Sparse ultrasound array

Abstract

A medical ultrasound system generating an output ultrasound image with reduced grating lobe artifacts, comprising: a) ultrasound transmitting tranducers that transmit frequency components of ultrasound waves at well separated frequency bands, including a different normal mode frequency of the transmitting transducers; b) ultrasound receiver elements, spaced further from their neighbors than half a wavelength for at least one frequency of the resonant frequency bands, that receive the frequency components of ultrasound waves after they reflect from body tissue, generating signals of each receiver element; and c) a controller to control the transmitting transducers to transmit the frequency components; record signals of the ultrasound received by each receiver element; determine, from signals of the receivers elements, information about reflectivity and how it differs for different frequency components for different locations; and reconstruct the output image, with pixel intensities depending on information about differing calculated corresponding voxel reflectivities for different frequency components.

Description

[0001] SPARSE ULTRASOUND ARRAY

[0002] RELATED APPLICATIONS

[0003] This application claims the benefit of priority of U. S. Provisional Patent Application No. 63 / 751,327 filed on January 30, 2025, the contents of which are incorporated herein by reference in their entirety.

[0004] FIELD AND BACKGROUND OF THE INVENTION

[0005] The present invention, in some embodiments thereof, relates to a method and system for reducing imaging artifacts in an ultrasound image produced with a sparse ultrasound array and, more particularly, but not exclusively, to reducing side lobes and clutter in a high resolution 3-D medical ultrasound image generated using a sparse 2-D ultrasound array.

[0006] To obtain a high resolution 3-D ultrasound image, for example a medical ultrasound image that covers a large field of view, an ultrasound array with a large number of receiver transducers may be used. For example, to satisfy the Nyquist criterion, the ultrasound transducers may be spaced less than half an ultrasound wavelength apart, with the dimensions of the array comparable to its distance from the field of view, and the resolution may be determined by the ultrasound wavelength. But using such a large number of transducers may be too expensive to be practical, and processing the large amount of data received by so many transducers may also be impractical. Accordingly there is a need for an ultrasound image processing method that produces good quality images, relatively free of imaging artifacts such as grating lobes and clutter, even when the ultrasound array is sparse, with spacing between transducers that does not satisfy the Nyquist criterion.

[0007] U. S. Patent Application Publication No. 2022 / 0354465 to Shin et al describes an ultrasound transducer array that emits ultrasound waves, such as a cylindrical array arranged around an intravascular catheter, and receives echoes of the emitted ultrasound waves from discontinuities arising from tissue structures, red blood cells, and other features of interest. The array may not satisfy the Nyquist criterion. The received data is duplicated to produce a plurality of duplicate data sets. The data sets are passed through different narrowband filters, producing sets of receiver data each covering a different narrow frequency range, each narrower than the frequency range of the emitted ultrasound. Ultrasound images are generated from each of the filtered duplicate data sets. A processor generates a grating-lobe minimized image, with reduced image artifacts and clutter, by taking the minimum intensity, at each pixel, from the images generated from the differentfiltered data sets. The grating-lobe minimized image can be displayed, or combined with the original image to include image features lost or reduced in the grating-lobe minimized image.

[0008] U. S. Patent No. 11,408,987 to Vignon et al, which includes some of the same inventors as U. S. Patent Application Publication No. 2022 / 0354465, describes some similar devices and methods for producing grating-mode minimized ultrasound images.

[0009] Other devices and methods for producing images with reduced grating lobes or side lobes from sparse ultrasound arrays that do not satisfy the Nyquist criterion are given by U. S. Patent No.

[0010] 5,808,962 to Steinberg et al; by U. S. Patent 6,524,253 to Abend; by U. S. Patent Application Publication No. 2004 / 0193050 to Ogawa; and by U. S. Patent No. 10,281,568 to Oelze.

[0011] Lateral mode ultrasound transducers are described by Victor Giurgiutiu and Andrei Zagrai, “Characterization of Piezoelectric Wafer Active Sensors,” Journal of Intelligent Material Systems and Structures 11, 959-975 (2000); by Suresh Bhalla et al, “Electromechanical Impedance Modeling for Adhesively Bonded Piezo-transducers,” Journal of Intelligent Material Systems and Structures 15, 955-972 (2004); and by B. Lin et al, “Modeling and testing of PZT and PVDF piezoelectric wafer active sensors,” Smart Mater. Struct. 15, 1085-1093 (2006).

[0012] Using a spatial coherence factor to improve image quality in ultrasound imaging is described by Zainab Alomari, “The Effect of Imaging Parameters on the Performance of Coherence Factor in Plane-Wave Imaging,” Proceedings of the 9th World Congress on Electrical Engineering and Computer Systems and Sciences (EECSS ’23), Brunel University, London, United Kingdom - August 03-05, 2023, Paper No. ICBES 129, available at doi:10.11159 / icbes23.129; and by James Long, Gregg Trahey, and Nick Bottenus, “Spatial Coherence in Medical Ultrasound: A Review,” Ultrasound Med. Biol. (2022 June) 48(6): 975-996, available at doi: 10.1016 / j(dot)ultrasmedbio.2022.01.009.

[0013] Tung Manh et al, “Dual Frequency Hybrid Ultrasonic Transducers — Design and Simulations,” in 2016 IEEE Ultrasonics Symposium Proceedings, studies a design concept of a dual frequency ultrasonic transducer, in which different technogies are used for each band. A hybrid structure consists of a piezoelectric stack which is used for the lower frequency band, and a Capacitive Micromachined Ultrasound Transducer (CMUT) array placed on the top surface is used for the higher frequency band. The work shows a potential of combining piezoelectric and MUT technologies in ultrasound transducer design for use in medical applications.

[0014] Xiaofei Luo et al, “Stack-Layer Dual-Element Ultrasonic Transducer for Broadband Functional Photoacoustic Tomography,” Frontiers in Bioengineering and Biotechnology 9, Article 786376 (Oct. 2021), describes a stack-layer dual element ultrasonic transducer where the central frequencies of the two piezoelectric elements are 3.06 MHz (1.54 to 4.58 MHz at -6 dB)and 11.07 MHz (6.35 to 15.78 MHz at -6 dB). The transducer bridges the sensitivity capability of ultrasound and the high contrast of optical methods in functional photoacoustic tomography.

[0015] Additional background art includes U. S. Patent No. 6,457,365 to Stephens et al; Jeffrey Elloian et al, “Flexible ultrasound transceiver array for non-invasive surface-conformable imaging enabled by geometric phase correction,” Scientific Reports 12, 16184 (2022), published by Nature Portfolio, available at https(colon) (slash)(slash)doi(dot)org(slash)10.1038(slash)s41598-022-20721-7; Wang et al, “An array transmitter for dual-frequency contrast enhanced intravascular ultrasound imaging,” in 2014 IEEE International Ultrasonics Symposium Proceedings, 2104-2107; and Zhuochen Wang, K. Heath Martin, Paul A. Dayton, and Xiaoning Jiang, “Real-time ultrasound angiography using superharmonic dual-frequency (2.25 MHz / 30 MHz) cylindrical array: in vitro study,” Ultrasonics 82:298-303 (2018 January), doi: 10.1016 / j.ultras.2017.09.012.

[0016] SUMMARY OF THE INVENTION

[0017] An aspect of an embodiment of the invention concerns a method of generating an ultrasound image with reduced grating lobes, in spite of using a sparse ultrasound receiver array, by comparing images generated from echoes of transmitted ultrasound at different frequency ranges.

[0018] There is thus provided, in accordance with an exemplary embodiment of the invention, a medical ultrasound system that, when it is in good acoustic contact with a body surface, generates an output ultrasound image of a field of view of tissue inside the body, with reduced grating lobe artifacts, the system comprising:

[0019] a) an ultrasound transmitting module comprising one or more ultrasound transmitting transducers that, when the system is in good acoustic contact with the body surface, transmits into the field of view two or more different frequency components of ultrasound waves, each frequency component limited to a different one of well separated resonant frequency bands, each of the resonant frequency bands including a different normal mode frequency, associated with that frequency component, of the one or more ultrasound transmitting transducers;

[0020] b) an array of ultrasound receiver elements, at least some of them spaced further apart from their neighbors than half a wavelength for at least one frequency of the resonant frequency bands at a sound speed of 1540 meters per second, that, when the system is in good acoustic contact with the body surface, receive the frequency components of ultrasound waves after they reflect from tissue in the field of view, the array generating signals of the ultrasound received by each receiver element; andc) a controller configured to 1) control the transmitting module to transmit the frequency components of ultrasound waves, 2) record a signal of the ultrasound received by each of the ultrasound receiver elements, 3) calculate, from the generated signals of the receivers elements, information about reflectivity and how it differs for different frequency components for given locations in the field of view, and 4) reconstruct the output image, with pixel intensities in the output image depending on the information about differing calculated corresponding voxel reflectivities for the different frequency components.

[0021] Optionally, at least one of the one or more ultrasound transmitting transducers is a lateral mode transducer.

[0022] Optionally, the lateral mode transducer comprises a thin slab of piezoelectric material, and electrodes that cover opposite faces of the slab that produce an electric field in the slab, uniform within 20%, in a thickness direction of the slab, that causes the transducer to generate ultrasound waves when the electric field has a frequency equal to a lateral mode resonant frequency of the slab, wherein the slab has an anisotropic shape in a plane perpendicular to the thickness direction.

[0023] Optionally, the shape in the plane perpendicular to the thickness direction is not symmetric under 90-degree rotation in the plane.

[0024] Optionally, the slab has a non-square rectangular shape in the plane perpendicular to the thickness direction.

[0025] Optionally, two frequency components have frequencies at different lateral resonant modes of one transducer of the one or more transducers.

[0026] Optionally, a ratio of the frequencies of the two frequency components differ by more than 3% from 1, 4 / 3, 3 / 2, 5 / 3, 2, 5 / 2, and 3.

[0027] Optionally, the ultrasound transmitting module comprises a plurality of ultrasound transmitting transducers, having between them normal mode frequencies in each of the resonant frequency bands, that transmit ultrasound waves in those resonant frequency bands into the field of view when the system is in good acoustic contact with the body surface, at same times or at different times for different ultrasound transmitting transducers, the ultrasound waves reflecting from the tissue in the field of view, and contributing to the ultrasound waves received by the array of ultrasound receiver elements, and to the signals of received ultrasound waves generated by the array.

[0028] Optionally, the controller is configured to control the transmitter to transmit the different frequency components at different non-overlapping times.

[0029] Optionally, the controller is configured to control the transmitter to transmit the different frequency components at times that at least partially overlap, and the array, the controller or bothare configured to generate and record signals for different frequency components by using hardware or software or both to separate different frequency components in the signals generated by the array of ultrasound receiver elements.

[0030] Optionally, at least one of the frequency components transmitted by the one or more ultrasound transmitting transducers is associated with a thickness mode resonant frequency of one of the one or more ultrasound transmitting transducers.

[0031] Optionally, all of the frequency components transmitted by the ultrasound transmitting module are associated with lateral mode resonant frequencies of the one or more ultrasound transmitting transducers.

[0032] There is further provided, according to an exemplary embodiment of the invention, a method of generating an ultrasound output image of a field of view of tissue inside a body, using a sparse array of ultrasound receiver elements, the method comprising:

[0033] a) transmitting a plurality of ultrasound wave components into the field of view, from one or more transmitting locations for each component, with each component limited to a different one of well separated resonant frequency bands;

[0034] b) receiving reflections of the transmitted ultrasound components from the field of view by the sparse array of ultrasound receiver elements, and recording signals of the ultrasound received by each receiver element as a function of time;

[0035] c) finding from the recorded receiver signals information about corresponding calculated voxel reflectivity as a function of transmitted ultrasound component; and

[0036] d) assigning an intensity to each pixel of the output image, based on the information about the corresponding calculated voxel reflectivity as a function of ultrasound component.

[0037] Optionally, assigning an intensity to each pixel in the output image comprises assigning intensities for which an estimated amplitude of grating lobes in the output image is less than half an estimated amplitude of grating lobes in images made from voxel reflectivities for one ultrasound component.

[0038] Optionally, transmitting the ultrasound wave components is done by one or more ultrasound transmitting transducers, and each of the resonant frequency bands corresponds to a different normal mode frequency of the one or more ultrasound transmitting transducers, corresponding to different normal modes of a same transmitting transducer, or to different normal modes of different transmitting transducers, or to both.

[0039] Optionally, at least one of the normal mode frequencies is a lateral mode of a lateral mode ultrasound transducer.Optionally, the one or more ultrasound transmitting transducers comprise at least two transmitting transducers of different designs with different normal modes, and at least two of the resonant frequency bands each correspond to a normal mode of a different one of the two transmitting transducers with different normal modes.

[0040] Optionally, at least some of the resonant frequency bands each correspond to a different normal mode of one of the transmitting transducers.

[0041] Optionally, transmitting the ultrasound wave components comprises transmitting different components at different time periods that do not overlap.

[0042] Optionally, transmitting the ultrasound wave components comprises transmitting different components at time periods that partly or completely overlap.

[0043] Optionally, finding information about different corresponding voxel reflectivities for each of the transmitted ultrasound components, from the recorded receiver signals, comprises finding a reflectivity for each voxel for each ultrasound component from a beam-formed signal of the receiver signals, filtered by a transmitted pulse shape for that component, and normalized by an energy of the transmitted pulse shape.

[0044] Optionally, assigning an intensity to each pixel of the output image comprises using a decision function that assigns an intensity to a pixel based on a distribution of the corresponding calculated voxel reflectivities for the different ultrasound components, and substantially the same decision function is used for almost all of the pixels in the output image.

[0045] Optionally, a corresponding calculated voxel reflectivity for one or more of the pixels in the output image is a calculated reflectivity for a voxel at a closest location to the pixel in the field of view.

[0046] Optionally, a corresponding calculated voxel reflectivity for one or more of the pixels in the output image is an interpolation of calculated reflectivities for a plurality of voxels at locations near the pixel in the field of view.

[0047] Optionally, the information about differing corresponding calculated voxel reflectivities for the different transmitted ultrasound components comprises calculated reflectivity as a function of voxel and component, and the intensity assigned to one or more of the pixels in the output image is based on corresponding calculated voxel reflectivities for the different ultrasound components that are normalized to a range of calculated voxel reflectivities for other voxels for the same ultrasound component.

[0048] Optionally, the information about differing corresponding calculated voxel reflectivities for the different transmitted ultrasound components comprises calculated reflectivity as a function of voxel and component, and the intensity assigned to one or more of the pixels in the output imageis based on corresponding calculated voxel reflectivities for the different ultrasound components that are smoothed over nearby voxels for the same ultrasound component.

[0049] Optionally, the method comprises normalizing the output image after assigning an intensity to each pixel of the output image.

[0050] Optionally, the method comprises smoothing the output image after assigning an intensity to each pixel of the output image.

[0051] Optionally, the array of ultrasound receivers elements is in acoustic contact with a surface of the body outside the ribs, extending over a plurality of ribs and a plurality of spaces between ribs, and the field of view comprises tissue inside the rib cage, and reflections of ultrasound from the tissue inside the rib cage are received by the array of ultrasound receiver elements substantially only when the ultrasound propagates between the ribs, and not through the ribs.

[0052] Optionally, the information about calculated voxel reflectivities for the different transmitted ultrasound components comprises a reflectivity image of the field of view for each component.

[0053] Optionally, at least some of the receiver elements in the sparse array are spaced further apart from their neighbors than half a wavelength for at least one frequency of the resonant frequency bands at a sound speed of 1540 meters per second.

[0054] Unless otherwise defined, all technical and / or scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the invention pertains. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of embodiments of the invention, exemplary methods and / or materials are described below. In case of conflict, the patent specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and are not intended to be necessarily limiting.

[0055] Implementation of the method and / or system of embodiments of the invention can involve performing or completing selected tasks manually, automatically, or a combination thereof. Moreover, according to actual instrumentation and equipment of embodiments of the method and / or system of the invention, several selected tasks could be implemented by hardware, by software or by firmware or by a combination thereof using an operating system.

[0056] For example, hardware for performing selected tasks according to embodiments of the invention could be implemented as a chip or a circuit. As software, selected tasks according to embodiments of the invention could be implemented as a plurality of software instructions being executed by a computer using any suitable operating system. In an exemplary embodiment of the invention, one or more tasks according to exemplary embodiments of method and / or system asdescribed herein are performed by a data processor, such as a computing platform for executing a plurality of instructions. Optionally, the data processor includes a volatile memory for storing instructions and / or data and / or a non-volatile storage, for example, a magnetic hard-disk and / or removable media, for storing instructions and / or data. Optionally, a network connection is provided as well. A display and / or a user input device such as a keyboard or mouse are optionally provided as well.

[0057] BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING(S)

[0058] Some embodiments of the invention are herein described, by way of example only, with reference to the accompanying figures, including drawings and images, which will be referred to as “drawings.” With specific reference now to the drawings in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of embodiments of the invention. In this regard, the description taken with the drawings makes apparent to those skilled in the art how embodiments of the invention may be practiced.

[0059] In the drawings:

[0060] FIG. 1A schematically shows a field of view of a body and a system for making an ultrasound image of it, according to an exemplary embodiment of the invention;

[0061] FIG. IB schematically shows a spectrum of ultrasound waves transmitted into the field of view, with two well separated resonant frequency bands, according to an exemplary embodiment of the invention;

[0062] FIG. 2 shows a flowchart for a method for making an ultrasound image of a field of view using the system of FIG. 1A, for example using a spectrum of ultrasound waves such as that shown in FIG. IB, according to an exemplary embodiment of the invention;

[0063] FIG. 3A schematically shows details of a transmitting transducer, an array of receiving transducers, and a point in the field of view from which the transmitted ultrasound reflects, according to the system of FIG. 1A;

[0064] FIG. 3B shows an example of an ultrasound pulse shape that could be used by the system shown in FIG. 1 A and FIG. 3A, using the method of FIG. 2 and FIG. 4;

[0065] FIG. 4 shows a flowchart with details of how voxel reflectivities are generated from the received signal data, according to the method of FIG. 2;

[0066] FIG. 5 shows an array of receiver transducers, with transducers missing from a horizontal stripe in the center of the array, which may represent a rib blocking some of the ultrasound reflected from a field of view inside the rib cage;FIG. 6 shows a two-dimensional slice of a field of view, with a single point reflector at the center;

[0067] FIG. 7 shows an image of the field of view of FIG. 6 made using the receiver array of FIG.

[0068] 5, using a higher frequency band of ultrasound, around 1.08 MHz, showing grating lobes both horizontally and vertically, due to the array not satisfying the Nyquist criterion both vertically and horizontally;

[0069] FIG. 8 shows an image of the field of view of FIG. 6 made using the receiver array of FIG.

[0070] 5, using a lower frequency band of ultrasound, around 0.54 MHz, showing grating lobes especially in a vertical direction, due to the array not satisfying the Nyquist criterion vertically, due to the missing transducers in the horizontal stripe;

[0071] FIG. 9 shows a reduced image of the field of view of FIG. 6 made using the receiver array of FIG. 5, by taking the minimum intensity at each voxel of the image in FIG. 7 and the image in FIG. 8; and

[0072] FIG. 10 is a block diagram giving details of the controller shown in FIG. 1A.

[0073] DESCRIPTION OF SPECIFIC EMBODIMENTS OF THE INVENTION

[0074] The present invention, in some embodiments thereof, relates to a method and system for reducing imaging artifacts in an ultrasound image produced with a sparse ultrasound array and, more particularly, but not exclusively, to reducing side lobes, such as grating lobes, and clutter in a high resolution 3-D medical ultrasound image generated using a sparse 2-D ultrasound receiver array.

[0075] An aspect of some embodiments of the invention concerns a system and method for generating ultrasound images, using a sparse receiver array which does not satisfy the Nyquist criterion, with reduced grating lobes. The field of view is illuminated with two or more components of transmitted ultrasound waves, each component in a different frequency band, well separated from the frequency bands of the other components. Echoes of the transmitted ultrasound from the field of view are received by receiver elements of the ultrasound receiver array, and the received signals for the different elements are recorded. The recorded received signals for the different elements are used to obtain information about voxel reflectivities in the field of view, including information about corresponding calculated voxel reflectivity as a function of transmitted ultrasound component. In some embodiments of the invention this information includes the corresponding calculated voxel reflectivity for each ultrasound component, for each voxel, but in other embodiments of the invention the information may include only more general information, such as the mean corresponding calculated voxel reflectivity over all ultrasound components, foreach voxel. A reduced grating lobe image is then constructed from the information about corresponding voxel reflectivity as a function of transmitted ultrasound component. This method can be used to reduce the grating lobes because the locations of the grating lobes in an ultrasound image made with a sparse array depend on the ratio of ultrasound wavelength to the pitch of the transducers in the array, so ultrasound of different well separated frequency ranges, and hence different wavelengths, tends to produce images with grating lobes at different locations, which can be averaged out.

[0076] Optionally, all the transducers have the same resonant normal modes, for example they have identical design, optionally including both the transmitting and receiving transducers, and the different resonant frequency bands are generated by exciting different resonant normal modes of the transducers. Optionally the transducers are lateral mode transducers that have anisotropic shape in their lateral plane, with different lowest normal modes along different axes in the lateral plane. For example, the transducers are rectangular in shape in their lateral plane, with different length and breadth along their two lateral axes.

[0077] Alternatively, there are at least two different designs of transducers, which have different sets of normal modes, and the at least two frequency bands of ultrasound are generated by exciting normal modes of the different transducer designs. Optionally, all the normal modes used for the different frequency bands are lateral modes of the transducers. Alternatively, at least some normal modes excited for the different frequency bands are thickness modes of the transducers. If thickness modes are used, then optionally the transducers are all lateral mode tranducers, with thickness much less than their lateral dimensions, and they have thickness modes at much higher frequencies than their lateral mode frequencies. Alternatively, at least some of the transducers are not lateral mode transducers, but have thickness comparable to their lateral dimensions, with thickness modes at frequencies comparable to their lateral modes, optionally with no sharp distinction between thickness modes and lateral modes.

[0078] Before explaining at least one embodiment of the invention in detail, it is to be understood that the invention is not necessarily limited in its application to the details of construction and the arrangement of the components and / or methods set forth in the following description and / or illustrated in the drawings and / or the Examples. The invention is capable of other embodiments or of being practiced or carried out in various ways.

[0079] Method for producing reduced grating lobe images

[0080] Referring now to the drawings, FIG. 1A schematically shows a system 100 for producing an ultrasound image of a field of view 102, with reduced grating lobes. The field of view is, forexample, part of the body of a patient, for whom a medical ultrasound image is being made. An ultrasound transmitting transducer 104 transmits ultrasound waves into field of view 102, shown schematically in FIG. 1A as arrows radiating from transducer 104 into field of view 102. Optionally, there are one or more other transmitting transducers, for example transducer 105 in FIG. 1 A, that may be located some distance away from transducer 104, that also transmit ultrasound waves into field of view 102. Various small regions (voxels) in field of view 102 reflect the transmitted ultrasound waves as echoes, which are detected by an array 106 of ultrasound receiver elements. Each receiver element in array 106 is an ultrasound transducer, and optionally at least some of the receiver elements have the same design as at least some of the one or more ultrasound transmitting transducers. Array 106 generates signals corresponding to the reflected ultrasound waves received by each receiver element in the array, which are communicated to a controller 108, such as a computer, which records them. Controller 108 uses the recorded signal data to calculate an ultrasound image 110 of the field of view, with reduced grating lobes. Further details of the hardware of controller 108, in an exemplary embodiment of the invention, are shown below in block diagram 1000 in FIG. 10. The receiver array is a sparse array, with transducers that are far enough apart, relative to an ultrasound wavelength, that it does not satisfy the Nyquist criterion, and without using special image reconstruction methods the ultrasound image it generates would have strong grating lobes. For example, neighboring transducers are more than half a wavelength apart along one or both axes in the plane of the array, or neighboring transducers are more than one wavelength apart, or more than two wavelengths apart, or more than three wavelengths apart along one or both axes. Controller 108 also optionally calculates one or more other images 112, for example images based on ultrasound in different frequency bands. The images may be displayed on a monitor, for example, or printed, so they can be examined by a radiologist in order to diagnose a medical condition of the patient. Controller 108 optionally also controls transmitting transducer 104 and any other transmitting transducers, for example controlling the timing of the ultrasound transmission, and other characteristics of the transmitted ultrasound, such as its frequency spectrum, and the power as a function of time. Details of how the received signal data is used to generate the ultrasound images are provided below, starting with flowchart 200 in FIG. 2.

[0081] Optionally the transmitted ultrasound from the transmitting transducer or transducers has a frequency spectrum that consists of a plurality of well separated resonant frequency bands, for example with each band corresponding to a normal resonant mode of the one or more transmitting transducers. The use of this plurality of well separated frequency bands may be important for producing a reduced grating lobe image when the receiver array is sparse, as will be describedbelow, and may provide advantages in this respect over the prior art of Shin et al, cited above, where only a single frequency band associated with a single normal mode is used.

[0082] Figure 1B shows a plot 120 of an exemplary ultrasound spectrum that illustrates how these “well separated frequency bands” may be defined. Horizontal axis 122 shows the frequency of the ultrasound, and vertical axis 124 shows the transmitted ultrasound power, both in arbitrary units. Curve 126 shows the transmitted power as a function of frequency. There is a first normal resonant frequency 128, and a second normal resonant frequency 130. Around frequency 128 there is a resonant frequency band 132, and around frequency 130 there is a resonant frequency band 134. Between frequency bands 132 and 134, there are no other resonant frequency bands, and frequency bands 132 and 134 are separated by a separation range 136. The ultrasound in frequency band 132 is defined as the first frequency component of the transmitted ultrasound, and the ultrasound in frequency band 134 is defined as the second frequency component of the transmitted ultrasound. The integrated transmitted ultrasound power in each of bands 132 and 134 is optionally at least 1.5 times as great, or at least 2 times as great, or at least 3 times as great, or at least 5 times as great, or at least 10 times as great, as the integrated transmitted ultrasound power in separation range 136. Figure 1B has been drawn to show the transmitted power in bands 132 and 134 only a few times greater than the transmitted power in separation range 136, to illustrate these ratios, but in practice there is likely to be much less power in the separation range compared to the power in the frequency bands. For example, if the pulses have the gaussian pulse shape shown below in FIG. 3B, with full width at half peak of 1 microsecond, and if the peak frequencies in the frequency bands are separated by at least 1 MHz, then the ultrasound power in the separation range may be less than 1% of the power in each frequency band. The width of separation range 136 is optionally at least 1.5 times as great, or at least 2 times as great, or at least 3 times as great, or at least 5 times as great, or at least 10 times as great, as the width of band 132 and the width of band 134, defined for example as the full width at half peak. Having the power in the separation range that is smaller relative to the power in the frequency bands has the potential advantage that the images made from the two frequency bands will have grating lobe artifacts that are more well defined and more different from each other, than if the power in the separation range is greater relative to the power in the frequency bands, and this may make it possible for the two images to produce an image with grating lobes that are more reduced. Similarly, having the separation range greater relative to the width of the frequency bands has the potential advantage that the images made from the two frequency bands will have grating lobe artifacts that are more well defined and more different from each other, than if the separation range is smaller relative to the width of the frequency bands. But there may be a practical limit to how much the power in the separation range can be reduced, dueto noise for example. And there may be a practical limit to how narrow the frequency bands can be made, due to damping effects in the transducers for example.

[0083] Optionally, the natural width of one or both of the resonance frequencies is much less than the frequency, for example the Q of one or both of the resonance frequencies is greater than 3, or greater than 5, or greater than 10. Optionally, the first resonant frequency band 132 has a width smaller than or comparable to the natural width of first resonant frequency 128, or at least most of the transmitted power in band 132 is within the natural width of first resonant frequency 128, and optionally the second resonant frequency band 134 has a width smaller than or comparable to the natural width of second resonant frequency 130, or at least most of the transmitted power in band 134 is within the natural width of second resonant frequency 130. This has the potential advantage that the ultrasound may be generated more efficiently than if the transducers are excited substantially at frequencies that are outside the natural widths of their resonant frequencies.

[0084] Optionally there are one or more other resonant frequency bands, corresponding to other resonant modes of the one or more transducers, at frequencies below band 132, and / or at frequencies above band 134. The ultrasound in each of those other resonant frequency bands is defined as a separate frequency component of the transmitted ultrasound. In this case, optionally any two neighboring resonant frequency bands both have at least 1.5 times the integrated power, or at least 2 times the integrated power, or at least 3 times the integrated power, or at least 5 times the integrated power, or at least 10 times the integrated power, of the separation range between them, and optionally the width of the separation range between any two neighboring resonant frequency bands is at least 1.5 times as great, or at least 2 times as great, or at least 3 times as great, or at least 5 times as great, or at least 10 times as great, as the width of either of those neighboring resonant frequency bands. As will be described below, the method of generating a reduced grating lobe image may involve determining or estimating how the calculated reflectivity of a voxel varies for different frequency components of the ultrasound.

[0085] Optionally, the resonant frequencies used are between 0.25 MHz and 1 MHz, or between 0.5 MHz and 2 MHz, or between 1 MHz and 5 MHz, or between 2 MHz and 10 MHz, or in more than one of these ranges, or at lower or higher frequencies. Using an ultrasound frequency at least 0.25 MHz has the potential advantage that the wavelength is short enough to produce an image with resolution of a few millimeters or better. Using an ultrasound frequency less than 10 MHz has the potential advantage that the ultrasound can propagate for at least a few centimeters into the body without very great attenutation, from an ultrasound array on the surface of the body.

[0086] Figure 2 shows a flowchart 200 of a method of acquiring ultrasound imaging data and using it to generate ultrasound images, according to an exemplary embodiment of the invention, forexample using system 100 of FIG. 1A. At 202, transmitting transducer 104, or a plurality of transmitting transducers, transmits a plurality of frequency components of ultrasound, at well separated resonant frequency ranges, into the field of view. Optionally, different components of ultrasound are transmitted at different ranges of time, with no overlap, and / or are received by the receiver array at different ranges of time with no overlap for any of the receiver elements. This has the potential advantage that it may be easier to distinguish the echoes from the different frequency components, in the recorded receiver signal data, if they are transmitted at different times, and received by the receiver array at different times. Alternatively, the different frequency components of ultrasound are transmitted at times that overlap at least partly, or overlap completely, and / or the different frequency components are received by at least some elements in the receiver array at times that overlap at least partly, or overlap completely. In this case, the contributions of the different frequency components to the voxel reflectivities may be distinguished by filtering of the receiver signals, as will be explained below in FIG. 4. This has the potential advantage that it may be possible to acquire data for all the frequency components more quickly than if the different frequency components are transmitted and / or received at non-overlapping times.

[0087] At 204 echoes of the transmitted ultrasound from the field of view are received by the receiver elements in receiver array 106. The receiver array sends signals to controller 108, with signal data of the ultrasound echoes as a function of time received by each of the receiver elements, and the data is recorded by the controller. Optionally, at least some of the receiving elements are resonant at one or more of the resonant frequency bands transmitted into the field of view. For example, those receiver elements have a same or similar design to a transmitting transducer that has resonant frequency associated with that resonant frequency band, and that transmits that frequency component of ultrasound, and the receiver element also has a resonance frequency at or near that frequency, optionally with the same or a similar natural width. Optionally, different receiver elements are resonant at different ones of the resonant frequency bands. Alternatively, at least some receiver elements are resonant at two or more different resonant frequency bands. Having receiver elements that are resonant at one of the resonant frequency bands has the potential advantage that the receiver element may be more sensitive to ultrasound in that resonant frequency band. Alternatively, at least some of the receiver elements have a relatively flat frequency response to ultrasound over the transmitted spectrum, which has the potential advantage that they may be able to compare the received ultrasound power at different frequencies more accurately. Optionally, most of the receiver elements are not highly resonant at frequencies outside the ranges of frequency where most of the ultrasound power is being transmitted, or at least they are not highly resonantonly at those frequencies, which might result in the receiver elements being very insensitive to ultrasound in the ranges of frequency where most of the ultrasound is being transmitted.

[0088] At 206, the recorded signal data from the receiver elements is used to obtain information about calculated voxel reflectivities in the field of view, and in particular to obtain information about calculated voxel reflectivity as a function of frequency component of the ultrasound. Methods for obtaining this information are described below in the description of FIG. 4. As used herein, “calculated” can include computed, or mathematically determined by any method.

[0089] Here, “calculated voxel reflectivities” refers to reflectivities as calculated, for example using the methods of FIG. 4, which includes artifacts such as grating lobes, when the receiver array is sparse. The actual reflectivities, without artifacts, may be different from these calculated reflectivities. Since the grating lobes and other artifacts may depend on the frequency of the ultrasound, it may be possible to reduce or eliminate the artifacts by seeing how the calculated reflectivity varies for different frequency components of the ultrasound, for a given voxel.

[0090] The field of view is divided into voxels, small volume elements, to generate images, but the way the field of view is divided into voxels may be different for different frequency components, and optionally the comparison of calculated reflectivities is made, not for the same voxel for different frequency components, but the comparison is made for corresponding voxel reflectivities for the different frequency components. For example, the corresponding voxel reflectivity for one frequency component, for the reflectivity at a given voxel for a different frequency component, may be based on the reflectivity of the closest voxel, or it may be based on an interpolation of reflectivities for nearby voxels.

[0091] Optionally, a set of calculated voxel reflectivities is found, for example using the methods of FIG. 4, for each frequency component of the ultrasound, and this set of calculated voxel reflectivities may be used, for example, to generate a preliminary reflectivity image for each frequency component of the ultrasound. Information about how the calculated corresponding voxel reflectivities vary with frequency component may then be found by comparing the calculated corresponding voxel reflectivities in the different preliminary images. Alternatively, information about the calculated corresponding voxel reflectivity as a function of frequency component may be found without ever finding the calculated voxel reflectivities themselves for each frequency component. For example, a mean value of calculated corresponding voxel reflectivities may be found, for example, for each voxel, according to one way of dividing the field of view into voxels. Methods for doing this are also described below in the description of FIG. 4.

[0092] At 208, the information about the calculated corresponding voxel reflectivity as a function of frequency component is used to generate an output image with reduced image artifacts, forexample with reduced grating lobes. For example, for each pixel of the output image, a decision function sets the intensity of the output image to a value that depends on the distribution of the calculated corresponding voxel intensities. It should be noted that generally, herein, the term “voxel” is used to refer to an element of the field of view, and the term “pixel” is used to refer to an element of the output image. The same decision function is used, or substantially the same decision function is used, for all or most of the pixels of the output image, for example for at least 90% of the pixels, or at least 95% of the pixels, or at least 80% of the pixels. To the extent that the calculated corresponding voxel reflectivities vary for different frequency components primarily because of the grating lobes, which depend on frequency, such a procedure may tend to produce an output image with reduced grating lobe artifacts, for example with reducing grating lobe artifacts compared to an image that would be generated using only one of the frequency components. For example, the intensity is set equal to a mean value of the calculated corresponding voxel reflectivities, or the intensity is set equal to a same linear combination of the mean value and the standard deviation. For example, for each voxel of the output image, the intensity is set equal to the mean value minus the standard deviation, which may be a reasonably good approximation to the minimum value of the calculated corresponding voxel reflectivities for all the frequency components, or the intensity is set equal to the mean value plus the standard deviation, which may be a reasonably good approximation to the maximum value of the calculated corresponding voxel reflectivities for all the frequency components. Alternatively, the intensity for each voxel of the output image is set equal to the actual minimum value, or the intensity for each voxel of the output image is set equal to the actual maximum value, of the calculated corresponding voxel reflectivities for all frequency components. Alternatively, the intensity for each pixel in the output image is set equal to a particular percentile value of the calculated corresponding voxel reflectivities for all frequency components, for example the 25thpercentile or the 75thpercentile. But in order to do that, it may be necessary to find the complete set of calculated corresponding voxel reflectivities for all frequency components, rather than only finding the mean value and standard deviation. It should be noted that if only a small number of frequency components are used, for example if only two frequency components are used, then some of these different methods of finding the intensity of each voxel may be equivalent to each other.

[0093] There are various ways to define exactly what is meant by saying that the decision function is “substantially the same” for different pixels. An example of such a definition, which the inventor believes may be useful, is as follows. The decision function, for each pixel, is a function of the B-dimensional distribution space of calculated corresponding voxel intensities for the B different frequency bands, and its output is a single number, the output image intensity for that pixel. Toevaluate whether the decision function is “substantially the same” for different pixels, a coarsegrained evaluation of the decision function over the distribution space is made for each of the pixels. This coarse-grained evaluation of the decision function is made, for example, over a large region of the distribution space that covers typical values of calculated corresponding voxel intensities that are found when making ultrasound images of a particular modality of interest, for example imaging a particular FOV of a particular body part in different patients of a same gender, age, and body type, using a particular set of ultrasound imaging parameters. The large region of distribution space is divided into small regions, which are each, for example, less than 5%, or less than 10%, or less than 20%, or less than 30%, of the full range of the large region of distribution space in each of its B dimensions. Within each small region of the distribution space, for each of the pixels, the output of the decision function is found for each of many points well distributed in the small region, and a mean value of those outputs is found. A determination is then made of how much this mean value varies among the pixels being considered. For example, a standard deviation of the mean values is found, among the pixels being considered, for that small region of the distribution space. The decision function is considered to be “substantially the same” for those pixels, if the standard deviation among the pixels is small enough for each of the small regions of the distribution space, or on average for all the small regions of the distribution space. For example, the decision function is considered to be “substantially the same” if the standard deviation, for each of the small regions or averaged over all the small regions, is less than 5% or less than 10% or less than 20% or less than 30% of a typical range of values, or a typical standard deviation of values, of output pixel intensities found using this decision function for ultrasound images made with this modality. It should be noted that the decision function need not be a smooth function of position in the distribution space. But within each small region of the distribution space, the decision function should be evaluated at enough different points, sufficiently spread out, so that the standard deviation in the mean value, due to any lack of smoothness in the decision function, is small enough that it will not lead to a spurious variation in the mean values among different pixels that will make it appear that the decision function is not substantially the same for the different pixels.

[0094] Optionally, before using the information about calculated corresponding voxel reflectivities to generate an output image with reduced artifacts, the calculated voxel reflectivities are pre-processed, in a way that reduces effects other than grating lobes that may cause the calculated reflectivity to vary for different frequency components. For example, these other effects are reduced sufficiently that grating lobes are the dominant reason for the calculated reflectivity to vary for different frequency components. Other effects, that may cause the calculated reflectivity to vary for different frequency components, may include, for example, 1) noise in the image; 2)dependence of sensitivity of the receiver elements on ultrasound frequency; and 3) dependence of reflectivity of the tissue in the field of view on ultrasound frequency.

[0095] To reduce the effects of noise, the calculated voxel reflectivities may be denoised for each frequency component, for example, by averaging each voxel reflectivity with the reflectivities of neighboring voxels, for the same frequency component. Optionally, an assessment is made of the noise level of the reflectivity, before performing the denoising, and the denoising is only done to an extent necessary to reduce the noise level well below the amplitude of the grating lobes. The amplitude of the grating lobes may be estimated, for example, by modeling the sparsity of the receiver array and calculating the amplitude of the grating lobes it is expected to produce.

[0096] To reduce the effects of dependence of sensitivity of the receiver elements on ultrasound frequency, and the effects of dependence of tissue reflectivity on ultrasound frequency, the calculated voxel reflectivities are optionally normalized, for each frequency component. For example, the normalization is optionally based on the observed distribution of reflectivities over the field of view, for each frequency component. For example, the normalization optionally makes the mean reflectivity, over the field of view, the same for all frequency components, or optionally makes the minimum reflectivity over the field of view the same, or makes the maximum reflectivity over the field of view, or makes the range of reflectivity over the field of view the same, for all frequency components, or optionally makes the reflectivity at a particular percentile value over the field of view the same for all frequency components. Alternatively, the normalization is based on modeling how the sensitivity of the receiver elements depends on frequency, and / or on modeling how the reflectivity of the tissue in the field of view depends on frequency. Alternatively, a combination of these factors is used to determine the normalization. Optionally, the normalization is nonlinear.

[0097] At 210, information about the calculated voxel reflectivities for different frequency components, obtained from the recorded receiver signal data, is used to generate one or more additional images of the field of view. For example, the preliminary images are generated for the different frequency components, optionally after denoising and / or normalizing them. Optionally, the preliminary images are segmented before normalizing them, which may make the normalization more accurate. Optionally, normalizing the preliminary images may comprise using a nonlinear transformation. Optionally, even if the calculated voxel reflectivities were denoised and / or normalized before generating the output image with reduced grating lobes, optionally noise is added back into one or more of the additional images, for example in order to show the texture of the tissue. Optionally, the normalization is removed, or modified, for example in order to show how the tissue reflectivity varies with frequency. These additional images may not have the reducedgrating lobes that the main output image has, but may have other features that are missing from the main output image, that are still medically useful, for example for diagnosing a medical condition. It is potentially advantageous to make both the main output image and these additional images available to a radiologist.

[0098] At 212, the main output image with reduced grating lobes, and any of the additional images generated in 210, are outputted, for example as an image file that can be displayed on a monitor or printed. Optionally, two or more of these images, for example the main output image and one or more of the additional images, are combined to form a single image, for example by taking an average or a weighted average of them, for example a single image that has both reduced grating lobes and features of the one or more additional images that are medically useful.

[0099] Method for reconstructing each preliminary image from receiver data

[0100] Figure 3A shows a schematic perspective view 300, of an ultrasound transmitting transducer 104, an array of ultrasound receiving transducers 106, and a point in a field of view 102, to introduce the terminology used in describing how the receiver data is used to find a preliminary image of the field of view, in the description of FIG. 4. The transmitting transducer, which optionally is located in array 106, has the label “0”, while the receiver transducers are labeled 1, 2, 3,... up to a maximum label N. The index n is used to indicate the label of a receiver transducer. The image is defined as an ultrasound reflectivity R at each of a set of points in the field of view. For example, the set of points are grid points in a grid that fills the field of view, for example a Cartesian grid. The index i is used to indicate the label of grid point, with i going from 1 to M. The ultrasound amplitude at each receiver transducer n is modeled as the sum, over points i, of ultrasound waves transmitted from transducer 0, reflected from point i, and received by transducer n. The transmitter is assumed to emit a pulse f(t), for example a gaussian pulse centered at a frequency coo, which travels in a straight path 302, from transducer 0 to point i, at a constant sound speed c, and then, after reflecting from point i with a reflectivity Ri, travels in a straight path 304 at speed c to transducer n. The amplitude of the pulse is also assumed to be reduced by the inverse of the distance it travels from transducer 0 to point i, and by the inverse of the distance it travels from point i to transducer n, and by directivity functions Dir0,i and Diri,n that depend on the orientation of paths 302 and 304 relative to the orientation of the surfaces of the transducers, which are, for example, parallel to the plane of the array, as shown in FIG. 3A. It should be noted that the directivity functions Dir0,i and Diri,n are generally dependent on frequency, because, for example, the angular spread of the sensitivity of a transducer that receives ultrasound, and the angular spreadof the power of ultrasound transmitted by a transducer, will generally depend on how the lateral dimensions of the transducer compare to a wavelength of the ultrasound.

[0101] Figure 3B schematically shows a plot of an exemplary pulse shape for one of the ultrasound frequency components transmitted by the transducer shown in FIG. 1A and FIG. 3A, using the method of FIG. 2. In this example, the pulse has a gaussian envelope with a full width at half maximum of about 1.0 microseconds, with a peak frequency of about 2 MHz, or 4π × 106radians per second. The spectrum is gaussian with a full width at half maximum of about 106radians per second, or 0.16 MHz, about 8% of its peak frequency. Almost all of the energy of the pulse, more than 99%, is found within ±0.16 MHz of the peak frequency. If the frequency band for this pulse is taken to extend from 1.84 MHz to 2.16 MHz, a range that includes almost all of its energy, then there would be plenty of room for it to be well-separated from another frequency component with a similar pulse shape, for example with a peak frequency of 1 MHz or 3 MHz.

[0102] It should be noted that if each frequency component of the ultrasound has a spectrum that is strictly confined to its resonant frequency band, as stated, then its pulse shape f(t), which is the Fourier transform of its spectrum, will necessarily extend infinitely in time, and in principle it would take an infinite time to measure the signal of its echo. In practice, each transmitter only transmits each ultrasound component for a finite time, and each receiver element only generates a signal for a finite time, and any part of the resulting spectrum that is outside the resonant frequency band is assumed to have negligible power and is ignored. In the example given above, less than 1% of the energy of the pulse is outside the resonant frequency band, and more than 99% of the energy of the pulse is transmitted within a time of 2 microseconds.

[0103] Figure 4 is a flow chart 400 of further details of a method for finding reflectivities Ri for the grid points i in the field of view, from a set of receiver signal data Sn(t) for each receiver transducer n as a function of time t, when an ultrasound pulse f(t) is transmitted to the grid points i in the field of view. This method can be used, for example, to obtain the information about the calculated corresponding voxel reflectivity as a function of frequency component at 206 in FIG. 2.

[0104] It should be understood that a reflectivity Ri need not be found for every grid point i in the field of view. Optionally, reflectivities Ri are found only for some of the grid points i, optionally only for a small fraction of the grid points. However, if reflectivities are found for too few of the grid points, then the resulting images may not look very good, for example they may have low resolution or look noisy or not adequately cover the field of view.

[0105] At 402 the transmitted pulse is modeled, optionally by an analytic expression. Alternatively, the transmitted pulse is modeled numerically, using a table of values of the pulseamplitude as a function of time. For example, the pulse is modeled by a real-valued Gabor function of the form

[0106] _ f eaLt2cos(a)0t + <p); t < 0

[0107] JG\t) ) 4-2 Xx

[0108]

[0109] fe cos(m0t + <p); t > 0

[0110] with 0 representing the parameters (<zL, aR, m0, <p). Here, (<zL, aR) describes the bandwidth factor, m0= 2TT0is the center frequency, and cp the phase of the pulse. It is a non-symmetric Gaussian-weighted cosine function, with aLnot necessarily equal to aR. For example, the pulse shape shown in FIG. 3B is of this form, but is a symmetric Gaussian-weighted cosine function, with aLequal to aR. At 404 travel times are found for the pulse to travel from transmitting transducer 104 to point i in the field of view, and from point i to receiving transducer n, for each i and n. For example, the travel time of the pulse from transmitting transducer 104 to point i to receiver transducer n is given by

[0111] T" i -T0,i +Ti,n

[0112] where To iis the travel time from transmitting transducer 104 to point i, and Ti nis the travel time from point i to receiving transducer n. The pulse is assumed to travel in a straight line at a constant sound speed c, so To iis the length of path 302 divided by c, and Ti nis the length of path 304 divided by c.

[0113] At 406, a Green’s function, relating the transmitted pulse to the received signal, when the transmitted pulse reflects from point i, is found for each i and n, and it is given for example by fD ir0 _, _lDiri —,n

[0114] ui,n 2m rr.

[0115] c T Tj „

[0116]

[0117] 0,il’n

[0118] The amplitude is reduced by the inverse of the distance cTo,i from the transmitter to point i, by the inverse of the distance cTi,nfrom point i to receiver transducer n, and by the directivity function Diro.i which represents the dependence of transmission amplitude on the direction of transmission relative to the normal to the surface of the transmitter, and the directivity function Diri,nwhich represents the dependence of sensitivity of the receiving transducer on the direction of the received ultrasound relative to the normal to the surface of the receiving transducer. It should be noted that the Green’s functions Gi,nwill generally depend on frequency through the frequency dependence of the directivity functions Diro.i and D,n, even though the travel times To,i and Ti,nare independent of frequency in the approximation of straight line propagation of ultrasound waves at a constant speed c.In order to reconstruct an image of the field of view from a set of received signal data of each receiver transducer, the signal Sn(t) measured at the element n is modeled as a sum of reflections of the transmitted signale(t) by the various reflectors {R; i G {1,..., M}.

[0119]

[0120] The image will consist of a reflectivity Ri for each grid point i in the field of view. It should be noted that the actual reflection coefficient for a given point i may in general vary depending on the location of the receiver transducer n, as well as on the location of the transmitting transducer if there is more than one of them, depending on the angle of reflection, the angle of incidence, and the angle of orientation of any smooth surface that the ultrasound is reflecting from, for example the surface of a bone or an organ. But because it is difficult to model those effects so that they can be included in the Green’s function, the reflection coefficient Ri may be modeled as a constant for each grid point i, with the understanding that when the reflection coefficients are calculated from the received signals, the result for each grid point i will be an average of the reflection coefficients obtained by ultrasound received by the receiver transducers n. This is true for smooth surfaces, smooth on the scale of an ultrasound wavelength. However, a typical ultrasound image of body tissue also includes a large of number of what are effectively point reflectors, much smaller than an ultrasound wavelength, that produce speckles in the image that contribute to the texture of the image. For these speckles, the reflectivity is nearly isotropic, independent of the angle of incidence, so the calculated reflection coefficients will be accurate.

[0121] The reflection coefficients {R, i G {1,, M are optionally found for each grid point z by fitting the reflection coefficients {R, i G {1,

[0122]

[0123] to the measurements {Sn(t)}, n G {1,...,1V}. This is accomplished by minimizing a target function Q, where

[0124]

[0125] Or

[0126] Q = l Sn\t) - 2 ^Ri^ GiMt - Tni)Sn(t) + ^RiRj^ Gi nGj nfg(t - - T^)]

[0127]

[0128] In the delay and sum (DAS) approximation, which is optionally used,

[0129] J dtfg(t - = 8^ J cZt / 02(t)

[0130]

[0131] Within the framework of this approximation, the problem decouples into M independent problems:

[0132] M

[0133] Q = ^(Qo + Qi)

[0134]

[0135] 1=1

[0136] where

[0137] Q°=N jdt^

[0138]

[0139] n

[0140] and

[0141] Qt= -2R^ Gi,nf dtf (t - Tn;)Sn(t) + R2Gi:n2J dtf2(t)

[0142]

[0143] n 6 n

[0144] For a given point z, the estimated reflectivity (Ri) is given by

[0145] = argmin{Q } = argmin [ -2RtV Gi n| dtf (t - Tn^Sn(t) + R2\ Gi n2| dtf2(t}\ Ri Ri JQJ

[0146]

[0147] where 0 designates the set of pulse parameters (aL, aR, m0, <p). This equation can be written as = argmin£<4j7?j2—

[0148]

[0149] Ri

[0150] where

[0151] dt / 02(t)

[0152]

[0153] and

[0154] Bi ft - T i dtf (t)Sn(t + rnJ

[0155]

[0156] e 0

[0157] Then

[0158] Bi

[0159] (R^) = - —Bi2= -A^ / RffT) X)2

[0160]

[0161] Aj A,

[0162] and

[0163] Yn Glnf dt f (t)Sn(t + rnj)

[0164] <

[0165]

[0166] Ri(9)> =ZXTPoTw

[0167] This equation for the estimated reflectivity

[0168]

[0169] may be expressed in terms of the DAS beamformed signal

[0170] DAS(-t\ _ ^n=l GtnSn(t 4

[0171] f ' y / V p 2

[0172]

[0173] 2JTI=I 'Ji,n“filtered” by the pulse shape fe(t):

[0174] f dtfg(t')SiDAS(<t')

[0175]

[0176] f dtfg2(t)

[0177] and normalized by the energy of the pulse shape, J dtfg2(t). Alternatively, the estimated reflectivity is found from the Hilbert transform based envelope of the DAS beamformed signal SLDAS(t)

[0178] V

[0179]

[0180] £DAS(t) = |S£DAS(t) + j%(S£DAS(t))|

[0181] instead of S£DAS(t).

[0182] If there is more than one transmitting transducer, at different locations, then the sum over receiver elements n is optionally replaced by a sum over both receiver elements and transmitting transducers, both in the numerator and in the denominator of the expression above for the DAS beam-formed signal S£D74S(t). Note that the Green’s function Gi nand the travel time rn£differ for different transmitting transducers at different locations, for a given z and n, because they depend on the distance from the transmitting transducer to the grid point z. Alternatively, the ultrasound from the different transmitting transducers may be treated as different components of ultrasound, and their reflectivities may be calculated separately. At 408, the DAS beamformed signal S£Dj4S(t) is found for each grid point i in the field of view, from the signal data Sn(t) for each receiver transducer n, and from the Green’s functions Gi,n. In general, because, as noted above, the Green’s functions Gi,ndepend on frequency, the DAS beamformed signal SLDAS(t) will also depend on frequency, and is calculated separately for each ultrasound component. Optionally, the directivity functions, and hence the Green’s functions, are assumed not to vary very much over the width of the frequency band of each component, and, for example, the directivity functions are evaluated at a peak frequency or a central frequency of that component. This may be a good approximation for finding the DAS beamformed signal if the width of the frequency band is relatively small compared to the peak or central frequency. Alternatively, for example if the transmitted ultrasound consists of short pulses of only one or two cycles, which have a relatively large bandwidth compared to the peak or central frequency, then the directivity functions may be evaluated for those pulse shapes, for example numerically using finite element methods, or by measuring the angular dependence of receiver sensitivity and / or transmitted power.

[0183] At 410, the Hilbert transform based envelope ViDASt) of the DAS beamformed signal is optionally found. At 412, either S£D74S(t) or F£D / 1,s(t) is used to find the estimated reflectivity image

[0184] At 414, the coherence factor COFi is optionally found, for each grid point i. The coherence factor gives more weight to large coherent structures that dominate an image, and less weight toclutter. It can be used to reduce side lobes, grating lobes and clutter due to the ribs and the lungs in an ultrasound image, especially when the receiver array is sparse, not satisfying the Nyquist criterion. See, for example, the papers by Alomari, and by Long, Trahey and Bottenus, cited above, for more information on how the coherence factor is found and used.

[0185] If there is a single strong reflector at a point i that dominates the other reflectors in the field of view, then the time delayed signal Sn(t + Tnj) for that z, for each receiver transducer n, will be fully correlated with fe(t), meaning that

[0186] f dtf (t)sn(t + Tnt)

[0187] 3= +1

[0188] f dtf2(t) x f dtSn2(t + rnJ

[0189]

[0190] In general, whether or not there is a single strong reflector, we define, for each transducer n and point i,

[0191] f dtf (t)Sn(t + rf)

[0192] I Corr f l =3- f dtfg2(i) x f dtSn2(t + Tnf)

[0193]

[0194] Then in general

[0195] |Corrj”| < 1

[0196] with | Corral = 1 if z is the only reflector. The correlation factor for each point is defined as

[0197]

[0198] which is always between 0 and 1. Points z that are part of localized coherent structures that have large reflectivity will have higher COFi, while points that only contribute noise and clutter will have low COFi.

[0199] At 416, if the coherence factor has been calculated, the estimated reflectivity

[0200]

[0201] at each grid point i is optionally weighted by the coherence factor COFi, for example it is multiplied by the coherence factor. This tends to emphasize the coherent structures in the image, and reduces the noise and clutter, including side lobes and grating lobes. For the reduced grating lobe image shown in FIG. 9, the coherence factor has been included, and contributes to the reduction in grating lobe artifacts.

[0202] In some embodiments of the invention, instead of or in addition to using the method of 412 to find the reflectivity

[0203]

[0204] for each frequency component of the transmitted ultrasound, the method of 412 is used to find an average reflectivity Ri for two or more frequency components of the transmitted ultrasound. For example, the method of 412 is used to find an average voxel reflectivity of all the frequency components. This is done, for example by replacing the pulse shapefe(t) for one frequency component, used to find the filtered beam- formed signal at 412, with a pulse shape that is equal to the sum of the pulse shapes for all the frequency components for which the reflectivity is being averaged. If the average reflectivity is found in this way for all the frequency components, then the resulting average reflectivity can be used at 208 of FIG. 2 to find the output image with reduced grating lobes, rather than first finding the reflectivities for each frequency component using the method of 412, and then averaging corresponding voxel reflectivities to find the output image. In this case, the same set of grid points z and corresponding voxels is used for all frequency components. However, this method may be most accurate if the DAS beam-formed signal is to a good approximation independent of frequency, over the range of frequency components used. That may be true, for example, if there is relatively small range of frequencies used for the different components. It may also be true, for example, if the ultrasound wavelength is much greater than the lateral dimensions of the transducers, for the entire range of frequencies used, in which case the directivity function may be nearly isotropic for all frequencies used. But in many situations of interest, the DAS beam-formed signal may vary a lot with frequency, and this method of finding the average reflectivity over the different components may not give accurate results.

[0205] A potential advantage of finding the average reflectivity directly using the method of 412, without ever finding the reflectivities for each frequency component, is that it may save computation time and memory. Potential advantages of first finding the reflectivities for each frequency component using the method of 412, and then finding the average of corresponding voxel reflectivities to find the output image with reduced grating lobes at 208, is that the voxel reflectivities for each frequency component may be pre-processed, for example by normalizing them and / or denoising them, as described above for 208 in FIG. 2, and that other information about the distribution of corresponding voxel reflectivities for different frequency components may also be used in finding the output image, for example the variance of the reflectivities, which is a nonlinear function of the reflectivities of the different frequency components.

[0206] A similar method may be used to find an average of corresponding voxel reflectivities for two or more frequency components, not necessarily all the frequency components, using the method of 412, by adding up the pulse shapes fe(t) for the components that are being averaged, when filtering the beam-formed signal, to directly generate additional images at 210 that are based on the average reflectivity at two or more frequency components. This option has the same potential advantages and disadvantages as using the method of 412 to directly find an average reflectivity of all the frequency components. If the method of 412 is used to directly find an average voxel reflectivity at two or more frequency components, or at all the frequency components, then it ispotentially advantageous to select a ratio of transmitted powers for the different frequency components that will produce a weighted average of reflectivities that will have desired properties, for example a weighted average that will have optimized properties or almost optimized properties, for example a substantially reduced amplitude of grating lobes, or an image with features that are useful for diagnosing or evaluating a particular medical condition. This can be done, for example, by trying different ratios of powers at different frequency components, and seeing what works best for certain types of images. But, as noted above, the results may be inaccurate if the DAS beam-formed signal varies a lot with frequency over the range of the different frequency components, and if that is ignored.

[0207] Origin of grating lobes

[0208] When an ultrasound receiver array has adjacent receiver elements that are closer together, along both axes of the array, than half a wavelength of the ultrasound being used, then it is possible to tell, without ambiguity, from which direction an ultrasound wave is impinging of the array, from the phase difference seen at adjacent receivers. But if the array is sparse, with adjacent receiver elements further apart than half a wavelength, and especially if adjacent receiver elements are further apart than one wavelength, then it is possible for ultrasound waves coming from two different directions to produce the same phase difference between adjacent receiver elements. In this case, it is not always possible to tell, from the receiver signals, where an ultrasound wave is coming from. This situation leads to grating lobe artifacts in an image made by such a sparse array.

[0209] Further details on grating lobes may be found in the pdf file “Grating Lobes and Side Lobes, Olympus IMS,” downloaded from <https(colon)(slash)(slash)ims(dot)evidentscientific(dot)com(slash)en(slash)learn(slash)ndt-tutorials(slash)transducers(slash)lobes>, and in the Wikipedia article on “Grating Lobes,” downloaded from <https(colon)(slash)(slash)en(dot)Wikipedia(dot)org(slash)wiki(slash)Grating_lobes>, both downloaded on January 26, 2025. See also the material on grating lobes in Szabo, T. L., Diagnostic Ultrasound Imaging: Inside Out (2nd ed.), Academic Press, 2013. ISBN: 9780123964878.

[0210] Simulation result

[0211] Simulated images were calculated for a sparse array of the transducers, not satisfying the Nyquist criterion, to illustrate the method of FIGS. 2 and 4 for generating reduced grating lobe images. Figure 5 shows a rectangular array 500 of receiving transducers, with one transmitting transducer in the center. There are rows 502 of receiving transducers near the top of the array, androws 504 of receiving transducers near the bottom of the array. In a stripe 506 in the middle of the array, there are no receiving transducers. Array 500 could be used to model an array of receivers that is located outside a subject’s chest, viewing a field of view inside the rib cage, with part of the array, stripe 506, blocked by a rib and not receiving any reflected ultrasound from the field of view. To be consistent with that model, the transmitting transducer should be located outside stripe 506, since transmitted ultrasound would also be blocked by the rib, but it is expected that locating the transmitting transducer near the edge of stripe 506, just above or below the rib according to this model, will not make much difference in the appearance of the images simulated in FIGS. 7, 8, and 9.

[0212] In the simulated images, the transducers have a high frequency normal mode at 1.08 MHz, and a low frequency normal mode at 0.54 MHz. The width of each frequency band is somewhat smaller than the difference between them, so the two bands are well separated. Assuming an ultrasound speed in body tissue of 1.54 x 103m / s, the wavelength for the high frequency component would be about 1.4 mm, and the wavelength for the low frequency component would be about 2.8 mm. The transducers in array 500 are 4 mm apart in the x-direction, but across stripe 506 there is a gap of 20 mm between transducers in the y-direction. So the Nyquist criterion is marginally satisfied for the low frequency component in the x-direction, but far from satisfied for the high frequency component in the x-direction, and very far from satisfied for either the high frequency or the low frequency component in the y-direction.

[0213] Figure 6 shows a two-dimensional field of view 600 extending between +35 mm and -35 mm in the x and y directions, in a plane with z = 40 mm, parallel to the plane of the array and 40 mm from it. There is a single reflector, effectively a point reflector, in the center at x = 0 and y = 0.

[0214] Figure 7 was generated by simulating receiver signals from each of the receiver transducers in array 500, for ultrasound pulses at the high frequency transmitted from the transmitting transducer in the center of the array, reflecting from the point reflector shown in FIG. 6, and received by each of the receiving transducers. The simulated receiver signals were then processed according to the method of FIG. 2 and FIG. 4, to produce an image of the field of view. As expected, because the high frequency ultrasound fails to satisfy the Nyquist criterion in both x and y, FIG. 7 exhibits prominent grating lobes around the reflector at the center, both in the x direction and in the y direction.

[0215] Figure 8 was generated in the same way using the low frequency ultrasound. Because the low frequency ultrasound, with wavelength 2.8 mm, marginally satisfies the Nyquist criterion in the x-direction, with transducers separated by 4 mm, there are only a relatively weak grating lobesseen in FIG. 8 in the x direction. But in the y direction, where there is a gap of 20 mm between the rows of transducers in the middle of the array, the low frequency ultrasound also is far from satisfying the Nyquist criterion, and there are strong grating lobes in the y direction.

[0216] Figure 9 shows a reduced grating lobe image, generated by taking the minimum intensity of FIG. 7, the high frequency image, and FIG. 8, the low frequency image, for each voxel in the field of view. Because the grating lobes in FIGS. 7 and 8 tend to occur at different locations, taking the minimum intensity from both images tends to reduce the intensity of the grating lobes. A similar result could be obtained by taking the average intensity of both images, or the maximum intensity of both images, or any linear combination of the minimum, the maximum and the average.

[0217] Evaluating effectiveness of reducing grating lobes in an image

[0218] The effectiveness of the methods described herein, for reducing grating lobe artifacts, may be estimated by comparing the amplitude of grating lobes in the reduced grating lobe output image with the amplitude of grating lobes in an image made from only one of the frequency components. It may also be estimated by comparing the amplitude of grating lobes in the reduced grating lobe output image, with the amplitude of grating lobes in an image of the same field of view made with an ultrasound receiver array that is not sparse, for which the Nyquist criterion is well satisfied.

[0219] To estimate the amplitude of grating lobes in an image made with a single frequency component, an estimate may be made first of what the field of view would look like without grating lobes, for example by using an image of the field of view made with a non- sparse array of ultrasound receivers for which the Nyquist criterion is well satisfied. Alternatively, an image of the field of view without grating lobes may be obtained from an anatomical atlas. Using that image for the “ground truth” for the field of view, a simulation can then be made of an ultrasound image of that field of view that would be generated for that frequency component of ultrasound using the sparse array, for which the effectiveness of the method of reducing grating lobes is being evaluated. The amplitude of grating lobes in the image made from one frequency component may then be estimated by taking a pixel by pixel rms difference in intensities between the simulated image made using the sparse array, and the “ground truth” image of the field of view. Optionally, one or both these images are denoised before finding the rms difference, for example by averaging image intensity over a small neighborhood of each pixel. If images are not denoised first, the pixel by pixel rms difference in intensities may be dominated by noise, rather than by the grating lobes. It should be noted that the amplitude of grating lobes in a realistic ultrasound medical image will generally be smaller than the amplitude of grating lobes seen in an image of a single point reflector, such as the images shown in FIGS. 7, 8 and 9. But for evaluating the effectiveness of the methodfor reducing grating lobes, it may be more relevant to look at a realistic medical image than at an image of a single point reflector.

[0220] A similar method may be used to estimate the amplitude of grating lobes in the reduced grating lobe output image generated using the methods described herein. But for that purpose, it may be more important that the “ground truth” image of the field of view be almost completely free of grating lobe artifacts.

[0221] Using these techniques of estimating the amplitude of grating lobe artifacts in an output image generated using the methods described herein, the parameters used for the method of reducing grating lobe artifacts, for example the number of frequency components, and their peak frequencies and bandwidths, may then be chosen to reduce the grating lobe artifacts, or the parameters may be optimized to minimize the amplitude of grating lobes artifacts.

[0222] Once that has been done, a satisfactory set of parameters may be chosen, for use in a given type of ultrasound imaging. For example, the parameters can be chosen so that the grating lobe artifacts in the output image are reduced by at least a factor of 2, or at least a factor of 3, or at least a factor of 5, or at least a factor of 10, from the images made with a single frequency component.

[0223] To characterize the effectiveness of a particular ultrasound system at reducing grating lobe artifacts in general, rather than the effectiveness at reducing grating lobe artifacts for a particular field of view, a standard field of view may be chosen, simulated receiver signals are found for the standard field of view, the system is used to find an output image using the simulated receiver signals, and the effectiveness of the system at reducing the grating lobe artifacts for the standard field of view may be evaluated, using the methods described above. For example, the standard field of view is a single point reflector at the center of the field of view, with the field of view big enough that almost all of the grating lobe artifacts in the image fall within the field of view. Alternatively, the standard field of view could be a field of view from a medical atlas, characteristic of the kind of ultrasound imaging that the system is going to be used for, or several different fields of view from medical atlases may be used, covering a range of different types of ultrasound imaging, and the effectiveness of the system at reducing grating lobe artifacts can be evaluated separately for each type of ultrasound imaging.

[0224] Use of ultrasound transmitting transducers

[0225] One option for producing the two different frequency ranges of ultrasound used to reconstruct the two preliminary images in the method of FIG. 2, is to use two different normal mode resonant frequencies of the same ultrasound transmitting transducer. The frequency range for each normal mode is typically centered at that normal mode resonant frequency, and extendsover a range corresponding to the bandwidth of the pulse shape that is used. This is optionally done with a single transmitting transducer, as shown in FIG. 1A. Alternatively, two or more transmitting transducers of the same design are used, optionally simultaneously and located close together, with the same two normal mode resonant frequencies, which may allow a higher transmitted ultrasound power to be used than with a single transmitter. With a broad phased array of ultrasound transmitters, the ultrasound power can be concentrated over a narrower field of view than with a single small ultrasound transmitter.

[0226] Alternatively, the two frequency ranges are produced by two different transmitting transducers, or two different sets of identical transmitting transducers, for example with different designs and different normal mode frequencies. One of the preliminary images is made using one of the transducers to transmit ultrasound with a first frequency range around a normal mode resonant frequency of that transducer, and the other preliminary image is made using the other transducer to transmit ultrasound with a second frequency range around a normal mode resonant frequency of that transducer, different from and optionally not overlapping with the first frequency range. A potential advantage of using two normal modes of the same transducer to produce the two frequency ranges of ultrasound is that all of the transmitting transducers can be used to produce the ultrasound for each frequency range, simplifying the fabrication of the transducers and possibly transmitting a higher ultrasound power. A potential advantage of using two different transducers, of different design to transmit the ultrasound in the two frequency ranges, is that each transducer can be optimized for transmitting one frequency range.

[0227] It should be understood that in some embodiments of the invention three or more frequency ranges of ultrasound are used to produce three or more preliminary images, which are then used to produce the reduced grating lobe image, and in this case the different frequency ranges can be centered around different normal mode frequencies of the same transducer, or of different transducers, or of any combination of the same and different transducers.

[0228] Optionally, the ultrasound transducers, including the receiving and / or transmitting transducers, are rectangular lateral mode transducers that have different dimensions in the x and y directions, corresponding to the axes of the rectangle, and have different lowest lateral mode resonant frequencies along its x and y axes. These two normal mode frequencies can be used for the two frequency ranges used to produce the two preliminary images, using the same transducer for both frequency ranges. A potential advantage of using the lowest normal mode, along each axis, to produce the ultrasound in the two frequency ranges, is that the lowest normal modes may produce ultrasound with higher efficiency that the higher normal modes. If more than two frequency components are being used, then optionally higher normal mode resonance frequencies are usedalong one or both of the x and y axes of the rectangular lateral mode transducer, and / or additional transmitting transducers are used, for example lateral mode transducers of different design with different normal mode resonant frequencies. For example, the additional transducers are lateral mode transducers with different lateral dimensions and / or made of different materials than the transducer used for the first two frequency components, so they have different lowest normal mode resonant frequencies.

[0229] Design and method of manufacture of the ultrasound transducers

[0230] The following exemplary procedure is believed by the inventor to be suitable for fabricating an array of ultrasound transducers that could serve as array 106 in system 100 of FIG.

[0231] 1A. The piezoelectric elements (PZT-5H, CTS 3203) have CTS-standard NiCr or Au electrodes applied by sputtering. The piezoelectric elements are attached to copper pads on a PCB substrate using silver epoxy (8331-14G, made by M. G. Chemicals in Surrey, B. C., Canada). The properties of PZT-5H and other types are PZT, including their density, elastic modulus, dielectric constant, piezoelectric coupling constants, and typical bias electric field and bias stress, are given, for example, by the html document “Material Properties — Piezo Support,” available at <http(colon)(slash)(slash)support(dot)piezo(dot)com(slash)article(slash)62-material-properties>, downloaded on January 27, 2025. The piezoelectric elements are, for example, rectangular slabs, 3 mm long, 2 mm in breadth, and 0.3 mm thick, and their lowest lateral normal modes are, for example, 0.47 MHz and 0.74 MHz.

[0232] To minimize cross-talk between elements through the substrate, only the central 25% of each element is bonded. Small stud bumps, around 100 pm in size, are glued to the substrate to support the element and maintain parallelism, reducing lateral friction between the element and substrate. Wire bonding is used to connect the top (signal) side of the element to the ground (copper) pad on the substrate. Some details on how wire bonding may be done are given, for example, by the html file “Wire Bonding — Advanced PCB Design Blog — Cadence,” available at <http(colon)(slash)(slash)resources(dot)pcb(dot)cadence(dot)com(slash)blog(slash)2023-wire-bonding>, downloaded on January 28, 2025.

[0233] An 0.8 mm layer of epoxy (EPO-TEK 353ND from Epoxy Technology) was applied to the elements, serving as a matching layer between the PZT substrate and a silicone layer. A thin (1 mil) layer of silicone is then glued to the elements in order to seal the entire patch.

[0234] When a transducer transmits ultrasound waves, a power supply applies an AC voltage between the two electrodes, together with a DC bias electric field. With the thickness of PZT slab uniform over its area, the electrodes will produce a uniform electric field in the slab, in the z-direction. The piezoelectric coefficient, expressed as a rank-2 tensor d, gives the piezoelectric strain of the slab as a linear function of the electric field. Optionally, the AC electric field is small enough, for example no more than 50% or 30% or 20% of the DC bias electric field, so that the piezoelectric strain is well approximated as a linear function of the AC electric field, for example with nonlinear components of the piezoelectric strain no more than 20% or 10% or 5% or 3% of the linear component. Having small nonlinear components of the piezoelectric strain has the potential advantage that the piezoelectric strain, and the transmitted ultrasound waves, will be mostly at the frequency of the applied AC electric field, with little power going into ultrasound waves at harmonics of the frequency of the applied AC electric field. Since the electric field is in the z-direction, the d₃₃ component of the piezoelectric coefficient gives the piezoelectric strain in the z-direction, and the d₃₁ and d₃₂ components, which are generally equal to each other for PZT and similar piezoelectric materials, give the piezoelectric strains in the x and y directions. Optionally, the housing of the transducer produces a bias compressive stress in the slab in the z direction, for example with a value that produces a reasonably large d₃₃ and d₃₁, for example not too much less than the largest d₃₃ and d₃₁ possible, when the bias DC electric field and the bias compressive stress are optimized. The upper face of the slab is mechanically coupled to the body of the patient, and the time-varying strain of the slab in the z-direction, depending on d₃₃, launches ultrasound waves into the field of view in the body, propagating in the z-direction and spreading out spherically at large distances compared to the lateral dimensions of the slab. But the normal mode resonant frequencies of the slab depend mostly on d₃₁. The normal vibrational modes of a slab of piezoelectric material depend on the AC piezoelectric coefficient tensor d, the AC dielectric tensor s at zero AC stress, and the AC elasticity tensor s at zero field. Giurgiutiu and Zagrai, cited above, derive equations for the normal modes of a thin rectangular slab of piezoelectric material, with a uniform AC electric field and DC bias electric field applied in the thin direction, and with different lateral boundary conditions, for example free or clamped or constrained. The lowest normal mode frequencies are for the lateral modes, where, if the lateral dimensions of the slab are much greater than its thickness, most of the kinetic energy of the vibration is from motion in the lateral directions, and most of the elastic energy is from stress in the lateral directions. The AC electric field in the z-direction drives these motions and stresses in the x and y directions due to ds i - It may help to drive these lateral modes efficiently, if the AC electric field is applied at a frequency close to the normal mode resonant frequency for the lateral mode that is being driven.

[0235] Giurgiutiu and Zagrai give expressions for the lateral normal mode resonant frequencies, in the linear regime, of a thin rectangular slab excited by a uniform AC electric field in the thindirection, with a » b » t, for modes varying in the y-direction (along the length a of the slab), and for modes varying in the x-direction (along the breadth b of the slab), with free boundary conditions, with clamped boundary conditions, and with constrained boundary conditions. For the free boundary case, which is the case of primary interest for the ultrasound receiver array of FIG.

[0236] 1A, resonant normal mode frequencies occur at frequencies co = (2m+l)7tcs / a, where m = 0, 1, 2,..., for modes in the y-direction, and csis the sound speed in the piezoelectric material of the slab. If the material has an anisotropic sound speed, then csis the sound speed in the y-direction. For modes in the x-direction, the resonant normal mode frequencies are co = (2m+l )ncjb, where m = 0, 1, 2,..., and csis the sound speed in the x-direction if the sound speed is anisotropic. These expressions are valid for lateral modes, with frequencies much less than 7tcs / t. These results are given by Giurgiutiu and Zagrai on page 961 in the right hand column, following Eq. (8), for modes in the y-direction, and on page 964 in the right hand column for modes in the x-direction. In the absence of damping, for a finite electric field applied across the thickness of the slab at a normal mode resonant frequency, and ignoring nonlinear effects, the amplitude of the excited mode would be infinite. But with a small amount of damping, the amplitude at or near the normal mode resonant frequency would be large but finite, as explained by Giurgiutiu on page 962, right hand column.

[0237] If the length a is comparable to the breadth b, even though it is not much greater, and if a and b are both much greater than the thickness t of the slab, then we expect these expressions for the normal mode resonant frequencies to be approximately correct. To find the normal mode frequencies more precisely in this case, a two-dimensional partial differential equation could be solved, in the plane of the slab, with appropriate boundary conditions, for example free boundaries, at the edges of the slab. In general, there will also be lateral normal modes that have amplitude that varies significantly both in x and y. Alternatively, a few of the lowest normal mode frequencies can be measured experimentally, for a transducer made of a given piezoelectric material with a given length a, breadth b and thickness / . For the transducers simulated in FIGS. 7 and 8, with a = 3 mm, b = 2 mm, and / = 0.3 mm, the lowest lateral modes, respectively in the length and breadth directions, have frequencies 0.47 MHz and 0.74 MHz.

[0238] Details of controller

[0239] Figure 10 shows a block diagram 1000 of an exemplary configuration for controller 108 in FIG. 1A, that controls the transmission of ultrasound waves into the field of view in the subject’s body, receives signals of reflected ultrasound waves recorded by the receiver transducers, and uses the receiver signals to reconstruct an image of the field of view, including an image with reduced grating lobe artifacts.Transducer array 106 records signals of the reflected ultrasound waves each receiver transducer receives. These signals are optionally passed through an analog front end (AFE) 1002 to the field programmable gate array (FPGA) 1004. The AFE corrects the receiver signals for distortions caused by the ultrasound passing through body tissue. The FPGA is, for example, a Digital Beamformer Arria V FPGA, which is a family of FPGAs developed by Altera, now owned by Intel. The FPGA includes a receiving beamformer which combines the receiver signals into beamformed signals, for example the DAS beamformed signals as described in FIG. 4. Optionally, if there is more than one transmitting transducer, for example if there is a phased array of transmitting transducers, there is also a transmitting beamformer that assigns a transmitting waveform to each of the transmitting transducers, for example in order to produce a coherent beam of transmitted ultrasound that is focused on the volume of the field of view. A back-end processing unit optionally performs calculations needed to do the beamforming. A real time clock (RTC), which both the transmitting beamformer and the receiving beamformer have access to, ensures that the beamforming of the received signals is done with accurate values for the time delays between the transmitted waves and the received signals, as described in FIG. 4. A control block passes signal data, such as the beamformed signal data, from the receiving beamformer and an FPGA local memory, which processes the beamformed signals to obtain reflectivity images of the field of view, as described in FIG. 4. The FPGA local memory optionally stores the generated image data in an external LP DDR3 memory unit 1008, which may output images. A user optionally controls the processing of the receiver data, and the waveforms and timing of the transmitted ultrasound, through a control block in digital beamformer 1004, for example using an Android tablet 1006, or another suitable computer.

[0240] It is expected that during the life of a patent maturing from this application many relevant ultrasound transducers and transducer arrays will be developed and the scope of the terms ultrasound transducer and transducer array are intended to include all such new technologies a priori.

[0241] As used herein the term “about” refers to ± 10 %.

[0242] The terms "comprises", "comprising", "includes", "including", “having” and their conjugates mean "including but not limited to".

[0243] The term “consisting of’ means “including and limited to”.

[0244] The term "consisting essentially of" means that the composition, method or structure may include additional ingredients, steps and / or parts, but only if the additional ingredients, steps and / or parts do not materially alter the basic and novel characteristics of the claimed composition, method or structure.As used herein, the singular form "a", "an" and "the" include plural references unless the context clearly dictates otherwise. For example, the term "a compound" or "at least one compound" may include a plurality of compounds, including mixtures thereof.

[0245] Throughout this application, various embodiments of this invention may be presented in a range format. It should be understood that the description in range format is merely for convenience and brevity and should not be construed as an inflexible limitation on the scope of the invention. Accordingly, the description of a range should be considered to have specifically disclosed all the possible subranges as well as individual numerical values within that range. For example, description of a range such as from 1 to 6 should be considered to have specifically disclosed subranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6 etc., as well as individual numbers within that range, for example, 1, 2, 3, 4, 5, and 6. This applies regardless of the breadth of the range.

[0246] Whenever a numerical range is indicated herein, it is meant to include any cited numeral (fractional or integral) within the indicated range. The phrases “ranging / ranges between” a first indicate number and a second indicate number and “ranging / ranges from” a first indicate number “to” a second indicate number are used herein interchangeably and are meant to include the first and second indicated numbers and all the fractional and integral numerals therebetween.

[0247] It is appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention, which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable subcombination or as suitable in any other described embodiment of the invention. Certain features described in the context of various embodiments are not to be considered essential features of those embodiments, unless the embodiment is inoperative without those elements.

[0248] Although the invention has been described in conjunction with specific embodiments thereof, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications and variations that fall within the spirit and broad scope of the appended claims.

[0249] It is the intent of the Applicant(s) that all publications, patents and patent applications referred to in this specification are to be incorporated in their entirety by reference into the specification, as if each individual publication, patent or patent application was specifically and individually noted when referenced that it is to be incorporated herein by reference. In addition, citation or identification of any reference in this application shall not be construed as an admission that such reference is available as prior art to the present invention. To the extent that sectionheadings are used, they should not be construed as necessarily limiting. In addition, any priority document(s) of this application is / are hereby incorporated herein by reference in its / their entirety.

Claims

WHAT IS CLAIMED IS:

1. A medical ultrasound system that, when it is in good acoustic contact with a body surface, generates an ultrasound output image of a field of view of tissue inside the body, with reduced grating lobe artifacts, the system comprising:a) an ultrasound transmitting module comprising one or more ultrasound transmitting transducers that, when the system is in good acoustic contact with the body surface, transmits into the field of view two or more different frequency components of ultrasound waves, each frequency component limited to a different one of well separated resonant frequency bands, each of the resonant frequency bands including a different normal mode frequency, associated with that frequency component, of the one or more ultrasound transmitting transducers;b) an array of ultrasound receiver elements, at least some of them spaced further apart from their neighbors than half a wavelength for at least one frequency of the resonant frequency bands at a sound speed of 1540 meters per second, that, when the system is in good acoustic contact with the body surface, receive the frequency components of ultrasound waves after they reflect from tissue in the field of view, the array generating signals of the ultrasound received by each receiver element; andc) a controller configured to 1) control the transmitting module to transmit the frequency components of ultrasound waves, 2) record a signal of the ultrasound received by each of the ultrasound receiver elements, 3) determine, from the generated signals of the receivers elements, information about reflectivity and how it differs for different frequency components for given locations in the field of view, and 4) reconstruct the output image, with pixel intensities in the output image depending on the information about differing calculated corresponding voxel reflectivities for the different frequency components.

2. A system according to claim 1, wherein at least one of the one or more ultrasound transmitting transducers is a lateral mode transducer.

3. A system according to claim 2, wherein the lateral mode transducer comprises a thin slab of piezoelectric material, and electrodes that cover opposite faces of the slab that produce an electric field in the slab, uniform within 20%, in a thickness direction of the slab, that causes the transducer to generate ultrasound waves when the electric field has a frequency equal to a lateral mode resonant frequency of the slab, wherein the slab has an anisotropic shape in a plane perpendicular to the thickness direction.

4. A system according to claim 3, wherein the shape in the plane perpendicular to the thickness direction is not symmetric under 90-degree rotation in the plane.

5. A system according to claim 4, wherein the slab has a non-square rectangular shape in the plane perpendicular to the thickness direction.

6. A system according to any of claims 2 to 5, wherein two frequency components have frequencies at different lateral resonant modes of one transducer of the one or more transducers.

7. A system according to claim 5, wherein a ratio of the frequencies of the two frequency components differ by more than 3% from 1, 4 / 3, 3 / 2, 5 / 3, 2, 5 / 2, and 3.

8. A system according to any of the preceding claims, wherein the ultrasound transmitting module comprises a plurality of ultrasound transmitting transducers, having between them normal mode frequencies in each of the resonant frequency bands, that transmit ultrasound waves in those resonant frequency bands into the field of view when the system is in good acoustic contact with the body surface, at same times or at different times for different ultrasound transmitting transducers, the ultrasound waves reflecting from the tissue in the field of view, and contributing to the ultrasound waves received by the array of ultrasound receiver elements, and to the signals of received ultrasound waves generated by the array.

9. A system according to any of the preceding claims, wherein the controller is configured to control the transmitter to transmit the different frequency components at different non-overlapping times.

10. A system according to any of claims 1 to 8, wherein the controller is configured to control the transmitter to transmit the different frequency components at times that at least partially overlap, and the array, the controller or both are configured to generate and record signals for different frequency components by using hardware or software or both to separate different frequency components in the signals generated by the array of ultrasound receiver elements.

11. A system according to any of the preceding claims, wherein at least one of the frequency components transmitted by the one or more ultrasound transmitting transducers is associated with a thickness mode resonant frequency of one of the one or more ultrasound transmitting transducers.

12. A system according to any of claims 1 to 10, wherein all of the frequency components transmitted by the ultrasound transmitting module are associated with lateral mode resonant frequencies of the one or more ultrasound transmitting transducers.

13. A method of generating an ultrasound output image of a field of view of tissue inside a body, using a sparse array of ultrasound receiver elements, the method comprising: a) transmitting a plurality of ultrasound wave components into the field of view, from one or more transmitting locations for each component, with each component limited to a different one of well separated resonant frequency bands;b) receiving reflections of the transmitted ultrasound components from the field of view by the sparse array of ultrasound receiver elements, and recording signals of the ultrasound received by each receiver element as a function of time;c) finding from the recorded receiver signals information about corresponding calculated voxel reflectivity as a function of transmitted ultrasound component; andd) assigning an intensity to each pixel of the output image, based on the information about the corresponding calculated voxel reflectivity as a function of ultrasound component.

14. A method according to claim 13, wherein assigning an intensity to each pixel in the output image comprises assigning intensities for which an estimated amplitude of grating lobes in the output image is less than half an estimated amplitude of grating lobes in images made from voxel reflectivities for one ultrasound component.

15. A method according to claim 13 or 14, wherein transmitting the ultrasound wave components is done by one or more ultrasound transmitting transducers, and each of the resonant frequency bands corresponds to a different normal mode frequency of the one or more ultrasound transmitting transducers, corresponding to different normal modes of a same transmitting transducer, or to different normal modes of different transmitting transducers, or to both.

16. A method according to claim 15, wherein at least one of the normal mode frequencies is a lateral mode of a lateral mode ultrasound transducer.

17. A method according to claim 15 or 16, wherein the one or more ultrasound transmitting transducers comprise at least two transmitting transducers of different designs with different normal modes, and at least two of the resonant frequency bands each correspond to a normal mode of a different one of the two transmitting transducers with different normal modes.

18. A method according to any of claims 15-17, wherein at least some of the resonant frequency bands each correspond to a different normal mode of one of the transmitting transducers.

19. A method according to any of claims 13-18, wherein transmitting the ultrasound wave components comprises transmitting different components at different time periods that do not overlap.

20. A method according to any of claims 13-18, wherein transmitting the ultrasound wave components comprises transmitting different components at time periods that partly or completely overlap.

21. A method according to any of claims 13-20, wherein finding information about different corresponding voxel reflectivities as a function of transmitted ultrasound component, from the recorded receiver signals, comprises finding the voxel reflectivities as a function of ultrasound component from a beam-formed signal of the receiver signals, filtered by a transmitted pulse shape for the component.

22. A method according to any of claims 13-21, wherein assigning an intensity to each pixel of the output image comprises using a decision function that assigns an intensity to a pixel based on a distribution of the corresponding calculated voxel reflectivities for the different ultrasound components, and substantially the same decision function is used for almost all of the pixels in the output image.

23. A method according to any of claims 13-22, wherein a corresponding calculated voxel reflectivity for one or more of the pixels in the output image is a calculated reflectivity for a voxel at a closest location to the pixel in the field of view.

24. A method according to any of claims 13-23, wherein a corresponding calculated voxel reflectivity for one or more of the pixels in the output image is an interpolation of calculated reflectivities for a plurality of voxels at locations near the pixel in the field of view.

25. A method according to any of claims 13-24, wherein the information about differing corresponding calculated voxel reflectivities for the different transmitted ultrasound components comprises calculated reflectivity as a function of voxel and component, and the intensity assigned to one or more of the pixels in the output image is based on corresponding calculated voxel reflectivities for the different ultrasound components that are normalized to a range of calculated voxel reflectivities for other voxels for the same ultrasound component.

26. A method according to any of claims 13-25, wherein the information about differing corresponding calculated voxel reflectivities for the different transmitted ultrasound components comprises calculated reflectivity as a function of voxel and component, and the intensity assigned to one or more of the pixels in the output image is based on corresponding calculated voxels reflectivities for the different ultrasound components that are smoothed over nearby voxels for the same ultrasound component.

27. A method according to any of claims 13-26, comprising normalizing the output image after assigning an intensity to each pixel of the output image.

28. A method according to any of claims 13-27, comprising smoothing the output image after assigning an intensity to each pixel of the output image.

29. A method according to any of claims 13-28, wherein the array of ultrasound receivers elements is in acoustic contact with a surface of the body outside the ribs, extending over a plurality of ribs and a plurality of spaces between ribs, and the field of view comprises tissue inside the rib cage, and reflections of ultrasound from the tissue inside the rib cage are received by the array of ultrasound receiver elements substantially only when the ultrasound propagates between the ribs, and not through the ribs.

30. A method according to any of claims 13-29, wherein the information about calculated voxel reflectivities for the different transmitted ultrasound components comprises a reflectivity image of the field of view for each component.

31. A method according to any of claims 13-30, wherein at least some of the receiver elements in the sparse array are spaced further apart from their neighbors than half a wavelength for at least one frequency of the resonant frequency bands at a sound speed of 1540 meters per second.

32. A method according to any of claims 13-31, wherein transmitting the plurality of ultrasound components into the field of view comprises transmitting at least two of the ultrasound components limited respectively to a first and a second of the well separated frequency bands, with no other one of the well separated frequency bands between them in frequency, and wherein the integrated transmitted ultrasound power in each of the first and second frequency bands is at least 1.5 times the integrated ultrasound power transmitted at frequencies between the first and second frequency bands, and wherein a separation between the first and second frequency bands is at least 1.5 times greater than a width of each of the first and second frequency bands.

33. A method of generating ultrasound output images of a specified type, according to any of claims 13-32, the method comprising first choosing a set of values of parameters for one or more of transmitting the ultrasound components, receiving the reflections, finding the information, and assigning the intensity, by:1) evaluating each of a plurality of the sets of values for a same simulated standard field of view chosen according to the specified type of ultrasound images, by:i) finding simulated receiver signals for the standard field of view using that set of values; ii) generating an output image for that set of values from those simulated receiver signals; iii) generating an uncorrected image for that set of values from those simulated receiver signals but with the receiver signals from only one of the frequency bands; and iv) estimating an amplitude of grating lobe artifacts in the output image for that set of values and in the uncorrected image for that set of values;2) choosing a set of values for which the estimated amplitude of grating mode artifacts is lower for the output image than for the uncorrected image; andthen generating ultrasound output images of the specified type using the chosen set of values.

34. A system according to any of claims 1-12, wherein the controller is configured to reconstruct the output image so that, for a field of view consisting of a single centrally located point reflector, grating lobe artifacts are reduced relative to an image of the field of view reconstructed from voxel reflectivities for only one of the frequency components.

Images