible angles in the original ultrasonic scan planes.
[See [11, 12] for reviews of 3D-scanning techniques.]
3D-volume-matrix acquisition and reconstruction are performed in two basic ways: a) reconstruction by external positioning; b) sequential reconstruction
Each method has its own characteristics, strengths, and problems
Reconstruction methods for 3D volume using external positioning rely on a reference point and coordinate frame, recording all dimensions within the 3D volume matrix relative to these coordinates Typically, a scanning framework is employed, where the scanning volume is inserted, and a motorized transducer moves along a rail at a constant speed Rigid supports may also be utilized to ensure the ultrasonic transducer maintains accurate 3D coordinates While this reconstruction method is known for its accuracy and reliability, it can encounter issues related to the initial positioning of the transducer's support frame, motor speed variations, or changes in the positioning system during data collection.
A variant of external positioning systems exists, in which the transducer is guided manually and its position and direction are sensed with respect to a reference coordinate system
Sequential positioning and reconstruction methods rely on attaching subsequent scan planes to a 3D matrix based on the previous plane's position One approach utilizes the rate of change of image speckle in at least one dimension, but this method is based on assumptions that may not always hold true For instance, it often assumes motion is either a purely linear or purely angular sweep, with some commercial systems presuming a uniform linear sweep speed Therefore, conducting tests is crucial to assess the capabilities and limitations of these measurements in laboratory settings Issues arise when the transducer is not moved at a consistent speed or when its angle deviates from the previous orientation, leading to inaccuracies in the reconstructed volume, as the reconstruction scheme typically cannot adjust for these shifts.
Test parameters associated with reconstruction problems
For reconstruction by external positioning-system testing, either water-based or tissue- mimicking-based test objects may be used
To evaluate the test parameters, measure the reconstructed lengths in all three Cartesian coordinate directions from the reconstructed volume and compare them with the physical dimensions of the object in the same directions Additionally, verify the orientation of the ultrasonic transducer relative to the frame and reference point, as well as the motor speed and the distance between the ultrasonic scan planes.
Cartesian dimensions of reconstructed volume Procedures for measuring the following parameters are described below a) linear dimensions; b) areas; c) perimeters of areas; d) volumes
3D-spatial encoding systems that rely on image speckle necessitate a test object with consistent speckle backscatter and structures to evaluate registration accuracy and precision In certain ultrasound systems, it is essential to assess not only the accuracy of long-distance measurements but also the uniformity of the distance scale Additionally, many sequential encoding systems require further testing due to potential imprecisions in position encoding, which can lead to distortions in the image direction caused by local jumps or delays in the recorded position during scanning.
To evaluate the accuracy of the reconstructed volume, it is essential to measure the reconstructed lengths across all three Cartesian coordinates and compare them with the physical dimensions of the object in the same coordinates The procedures for measuring key parameters include linear dimensions (axes), areas, perimeter of areas, and volumes.
7.3.3 Test instruments (phantoms) for evaluation of 3D-reconstruction accuracy
The reconstruction method utilizing external positioning can employ a filament test object filled with a working liquid, as illustrated in Annex A, Figures A.1 and A.2 This system's independence from speckle correlation for positioning ultrasonic scan planes within a 3D matrix makes the filament matrix an effective tool for assessing reconstruction accuracy For each filament row, the distance \( r \) from the image location of a filament to a reference filament is measured and compared to the known distance \( r' \) in the phantom The maximum and root mean square (r.m.s.) deviations of the measured filament positions from the linear regression lines of \( r \) on \( r' \) are calculated, along with the slopes of the fitted lines.
To accurately assess the volume measurement algorithm of the system under test, it is essential to utilize a filament test object, as illustrated in Figures A.1 and A.2, especially when dealing with sharp corners and flat surfaces.
Annex B presents a second test object suitable for evaluating both types of systems, featuring ovoid-shaped tissue-mimicking material structures with minimal specular boundaries, as illustrated in Figures B.1 to B.4 This volumetric test object defines targets through variations in backscatter contrast, with borders determined solely by grey scale texture and average signal levels By focusing on volumetric backscatter instead of specular reflectors or point targets, this approach facilitates the assessment of image formation, display, and measurement aspects in the ultrasound system's reconstruction of volumetric targets within the body.
Test methods for measurement of 3D-reconstruction accuracy
For accurate 3D measurements using both conventional 2D scanners and 3D scanners, it is essential that the separation of recorded image planes is less than the thickness of the ultrasonic scan plane at its focus Ideally, this separation should be less than half the elevational focal width of the ultrasonic scan plane If available, it is recommended to utilize controls for scan-plane separation to achieve high-quality 3D imaging and testing.
7.4.2 Measurement methods and accuracy using the filament test object
Volumetric measurements can be derived from two orthogonal 2D images by capturing two views of a roughly spherical object and measuring its three major axes The volume can then be calculated using appropriate equations or ultrasound systems To validate the volumetric calculation, diameters of a presumed sphere can be measured through four or more filaments in the phantom on one image, and these measurements can be repeated on the adjacent image, which is assumed to be at a 90-degree angle The calculated volume of the sphere, based on the measured cross-section, is then compared to the volume of the assumed sphere, allowing for the computation of any error.
To ensure accurate angulation correction of the scan plane in the elevational direction and to verify the measurement algorithm from a 3D-sweep parallel to filaments, conduct a 3D-scan with the central image planes perpendicular to the filaments and the scan direction aligned with them, positioning the transducer in View B of Figure A.1 If the transducer can rotate in an arc in the elevational direction, perform the scan accordingly and, if feasible, display the reconstructed volume with images normal to the filaments In cases where reformatting is not possible with a sector scanner, adjust the filament separation based on the known viewing angle Additionally, carry out the measurements outlined in section 6.4.2 on the first, middle, and last images of the 3D set, documenting any errors and variances The ratios of mean filament spacings to known spacings in horizontal and vertical directions are termed the lateral- and longitudinal-dimension calibration factors, R_x and R_y, respectively Finally, verify that the mean spacings of filament groups expected to have uniform spacings are consistent across all image planes.
To calibrate the ultrasonic scan-plane separation in 3D imaging and evaluate the distortion of images reconstructed from a volume data set along the thickness direction of the ultrasonic scan-plane, the transducer is gradually moved or rotated in the elevational direction.
The ultrasonic scan is conducted in a direction normal to the filaments, with the transducer moving from left to right as shown in Transducer View A of Figure A.1 This movement follows the 3D-scanning guidelines provided by the ultrasound system manufacturer While typically only linear translation or sector sweeps are permitted, exploring deviations from these instructions can reveal the extent of error produced.
In the second scan (View A), reconstructed images of the filament rows are displayed perpendicularly to the filaments The analysis includes calculating and reporting the maximum and root mean square (r.m.s.) deviations for three key aspects: a) the differences between the measured filament spacings and their known values; b) the discrepancies in the measured filament positions compared to the fitted lines; and c) the variations in the slopes of the fitted lines from the expected values.
To assess the precision of curved lines and cross-sectional areas, closed figures are centrally traced on the display, covering approximately 75% of the field of view Measurements of lengths, circumferences, and areas are taken, and the ratios of measured to known areas are calculated Volume measurements are conducted with the sweep direction perpendicular to the filaments A known area, A, is marked in the reconstructed image perpendicular to the filaments, along with an ultrasound system-indicated length L' for the third dimension of a 3D volume enclosed by the filaments, which, in the case of Figure A.1, resembles a cylindrical rod The measured volume A' × L' is compared to the known volume A × L'/R_x, where R_x is the lateral-dimension calibration factor Additional examples of measurements using filament test objects are provided in reference [16].
7.4.3 Measurement accuracy using volumetric targets in a backscattering object phantom (Figure B.1) with a 2D-scanner
In the measurement process, the transducer is rotated and tilted to capture a circular cross-section of each target within the 3D test object The image plane is adjusted to identify the minor axis, which represents the largest diameter where the object maintains a circular appearance Calliper markers are then positioned at the ends of the largest vertical diameter, and this value is recorded A horizontal diameter is measured similarly, and the average of these two diameter measurements is designated as \( b \).
To measure the longest dimension of the ellipsoid, rotate the transducer by 90° and identify the length, denoted as \( a \), which is the sum of the lengths of each half of the egg (\( a_1 + a_2 \)) This process should also be repeated for the smaller 3D object, if feasible The findings are compiled in Table 2, allowing for a comparison between the measured values and the known diameters listed in Table 1.
To measure the perimeter of an ovoid target, start at a chosen point on its perimeter and place calliper markers along the image of the 3D object until the starting point is reached The spacing between the calliper markers should not exceed 1/20 of the estimated perimeter length, unless the ultrasound system performs a fit for an ellipse or a curved line The perimeter of each of the two half-ellipses can be approximated using a specific equation.
+ b a i π (1) where a i is either a 1 or a 2 , the semi-major axes for a given half of the ellipsoid; b is the mean minor axis of the ellipsoid
The perimeter of the entire egg-shaped object is the sum of the perimeters of the two halves
The perimeter of the circular cross-section is 2πb See Table 1 for expected values for the two objects in Figures B.1 and B.2
On virtually all machines, a value for the enclosed (cross-sectional) area is calculated from the same measurement points defined in the perimeter measurements (see 7.4.3.2)
The measured area values should be compared against the known areas for the 3D-object cross-sections In the largest elliptical and circular cross-sections, respectively, the areas are
A c = + =π + andA c = 0,79b 2 =π b 2 /4 The surface area of the ellipsoids is given by:
b a b a b (2) where ε 1 is the eccentricity (1 – (b/(2a 1 )) 2 ; ε 2 is the eccentricity (1 – (b/(2a 2 )) 2
See Table 1 for expected values for the two objects in Figure B.1
Volume measurements of consistently shaped objects can be determined by assessing their maximum linear dimensions across three orthogonal axes For 3D ellipsoidal objects, the actual volume can be calculated using a specific formula.
To determine the volume, measure the lengths \(a\) and \(b\), where \(a = a_1 + a_2\), and average two perpendicular measurements of \(b\) These measurements can be obtained from two image planes by rotating the transducer 90º between them This calculation method is primarily accurate for volumes resembling two half-ellipsoids with circular cross-sections For estimating the volume of any mass that resembles an ellipsoid with orthogonal axes \(a\), \(b\), and \(c\), use the equation \(V = 0.52abc\) Refer to Table 1 for expected values for the objects illustrated in Figure B.3.
To enhance accuracy and applicability to various shapes, a series of images is captured from evenly spaced ultrasonic scan planes For optimal precision, the distance between these planes should be less than the width of the ultrasonic scan-plane thickness at its focus, ideally under half of that width The volume of the 3D object can be estimated by treating it as a collection of cylinders, where the base area corresponds to the area measured in each ultrasonic scan plane and the height equals the separation between the planes This involves multiplying the cross-sectional area of the object in each plane by the scan separation and summing the volumes for all slices, although this process can be labor-intensive when done manually.
More sophisticated volume-measurement algorithms are implemented on most 3D cross- section imaging systems and should be tested
Table 1 – Expected values for the two ellipsoidal objects in Figure B.3
Perimeter Area Surface area Volume cm cm cm 2 cm 2 cm 3
The perimeters and half-perimeters here are calculated from the exact elliptical integrals, rather than from
Equation 1. a Half the perimeter is a test of curved path length
Table 2 – Suggested table of reported values
Volume cm % cm % cm 2 % cm 2 % cm 3 %
To analyze the measurement, input the value as a percentage of the expected value from Table 1 Each measurement algorithm under study should have this form completed It is beneficial to populate the table with values derived from linear dimensions \(a_1\), \(a_2\), and \(b\) using equations 1, 2, and 3, as well as from any other measurement methods available in the imaging system Similar methodologies are applied in 3D fetal biometry Additionally, this table can be utilized for measurements involving curved lines, areas, and volumes from filament test objects.
8 Measurement of point-spread and line-spread functions
General
The high-contrast resolution of imaging systems is primarily defined by the point-spread function (PSF), which represents the system's response to a high-contrast point target Typically, for most optical systems, the PSF is singular, symmetric, and isotropic, making it a reliable measure for characterizing the system's impulse response.
Ultrasound imaging differs from optical systems as it generates a point spread function (PSF) and line spread function (LSF) that are neither singular nor isotropic These functions are asymmetrical, exhibiting varying axial and lateral dimensions, and their characteristics change with distance from the transducer, or depth in the image Consequently, to accurately assess the imaging performance of the system at specific locations along the beam axis, multiple measurements of the PSF and LSF at various positions and depths are necessary.
The individual Point Spread Function (PSF) or Line Spread Function (LSF) cannot serve as the system's impulse response Variations in the PSF and LSF with depth result in differing resolution levels based on the position within the ultrasound scan.
Ultrasound imaging presents different challenges when dealing with high-contrast versus low-contrast targets High-contrast targets, such as thin wires or air microbubbles, allow ultrasound to capture small structures, resulting in characteristic smears linked to the Point Spread Function (PSF) or Line Spread Function (LSF) In contrast, low-contrast targets like tissue produce a grainy speckle pattern instead of distinct point images, due to the coherent nature of the ultrasound beam The PSF dimensions at a specific depth indicate the system's high-contrast resolution For low-contrast resolution, tissue-mimicking phantoms are utilized, and contrast-detail performance is assessed, though this document focuses solely on high-contrast behavior of the ultrasound PSF or LSF, with low-contrast measurements addressed in a future standard (IEC 61391-2).
This document measures the Point Spread Function (PSF) and Line Spread Function (LSF) at two specific levels: Full Width at Half Maximum (FWHM), which is defined at –6 dB from the peak, and at –20 dB Understanding the relationship between grey level and signal intensity is essential for these measurements (refer to section 8.4.5).
The resolution of a two-target phantom, whether axial, lateral, or elevational, can be estimated by analyzing the Point Spread Function (PSF) or Line Spread Function (LSF) This involves combining two PSF or LSF curves based on their distance apart and identifying the point of separation where a -6 dB dip occurs in the combined curves For a more precise simulation of the two-target response, it is essential to sum the radiofrequency (rf) signals at each point in the PSF or LSF for each assumed target.
Test methods
The test procedures necessitate specific items, including test objects outlined in Annex A and Annex C with targets positioned accurately, a tank filled with degassed liquid, and digitized images as detailed in sections 6.2.2 and 6.3.5.
The specifications of these devices are given in Annexes A and C.
Instruments
This subclause outlines the chosen instruments designed for testing real-time ultrasonic scanners, enabling evaluation without the need for electronic signals to be transmitted to or from the scanners.
A test object should contain structures that allow the following machine features to be measured: a) axial resolution; b) lateral resolution; c) scan slice thickness;
Examples of test objects are given in Annexes A and C
A tank of degassed working liquid is required Under circumstances that allow a deviation of the speed of sound (see 5) the working liquid can be replaced by water to facilitate handling
An image digitizer as described in 6.2.2 is required.
Test settings
Due to the numerous combinations of scanner settings and transducers, it is impractical to test every possible configuration Consequently, tests are conducted for specific settings and each transducer The scanner setup follows the outlined procedures, ensuring that the ultrasonic beam is focused over a wide range to achieve optimal average resolution across all visible targets.
8.4.2 Display settings (focus, brilliance, contrast)
To achieve optimal image quality, first, set the focus to sharp and lower the brilliance and contrast controls Gradually increase the brilliance until the echo-free zone reaches the faintest visible grey shade, as indicated by the grey-shade bar on the display Next, enhance the contrast to maximize the range of grey shades in the image Verify the sharpness of the focus, and if necessary, repeat the process for adjustments If the system supports multiple focal zones, ensure all are activated for the required measurement range.
In the sensitivity settings of a scanner, the nominal frequency of the transducer is established, and any suppression or reject control is fine-tuned to display the smallest signals possible The output power and gain are minimized to ensure an adequate image is produced, while the ATGC controls are adjusted to eliminate automatic gain control Finally, the overall gain control is increased until the target signal is clearly visible on the display.
Digitized images will be used, as described in 6.2.2 and 6.3.5
8.4.5 Calibration of system characteristic curve
A reliable technique for calibrating the system characteristic curve, which represents the output signal as a function of the acoustic-pressure signal at the transducer, involves using planar boundaries with graded reflectivity This method utilizes a straightforward calibration phantom that features specular reflectors arranged parallel to one another and perpendicular to the ultrasound beam's intended path, positioned at various distances from the transducer within a tissue-mimicking material Each distance includes at least two reflectors with varying levels of reflectivity.
One plate has a reflection coefficient that is 10 dB lower than the other, achieved by using 342 stainless steel for one plate and plexiglas for the other The steel plate is positioned on each step of staircase A, while the plexiglas is on staircase B, which is adjacent At standard diagnostic frequencies, the first step is located 1 cm below the surface, followed by steps at 2 cm and 5 cm depths, and additional steps at 10 cm and 18 cm depths The area above these steps is filled with tissue-mimicking material, characterized by a speed of sound of \(c = (1,540 \pm 6) \, \text{m/s}\), a density of \(\rho = (1.05 \pm 0.05) \, \text{g/cm}^3\), and an attenuation coefficient slope of \(\alpha/f = (0.5 \pm 0.05) \, \text{dB cm}^{-1} \text{MHz}^{-1}\).
Echo signals resulting from backscatter in tissue-mimicking materials must be at least 20 dB weaker than the echo signal from the weakest reflecting interface across all tested waveform frequencies.
To calibrate a system's sensitivity control, gain, or output, the phantom is imaged at a range beyond the transducer's focal zone, ensuring the weaker reflector on staircase B is about 6 dB above the electronic noise level If signal levels are excessive, select a deeper step and minimize the TGC Signal levels, recorded 10 dB apart from staircases A and B, are used to adjust the system's sensitivity control until the signal from staircase B matches that of staircase A at the previous setting Consequently, the signal from plate A is now 10 dB higher than that from plate B at the initial control setting.
By repeating this process, the entire signal dynamic range can be calibrated in 10-dB steps
Calibrating the system's sensitivity control by recording its settings enables relative adjustments Interpolating these control settings facilitates the calibration of signal changes in finer increments.
A secondary method for constructing the phantom involves using high-density polyethylene (LB-861) for one staircase and low-density polyethylene (NA-117) for the other This approach aims to generate weaker signals that fall within the linear range of most diagnostic ultrasound systems It is essential to compare this method against the primary one to ensure accuracy The impedance of the high-density material is 2.33 x 10^6 kgm⁻² s⁻¹, while the low-density material has an impedance of 1.79 x 10^6 kgm⁻² s⁻¹.
The reflection coefficients of tissue-mimicking gels must differ by at least 10 dB To achieve weaker echoes and finer gradations, additional staircases with varying reflection coefficients can be utilized, but they require calibration against the primary method In cases of low echogenicities, it is crucial to maintain precise temperature control of the medium to ensure relative echogenicity remains within ±0.5 dB.
Electrical signal injection offers a convenient method for testing transducers, either directly or through acoustic coupling with a calibrated transmitter and receiver While direct injection connectors can be costly and tailored to specific systems, acoustic coupling requires meticulous execution Nonetheless, when referenced to the primary testing method, signal injection proves practical for real-world applications.
Point targets for PSF measurements utilize various target sizes, complementing the primary method This approach highlights the high-frequency components of the scanner's signal, as the scattering cross-section increases with frequency Additionally, it is feasible to evaluate a scanner using this method, given that similar emphasis is observed in signals from distributions of small scatterers.
According to theory, targets significantly smaller than the wavelength exhibit uniform frequency dependence across all diameters This allows for the creation of a series of reflectors with specific scattering ratios For instance, using sphere diameters that increase by 26%—such as 10, 12.6, 15.87, 20, 25.2, 31.75, 40, 50.4, 63.5, and 80 microns—can yield reflectivities that increase in increments of 6 dB, covering a total range of 54 dB While this method has been applied to dense distributions of small scattering particles, detailed documentation on the practical implementation for individual particles remains limited.
Both materials can be sourced from USI Corporation in Marlboro, MA, along with other suppliers These products serve as suitable commercial options This information is provided for the convenience of users and does not imply any endorsement by IEC.
Similarly, a series of flat-ended wires can provide calibrated reflectivities Here the ratio of diameters should be 2 = 1,41, to obtain a step of 6 dB For 5 MHz, diameters in the range
200 à to 1 600 à could span a range of 36 dB with a precision of circa 1 dB.
Test parameters
To measure point- or line-spread functions, an appropriate target is scanned in a test tank
A standard point target, represented by the tip of a flat-ended wire viewed on end, and a standard line target, typically a filament or wire oriented perpendicular to ultrasound propagation, are essential for evaluating system performance The dimensions of the point-spread function (PSF) are measured in three directions for comprehensive assessment in various settings Line-spread functions (LSFs) are preferred in field specifications due to their efficiency, requiring fewer targets while providing necessary spatial spread-function measurements across the field-of-view, influenced by the system's focal and frequency settings Although actual resolution with target pairs may be favored for visual assessments and non-linear signal processing, PSF and LSF measurements are interconnected and crucial for determining high-contrast resolution in imaging systems Furthermore, PSF, LSF, and resolution can be derived from a complete set of measurements, provided the ultrasound system's characteristic curve is calibrated The term "spot size" is defined in this standard to clarify the width of a PSF or LSF for those unfamiliar with spread-function terminology.
Current PSFs and LSFs provide relative measurements, while quantitative assessments of an imaging system's sensitivity to a defined point target or pressure source are possible However, the limited experience with these techniques prevents standardization at this time These quantitative measurements are intricately linked to the high-contrast axial resolution and axial LSF of the imaging system at specific depths within a phantom.
For all such measures, the sensitivity and angulation of the scanner is adjusted so that the maximum signal from the target is clearly visualized during each test
When a line target is positioned perpendicularly to the ultrasonic scan plane, the intersection of the line with the scan plane defines the lateral line-spread function However, this setup has two key limitations: first, the line-spread function can only be assessed within the ultrasonic scan plane; second, the intensity is highly dependent on the orientation of the ultrasonic beam relative to the line target.
To overcome this limitation a spherical target has been used However, now interference between ultrasonic paths inside the sphere leads to complicated patterns of angular scattering
[24] unless the sphere is extremely small and a relatively weak reflector
An effective alternative for measuring resolution is a vertically mounted single filament or wire on a micromanipulator, enabling movement through the field-of-view This setup allows for precise measurement of the spot size at any location within the field-of-view Additionally, it eliminates the challenge of aligning multiple filaments perpendicularly to the ultrasound beam simultaneously, ensuring maximum signal recording from each filament.
When the point is significantly smaller than the wavelength or the diameter of the ultrasonic beam at the target's position, the point-spread function (PSF) is measured independently of the target's properties It is important to note that the response width from the PSF for a point target is approximately equal to that of the line spread function (LSF) for a line target Research indicates that scattering strength varies smoothly with angle and frequency, leading to proposed methods for constructing targets Two simpler alternatives are illustrated in Figure C.4.
From the theory of the scattering by the front plane (equation 3 of [25].), the value for the back- scattering cross section,σ, is found:
Z Z r σ k (4) where k is the circular wave number; (= 2 π / λ in which λ is the wavelength in the medium) r is the radius of the wire;
Z m is the characteristic acoustic impedance of the wire material;
Z w is the characteristic acoustic impedance of the surrounding medium (water)
Scattering is directly related to the square of the frequency and the fourth power of the diameter, provided the target's diameter is sufficiently small to avoid phase differences At small angles, the angle between the front plane and the beam direction has minimal impact, while at larger angles, scattering decreases Reducing the target size further lessens the effect of the angle on scattering.
At oblique angles, ultrasound scattering strength on the wire's front end diminishes Consequently, for sector scanners, reorienting the transducer or target is essential to compare reflected intensity measurements across different field areas However, quantitative beam width measurements can still be obtained without the need for reorientation.
Figure C.4 shows a sketch of two point targets made from dental wire (diameter 0,24 mm)
The left target is designed for pulsed ultrasound, capturing echoes from the wire tip and supporting tube at different times, while the supporting tube, with an outer diameter of 1.0 mm, ensures mechanical stability In contrast, the right target is meant for continuous wave (CW) ultrasound, where only the wire tip, located within the ultrasound beam, contributes to backscattering.
8.5.2 Axial PSF and LSF dimensions and axial resolution
To accurately measure the dimensions of the Point Spread Function (PSF) and Line Spread Function (LSF) in the axial direction, assessments should be conducted at various depths using appropriate phantoms, as illustrated in Figures C.3, C.4, A.1, A.2, and C.1 It is essential to utilize Figure C.4 when reverberations from the support wire interfere with measurements Measurements can be taken in increments smaller than 1/3 of the axial field-of-view, depending on the transducer's imaging range It is crucial to identify and document sources of variability, as outlined in section 6.4.2 Axial resolution at a specific depth can be determined by the minimum separation of two distinguishable filaments in Figure C.2, which occurs before they blur together, and quantitatively, by observing a minimum 6 dB drop in signal between the filaments.
To achieve optimal signal from the target(s), align the transducer with the beam perpendicular to any filament target, as this alignment is crucial Adjust the system sensitivity controls to ensure all targets are clearly visible without saturation Capture multiple B-mode images or image volumes of the target or phantom, then digitize the best image For certain phantoms, ensure the scan window is parallel to the filament targets Additionally, confirm that the ultrasonic transducer is securely positioned on the surface, and when using a curved transducer face, ensure the central scan line is perpendicular to the scan window, with ultrasound beams directed normal to the filaments.
To measure the Point Spread Function (PSF), a 2D scan of the point target is necessary at each location, unless proper alignment with the beam's central axis or image plane is achieved An image of the test object is captured, displaying the imaged targets across the entire depth range of the ultrasonic transducer Utilizing image-processing software, a vertical brightness profile is plotted across multiple scan lines, moving laterally until the maximum amplitude profile of the target is identified For this maximum amplitude, the full-width-at-half-maximum (FWHM) and –20 dB width are measured and recorded for the specific filament depth Subsequently, the axial response width is measured for all imaged-target positions within the phantom image After obtaining all axial response widths, a plot of FWHM or –20 dB width versus depth is created, illustrating the variation in axial PSF or LSF with depth Typically, most systems exhibit minimal variation in axial PSF or LSF as depth changes.
The profile length (PSF or LSF) at a relative signal level close to the anticipated axial clutter level, or at a specific number of wavelengths, provides valuable insights, provided that the phantom's artefacts permit such measurements.
8.5.3 Lateral PSF- and LSF-width and lateral resolution
The dimensions of the Point Spread Function (PSF) and Line Spread Function (LSF) in the lateral direction are crucial for determining the high-contrast lateral resolution of an imaging system at specific beam positions Appropriate phantoms for PSF measurement are illustrated in Figures C.3 and C.4, with the latter used to mitigate issues from support wire reverberations LSF measurements utilize phantoms shown in Figures A.1, A.2, or C.1 While multi-target lateral resolution measurements are uncommon and not covered here, measurements can be conducted across the entire axial imaging range of the ultrasonic transducer in increments smaller than 1/7 of the axial field-of-view, maintaining a distance of 1/8 of the image width from the lateral edge For volumetric imaging systems where PSF may vary with elevational position, measurements should be taken at a distance of 1/8 of the volume thickness from the elevational edge and at a corner of the volume, 1/8 of the volume width and thickness from the lateral and elevational edges, respectively It is essential to identify and document sources of variability as outlined in section 6.4.2.
To achieve optimal signal reception from the target, ensure the transducer is aligned with the beam perpendicular to any filament target Adjust the system sensitivity controls as outlined in section 8.5.2.2 Capture multiple B-mode images or image volumes of the target or phantom, then digitize the highest quality image It is essential that the ultrasonic transducer is securely positioned on the scanning surface of the test object, maintaining a perpendicular orientation to the object's surface.
For accurate PSF measurements, a 2D scan of the point target is essential at each location unless proper alignment with the beam's central axis or image plane is achieved The LSF or PSF is derived from an image of the test object, capturing targets across the entire depth range of the ultrasonic transducer Image processing software is utilized to create a horizontal brightness profile from multiple scan lines that traverse the imaged point target, moving vertically to identify the maximum amplitude profile Key metrics such as the full-width-at-half-maximum (FWHM) and the –20 dB width are recorded for the filament depth at this maximum amplitude Additionally, measuring the profile width at a signal level close to the expected transducer clutter level can provide valuable insights, provided that the phantom allows for such measurements.