Time-sharing transmission and continuous recording of ultrasound data for enhanced frame rates
Through time-sharing transmission and continuous recording design, the pseudo-random sequence overlap of ultrasonic beams is solved by using the problem of high frame rate and high signal-to-noise ratio in medical ultrasonic imaging, efficient ultrasonic data acquisition, and improved diagnostic imaging capabilities.
Patent Information
- Application Number
- CN202380026916.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-01-30
- Filing Date
- 2023-01-30
- Publication Date
- 2024-11-08
AI Technical Summary
It is difficult to achieve image acquisition with high frame rate and high signal-to-noise ratio simultaneously in medical ultrasound imaging. Traditional methods often sacrifice image resolution or signal-to-noise ratio when frame rate increases.
Using time-sharing transmission and continuous recording (mixed) design, pseudo-random sequences of multiple ultrasonic beams overlap in time and space, and using analog electronic devices and analog digital converters to record and process ultrasonic RF data to reduce data acquisition time and improve Frame rate.
The combination of high frame rate and high image quality is achieved, which significantly reduces the data acquisition time and data size, and improves the diagnostic ability of ultrasound imaging.
Smart Images

Figure CN118922132A_ABST
Abstract
Description
[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 304,633, filed on January 30, 2022, which is incorporated herein by reference for all purposes as if fully set forth herein. Technical Field
[0002] The present invention relates to the acquisition and processing of ultrasound data for medical applications. Specifically, the present invention addresses two competing needs in medical diagnostic imaging: (1) high frame rates for imaging active tissue and organs such as blood and heart, and (2) accurate detection of small lesions requiring sharp resolution and high signal-to-noise ratio. Background Art
[0003] Compared with computerized tomography (CT) and magnetic resonance imaging (MRI), medical ultrasound imaging for diagnostic purposes has advantages such as reasonable cost, real-time imaging, portability, and its harmless effects [1,2]. However, the resolution of ultrasound imaging systems is generally lower than that of CT and MRI systems [3]. Due to better data acquisition hardware and sophisticated processing software, ultrasound imaging technology is developing towards high quality and clear resolution [4].
[0004] Common ultrasound data acquisition methods for medical applications include focused beams and plane wave beams. In ultrasound data acquisition using focused beams, the time delay of each transmitter is electronically controlled in such a way that the transmitters used by this beam transmit waves that arrive at the focus at the same time. The acoustic transmission at the focus is very intense, and the acoustic transmission quickly fades away from the focus. In ultrasound data acquisition using a plane wave beam, the time advance of each transmitter is similar to that of a divergent beam, except that the virtual focus is far behind the transducer. All transmitters participate in the excitation of each plane wave beam. The acoustic transmission of a plane wave beam in the image domain is weak and uniform. Most commercial ultrasound scanners use focused beam data acquisition because the signal-to-noise ratio is much higher due to the focusing ability of a focused beam.
[0005] In most conventional data acquisition methods, ultrasound beams are acquired sequentially, one after the other. The time required to collect each beam is determined by the travel time of the echo from the transducer to the deepest tissue being examined. The total lapse time to collect a complete data set used to produce one frame of an image is the product of the number of beams and the time required for one beam. The temporal resolution, expressed in number of frames per second (NFS), is the inverse of the total lapse time. A typical value for the number of beams required for a high-quality image is about 200. A typical value for the data recording time for one beam is about 200 microseconds. For data acquisition using a focused beam, the frame rate is approximately 25. For plane wave data acquisition, one beam still requires about 200 microseconds to collect. If we use only one plane wave beam in beamforming, the frame rate can be as high as 5000 [5]. If we use 100 plane wave beams in beamforming, the frame rate drops to 50. The higher frame rates using plane wave beams are achieved at the expense of resolution and signal-to-noise ratio in the final image. The same is true for other proposals that use a few focused beams to achieve higher frame rates [6]. Summary of the invention
[0006] In one embodiment, the present application discloses a method for acquiring ultrasound radio frequency data using a time-sharing transmission and continuous recording (BLEND) design. The method includes: providing an ultrasound transducer, the ultrasound transducer including multiple elements; providing an ultrasound data acquisition system, the ultrasound data acquisition system including analog electronics, an analog-to-digital converter, and a CPU (central processing unit) or a GPU (graphics processing unit); using the ultrasound transducer and the transmission function of the ultrasound data acquisition system to transmit multiple ultrasound beams; using the ultrasound transducer and the receiving function of the ultrasound data acquisition system to receive the ultrasound beams; using the ultrasound data acquisition system to record raw radio frequency data; sending the raw radio frequency data to a processing unit; and demixing the raw radio frequency data into separate ultrasound beam records. The ultrasound beams are transmitted in a manner such that a subsequent ultrasound beam is transmitted before a previous ultrasound beam is received by the ultrasound transducer; and the ultrasound beams overlap in time according to a first pseudo-random sequence and overlap in space according to a second pseudo-random sequence.
[0007] In another embodiment, the ultrasonic transducer is a 1D linear array, a 1D curved array, a 1D phased array, or a 2D matrix array.
[0008] In another embodiment, the element is first used as a transmitter and then as a receiver.
[0009] In another embodiment, the elements include a first row of elements and a second row of elements; the first row of elements are dedicated to either transmitting or receiving; and the second row of elements are dedicated to either receiving or transmitting.
[0010] In another embodiment, the elements include odd-numbered elements and even-numbered elements; the odd-numbered elements are dedicated to transmission or reception; and the even-numbered elements are dedicated to reception or transmission.
[0011] In another embodiment, the data acquisition system includes analog electronics, one or more analog to digital converters, and one or more CPUs (or GPUs).
[0012] In another embodiment, the time gap (jitter time) between two adjacent ultrasound beams is randomly selected between 0 microseconds and 200 microseconds.
[0013] In another embodiment, demixing the ultrasound RF data includes: (i) taking the ultrasound RF data as input; (ii) extracting raw beam records by inverting jitter time and applying receiver apodization to the ultrasound beams; (iii) sorting the raw beam records into a common receiver domain; (iv) performing de-spike and random noise attenuation in the common receiver domain; and (v) re-sorting the processed data to obtain the individual ultrasound beam data.
[0014] In another embodiment, the present application provides a system for acquiring and processing ultrasound radio frequency (RF) data using a time-sharing transmission and continuous recording (hybrid) design. The system includes: an ultrasound transducer, the ultrasound transducer includes a plurality of elements; a data acquisition system, the data acquisition system includes analog electronics, an analog-to-digital converter, and a first CPU (central processing unit) or a first GPU (graphics processing unit); a display device; a keyboard; a pointing device; a data acquisition device, the data acquisition device includes an analog-to-digital converter (ADC); and a processing unit, the processing unit includes a second CPU or a second GPU. The first and second CPUs and the first and second GPUs are suitable for: acquiring raw RF data via the ultrasonic transducer and the data acquisition system by transmitting multiple ultrasonic beams in a manner such that a subsequent ultrasonic beam is transmitted before a previous ultrasonic beam is received by the ultrasonic transducer; recording the raw RF data using the data acquisition system; sending the raw RF data to the processing unit; demixing the raw RF data into separate ultrasonic beam data; processing the demixed ultrasonic beam data and sending the demixed ultrasonic beam data to a CPU memory or a GPU memory; beamforming the demixed ultrasonic beam data on the first and second CPUs or the first and second GPUs to obtain an ultrasonic image; and processing the ultrasonic image and sending the ultrasonic image to the display device.
[0015] In another embodiment, the display device is remotely connected to the processing unit via an internet connection, a wireless connection, or a satellite connection.
[0016] In another embodiment, the ultrasonic transducer is a 1D linear array, a 1D curved array, a 1D phased array, or a 2D matrix array.
[0017] In another embodiment, the keyboard is a wireless keyboard or a software keyboard installed on the processing unit.
[0018] In another embodiment, the element is first used as a transmitter and then as a receiver.
[0019] In another embodiment, the elements include a first row of elements and a second row of elements; the first row of elements are dedicated to either transmitting or receiving; and the second row of elements are dedicated to either receiving or transmitting.
[0020] In another embodiment, the elements include odd-numbered elements and even-numbered elements; the odd-numbered elements are dedicated to transmission or reception; and the even-numbered elements are dedicated to reception or transmission.
[0021] In another embodiment, demixing the raw RF data includes: (i) taking the ultrasound RF data as input; (ii) extracting raw beam records by inverting jitter time and applying receiver apodization to the ultrasound beams; (iii) sorting the raw beam records into a common receiver domain; (iv) performing de-spike and random noise attenuation in the common receiver domain; and (v) re-sorting the processed data to obtain the individual ultrasound beam data.
[0022] In another embodiment, the data acquisition system and the processing unit share the same CPU or the same GPU, and the first CPU is the same as the second CPU or the first GPU is the same as the second GPU.
[0023] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are intended to provide further explanation of the invention as claimed.
[0024] BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention.
[0026] In the attached picture:
[0027] Figure 1 Sketch of a hybrid recording for a transducer with 128 elements. The vertical axis is the element position. The horizontal axis is the long elapsed time (approximately 100,000 time samples). The different beams were recorded at randomly jittered times and in random order. In this diagram, B15 represents beam #15, B72 is beam #72, etc.
[0028] Figure 2 Illustration of the transmission design for hybrid ultrasound data acquisition using a vertically positioned 128-element linear array transducer. T0 is the time when the recording function starts. t0 is the jitter time for a given beam. Tx delay is the difference between the timestamp at the probe center and the timestamp at the transmitter.
[0029] Figure 3 Schematic diagram of the receive design for hybrid ultrasound data acquisition using a vertically positioned 128-element linear array transducer. The horizontal axis is elapsed time. The recording circuit is always on, continuously recording the signal arriving at the analog-to-digital converter (ADC). Each element is turned on and off multiple times for transmission and reception: turned off when the element is used for transmission, and turned on after the transmission is completed. In this diagram, the white areas are when reception is on (>95%), and the black areas are when reception is off (<5%).
[0030] Figure 4 The phantom model used for the simulation is shown: in the phantom, the white dots are point scatterers and the white lines are reflectors.
[0031] Figure 5 A comparison of raw data of a focused beam in conventional acquisition (left) and an extracted record of the same beam in hybrid acquisition (right) is shown. The hybrid record contains interference from many neighboring beams.
[0032] Figure 6 Shown are (a) a blended beam record, (b) a common receiver gather of the blended data, and (c) the same beam record after de-blending.
[0033] Figure 7 Comparison of a focused beam image with conventional data acquisition (left) and the same image with hybrid data with hybrid data acquisition (right): The same focused beam was used in both images. The data collection time in the conventional data acquisition (left) was 38.4 milliseconds, and the data collection time in the hybrid data acquisition (right) was 5.0 milliseconds. All displays are shown at 60 dB.
[0034] Figure 8 Schematic representation of a hybrid data acquisition and imaging architecture according to one embodiment of the present invention. DETAILED DESCRIPTION
[0035] Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings.
[0036] The present invention proposes a novel design for acquiring radio frequency ultrasound data (also called RF data) using a linear 1D array, a curved 1D array, a phased 1D array or a matrix 2D array transducer. In this design, the recording circuit of the data acquisition system is always turned on, thereby continuously recording the signal arriving at the analog-to-digital converter (ADC). At certain time intervals, short pulse excitations are transmitted to the tissue being examined. The transmitters used for the beams are switched to the receiving mode immediately when their transmission is completed, unless they are dedicated only to transmission. This process is repeated until all required beams are transmitted and received. The beams are jittered in time with a pseudo-random sequence. They are also separated in space with another pseudo-random sequence. We call this data acquisition design and process time-sharing transmission and continuous recording, or simply hybrid. This hybrid data acquisition method is unique and effective, thereby achieving much higher frame rates without perceptible image quality loss. Before beamforming, appropriate demixing of the hybrid recording is required. After demixing, the RF data is fed into a beamformer specific to each beam type. Excellent image quality and extremely high frame rates are achieved using our hybrid design and process.
[0037] We call our invention time-sharing transmission and continuous recording, or hybrid for short. This is a new way to acquire raw ultrasound data. We allow time-sharing in the transmission and reception of ultrasound beams. The time-sharing is designed in a way that will allow us to blend the measured data to recover the individual ultrasound beams. Our hybrid design and process is the fastest way to acquire the same amount of ultrasound beam data. The ability to quickly illuminate large amounts of tissue with insonification of many ultrasound beams at overlapping times makes our hybrid design and process unique and suitable for diagnosing cardiovascular disease, heart disease, blood blockages, and micro-blood flow around malignant tumors, all of which require high frame rates and high image quality.
[0038] Technical Description
[0039] Focused ultrasound beams are widely used in commercial diagnostic imaging of tissues and organs due to their high resolution and high signal-to-noise ratio [1,3]. Less common are diverging ultrasound beams and plane wave ultrasound beams. High frame rates in data acquisition are necessary for imaging objects in motion such as blood flow, heart, and microvibrations of tissue. Ultrafast imaging can be achieved using one or a few plane wave beams at the expense of image resolution and signal-to-noise ratio [5]. We propose a new data acquisition method that achieves both high frame rates and good image quality (resolution and signal-to-noise ratio). We call our special design and process time-sharing transmission and continuous recording (hybrid).
[0040] Part 1: Time-sharing transmission and continuous recording (hybrid)
[0041] 1.1 Definition of Mixing
[0042] In our time-sharing transmission and continuous recording (hybrid) design, the recording circuit is always on during the acquisition of one frame's worth of ultrasound data. The subsequent beam is transmitted before the previous beam is completely recorded. The spatial sequence of the beams is random and separated. The time gap between two adjacent beams (also called jitter time) is also randomly selected. This special design and process is intended to shorten the data acquisition time and reduce the data size. All beams are mixed and stored in a long data buffer, hence the nickname hybrid. In the hybrid data, there is no clear separation of one beam from the other beams ( Figure 1 ).
[0043] exist Figure 1As an illustrative example, we choose a linear array transducer with 128 elements. Let us assume that for one frame worth of ultrasound data, we need 256 focused beams, each beam having 3000 samples in time (or 150 microseconds, assuming a 20MHz data sampling rate). For conventional data acquisition, the total data size will be 256×128×3000×4=393.2MB, and the total data collection time will be: 150×256=38,400 microseconds. The maximum frame rate is less than 26. For our hybrid data acquisition, we use a jitter time of 20 microseconds, which will allow recording of 6 overlapping beams. The recording buffer has 10,000 samples in time, and each sample has 4 bytes. The data size will be 128 (channels)×100,000 (data samples)×4 (bytes)=51.2MB. And the total data collection time will be: 100,000 (data samples) x 0.05 (sampling rate in microseconds) = 5,000 microseconds. The maximum frame rate is now close to 200. For the same number of ultrasound beams, we achieve an 86.9% reduction in data size, an 86.9% reduction in data acquisition time, and an approximate 600% increase in frame rate.
[0044] 1.2 Hybrid Transmission Design
[0045] Figure 2 A transmission design for acquisition of hybrid ultrasound data is shown. The horizontal axis is elapsed time and the vertical axis is element position. We use a linear array of 128 elements as an example. The beam type is focused beam. The design is equally applicable to other array configurations, such as linear arrays with more or less than 128 elements, curved arrays with any number of elements, or phased arrays with any number of elements. The design is also equally applicable to all beam types: focused beam, diverging beam, plane wave beam, and other beam types. In Figure 2 In the figure, the transducer is placed vertically on the left side. The sound waves propagate into the human tissue from left to right. A set of focused beams are transmitted at different times (also called jitter times). Each beam has its own Tx aperture and Rx aperture settings. The jitter times are randomly spaced between (T0, Tmax), approximately at intervals of 20-40 microseconds. The transmission sequence of the beams is also randomized. In this way, we can achieve better demixing results in subsequent data processing steps before beamforming.
[0046] 1.3 Hybrid Receiver Design
[0047] Figure 3A receive design for hybrid ultrasound beam acquisition is shown. The horizontal axis is elapsed time and the vertical axis is element position. We use a linear array of 128 elements as an example. Other array configurations work equally well. The recording circuit is turned on at T0 and remains on at all times, continuously recording the signal reaching the ADC. Each receiving element on the transducer can be turned on and off multiple times: turned off when the element is used for transmission, and turned on after the transmission event is completed. This process is repeated until all beams are transmitted. The thick black lines are the time intervals when the elements are used for transmission. No useful data is recorded in the time intervals marked by the black lines. These black lines are data gaps. These gaps are small and random, so they have minimal impact on image quality.
[0048] 1.4 Other Design Considerations
[0049] There are several variations in the hybrid design. One design uses the same row of elements on the transducer for both transmission and reception, where the elements are first used as transmitters and then as receivers. Another design requires two rows of elements on the transducer, one row dedicated to transmission and the other dedicated to reception. Yet another design alternates the elements on the transducer: odd elements for transmission and even elements for reception, or vice versa. In all designs, the recording circuit continuously records the signal without interruption. In the second and third designs, there are no data gaps because there is no time sharing between transmission and reception for a given element.
[0050] Part II: Wavefield separation of mixed records
[0051] The hybrid recording contains all the ultrasound beams needed to produce a high-quality image. These beams overlap in time and space according to two pseudo-random sequences: one pseudo-random sequence for the dithering time and another pseudo-random sequence for the spatial position of each beam. In this section, we disclose a method for demixing the hybrid recording into the individual ultrasound beam data.
[0052] 2.1 Separating mixed records into beams
[0053] Receiver position x j The input data sample (B(x j , t)) is the superposition of all beams with time jitter and apodization:
[0054]
[0055] Where N b is the total number of beams, S i (x j ,t) represents the data of the i-th beam at the receiver position, δ i is the jitter time of the i-th beam, αij is the receiver apodization of the jth receiver for the i-th beam. The set of jitter times {δ i The order in which the beams are transmitted is in another pseudo-random sequence.
[0056] The individual beam data in equation (1) is another superposition of the acoustic response of the individual transmitters with a set of transmitter delays (also called Tx delays) and apodization:
[0057]
[0058] Where M is the total number of elements on the transducer, u(x k ,x j ,t) represents x k The transmitter at x j The individual acoustic response of the receiver at ik is the Tx delay of the kth transmitter for the i-th beam, α ik is the transmitter apodization of the kth transmitter for the i-th beam. If the transmitter or receiver is active, we usually set α ij or α ik are set to 1, otherwise we set them to zero.
[0059] Combining equations (1) and (2) we obtain an expression for the mixed record from the acoustic responses of the individual transmitters (also called the synthetic aperture radar record):
[0060]
[0061] Equation (3) is accurate for all beam types: focused beam, diverging beam, plane wave beam, to name a few. The difference between them is the Tx delay setting. Equation (3) is also accurate for all subapertures, since the apodization of both the transmitter and the receiver can be set arbitrarily. The formula in equation (3) applies equally to the 2D and 3D cases, since the coordinates xk and x j Can be a vector in the (x,y) plane.
[0062] The key to separating the mixed record into beams (according to equation (1)) or synthetic aperture radar records (according to equation (3)) is to exploit the pseudo-randomness of both the jitter time and the beam transmission sequence. Extracting the beam from the mixed record can be simply written as:
[0063] S i (x j ,t)=α ij B(x j ,t+δ i ) (4)
[0064] Equation (4) reverses the mixing process of equation (1). It is important to note that the extracted beam data will contain a lot of interference from other overlapping beams. If the technician classifies the same data into a domain other than the usual beam display, the interference will appear random. For example, the technician can first classify the extracted beam data into a common receiver domain. In the common receiver domain, the interference noise from other beams appears random by design. The signals belonging to the corresponding beams are still coherent. The random noise in the common receiver domain can be attenuated using simple methods such as median filters [7] or complex algorithms such as singular value decomposition [8,9] or any other similar method. Another classification is then performed to reverse the first classification, thereby putting the data back into the individual beams.
[0065] 2.2 Implementation Considerations
[0066] The implementation of hybrid data acquisition and processing requires some special considerations as summarized below:
[0067] 1. Time Gain Control (TGC) settings: TGC is used in commercial ultrasound scanners to enhance reflections from deep tissues before analog-to-digital conversion. We need to reduce or remove TGC in hybrid data acquisition because a new beam is transmitted before deep reflections from the previous beam are recorded. We may not need TGC at all for the following reasons: There is so much mixing of early (shallow) reflections with late (deep) reflections that the resulting signal amplitude does not vary dramatically based on elapsed time.
[0068] 2. Jitter time setting: We use (δi=α*i+b*random() in our jitter time design, where the integer i is the beam sequence number, a is usually between 20 microseconds and 40 microseconds, and b is usually set at 10 microseconds. The function random() is a random number generator, and its value is between -1 and +1.
[0069] 3. Beam sequence setting: To maximize the randomness in the mixing of the hybrid recordings, we use another pseudo-random sequence when selecting the beams for the sequence of jittering times. The selection criterion is to maximize the overall spatial separation between the beams.
[0070] 4. Random interference attenuation after beam extraction: We do not need to attenuate the random interference in the extracted beam data very aggressively. The reason is that the subsequent beamforming will automatically attenuate these random noises because they are not stationary and do not satisfy the imaging conditions implicit in all well-known beamforming algorithms.
[0071] 5. Transducer selection: If there are two rows of elements in the transducer, we can dedicate one row to transmit and the other row to receive. This is ideal for hybrid data acquisition. For conventional ultrasound transducers, we can use odd elements for transmit and even elements for receive, or we can use all elements for transmit and receive by allowing time gaps in hybrid recordings when elements are used for transmit.
[0072] 6. Selection of beam type: This design and process is applicable to all beam types. We prefer focused beams or focusing beams because we can maximize the spatial separation of the beams.
[0073] 7. Choice of beamformer: We prefer advanced beamformer algorithms over conventional scanline based dynamic focusing (in the receiver). Dual dynamic focusing in both transmitter and receiver has better capability in random noise attenuation.
[0074] Part III: Examples
[0075] 3.1 Echo data simulation
[0076] We used a modified version of the Fresnel simulator from Ultrasound Toolbox (USTB, https: / / www.ustb.co) to generate numerical ultrasound beam data. The use of this simulator is subject to the citation rules. We sincerely thank the authors for making it available in the public domain
[11] . The simulator is based on the Fresnel approximation of acoustic diffraction for rectangular transducers in linear time invariant (LTI) systems. The inputs to the simulator include phantom specifications, transducer specifications, and waveform specifications. The phantom used in this simulation contains:
[0077] Rectangular boxes with a depth range between 7mm and 9mm,
[0078] 4 flat continuous reflectors at 20mm, 40mm, 60mm and 80mm depths,
[0079] A highly echogenic target with a radius of 8 mm at a depth of 70 mm and a second most echogenic target with a radius of 6 mm at a depth of 50 mm,
[0080] A row of scatter points at 30 mm depth and a column of scatter points at the center of the model.
[0081] Figure 4is a depiction of the phantom. The transducer is a linear array with 256 elements (pitch size is 0.2 mm), and each element has a width of 0.18 mm and a height of 5 mm. The center frequency of the simulated echo data is 6 MHz, the effective bandwidth is 80%, and the sampling frequency is 48 MHz.
[0082] We have used a Tx subaperture of 128 elements and an Rx subaperture of 256 elements to simulate a total of 256 focused beams. The focal depth is set at 25mm. We add each beam to the blending buffer in a random order and with increasing jitter time. The jitter time contains a random time perturbation. We store the jitter time and beam coordinates for later use in blending. Figure 5 Shown is a focused beam near the center of the transducer (left) and a portion of a fully hybrid record that coincides with the focused beam (right). The extracted hybrid record contains the same information as the focused beam, plus additional interfering energy from adjacent beams. This interfering energy needs to be removed prior to beamforming.
[0083] 3.2 De-hybrid testing
[0084] Figure 6 Illustrate the process and results of demixing. The left figure shows a mixed beam record extracted from a fully mixed record. We first sort all mixed beam records into a common receiver gather, as shown in the middle figure. In the common receiver gather, the true acoustic signal is still coherent, while the interference energy appears random by design. We then apply a de-spike filter and a random noise removal filter to all common receiver gathers. Finally, we sort the data back into individual beam records. The right figure shows the same beam record after demixing. All interference energy from adjacent beams is effectively removed. The true acoustic signal is nearly intact.
[0085] 3.3 Beamforming Test
[0086] Figure 7 Comparison showing an image of 256 focused beams acquired sequentially (left) and another image containing hybrid data of the same 256 focused beams (right). We see similar resolution and image quality between conventional ultrasound data acquisition and our hybrid data acquisition. The benefit of our hybrid data acquisition is a significant reduction in data acquisition time. The sequential data acquisition on the left requires a wall clock time of 38.7 ms. The hybrid data acquisition on the right requires a wall clock time of 5 ms.
[0087] Figure 8Schematic representation of a hybrid data acquisition and imaging architecture for one embodiment of the present invention. The system and controller contain a data acquisition unit and one or more CPUs and one or more GPUs. One of the CPUs or GPUs sends instructions to the data acquisition system at a certain timestamp to transmit an acoustic pulse to each element within the transmission aperture with a time delay specifically designed for a given beam. After a short delay for transmission, the CPU or GPU sends instructions to the data acquisition system to switch each element of the transducer to receive mode. This process is repeated until all beams are transmitted. The recording function in the data acquisition system is always turned on, thereby continuously recording acoustic echoes reflected from tissue contrast. The echo signals contain interference from all beams because they are recorded into a long data buffer. They are sent to a special processing unit for demixing and beamforming on the CPU, GPU, or both. The final image is displayed on a local monitor or transmitted to a remote display device via TCP / IP.
[0088] It will be apparent to those skilled in the art that various modifications and variations may be made in the present invention without departing from the spirit or scope of the present invention. Therefore, the present invention is intended to cover modifications and variations of the present invention as long as they are within the scope of the appended claims and their equivalents.
[0089] References:
[0090] [1] Richard SCCobbold (2007), Foundations of Biomedical Ultrasound, Oxford University Press, pp. 431-437.
[0091] [2]OH Schuck (1957), Variable Focus Transducer, U.S. Patent 3,090,030, May 14, 1963.
[0092] [3] BS Hertzberg and WD Middleton (2016), Ultrasound: The Requisites, 3rd ed., Elsevier. Chapter 1, pp. 3–31. Also on expertconsult.com.
[0093] [4] P. Suetens (2009), Fundamentals of Medical Imaging. 2nd edition, Cambridge University Press, pp. 33-158.
[0094] [5] G. Montaldo, M. Tanter, J. Bercoff, N. Benech, and M. Fink (2009), Coherent Plane wave Compounding for very high Frame Rate Ultrasonography and Transient Elastography, IEEE Trans Ultrasonics, Ferroelectronics and Frequency Control, Vol. 56, pp. 489–506.
[0095] [6] D. J. Napolitano, B. D. DeBusschere, G. W. McLaughlin, L. Y. Mo, C. H. Chou, T. L. Ji, R. W. Steins (2011), Continuous Transmit Focusing Method and Apparatus for Ultrasound Imaging Systems, U.S. Patent 8,002,705, issued in August 2011.
[0096] [7] M. A. Gungor and I. Karagoz (2016), The effects of the median filter with different window sizes for ultrasound image, 2nd IEEE International Conference on Computer and Communications (ICCC), 549-552, doi:10.1109 / CompComm.2016.7924761.
[0097] [8] C. Demené, T. Deffieux, M. Pernot, BF Osmanski et al. (2015), Spatiotemporal Clutter Filtering of Ultrafast Ultrasound Data Highly Increases Doppler and fUltrasound Sensitivity., IEEE Trans. Med. Imaging 34, 2271–2285.
[0099] [9]P. Song, J. D. Trzasko, A. Manduca, B. Qiang, R. Kadirvel, D. F. Kallmes, and S. Chen (2017), Accelerated Singular Value-Based Ultrasound Blood Flow Clutter Filtering With Randomized Singular Value Decomposition and Randomized Spatial Downsampling, IEEE Trans Ultrason Ferroelectr Freq Control. 64, 706-716. doi:10.1109 / TUFFC.2017.2665342。
[0100]
[10] A. Rodriguez-Molares, Fresnel simulator, http: / / www.ustb.no / examples / fresnel / 。
Claims
1. A method for acquiring ultrasound radio frequency data using a time-sharing transmission and continuous recording (hybrid) design, the method include: providing an ultrasonic transducer comprising a plurality of elements; An ultrasound data acquisition system is provided, the ultrasound data acquisition system comprising analog electronics, an analog-to-digital converter, and a CPU (central processing unit) or a GPU (graphics processing unit); transmitting a plurality of ultrasound beams using the ultrasound transducer and a transmission function of the ultrasound data acquisition system; receiving the ultrasound beam using the ultrasound transducer and a receiving function of the ultrasound data acquisition system; recording raw radio frequency data using the ultrasound data acquisition system; sending the raw radio frequency data to a processing unit; and demixing the raw RF data into individual ultrasound beam records, wherein the ultrasound beams are transmitted in a manner such that a subsequent ultrasound beam is transmitted before a previous ultrasound beam is received by the ultrasound transducer; and The ultrasound beams overlap in time according to a first pseudo-random sequence and in space according to a second pseudo-random sequence.
2. The method according to claim 1, wherein the ultrasonic transducer is a 1D linear array, a 1D curved array, a 1D phased array or a 2D matrix array.
3. The method of claim 1, wherein the element is first used as a transmitter and subsequently as a receiver.
4. The method of claim 1, wherein the elements include a first row of elements and a second row of elements; the first row of elements are dedicated to either transmitting or receiving; and the second row of elements are dedicated to either receiving or transmitting.
5. The method of claim 1, wherein the elements include odd-numbered elements and even-numbered elements; the odd-numbered elements are dedicated to transmission or reception; and the even-numbered elements are dedicated to reception or transmission.
6. The method of claim 1, wherein the data acquisition system comprises analog electronics, one or more analog to digital converters, and one or more CPUs (or GPUs).
7. The method of claim 1, wherein the time gap (jitter time) between two adjacent ultrasound beams is randomly selected between 0 and 200 microseconds.
8. The method of claim 1, wherein the ultrasound radio frequency data is demixed include: (i) taking the ultrasound radio frequency data as input; (ii) extracting the original beam record by inverting the jitter time and applying receiver apodization to the ultrasound beam; (iii) sorting the raw beam records into a common receiver domain; (iv) performing de-spiking and random noise attenuation in the common receiver domain; as well as (V) reclassifying the processed data to obtain the individual ultrasound beam data.
9. A system for acquiring and processing ultrasound radio frequency data using a time-sharing transmission and continuous recording (BLEND) design, the system include: an ultrasonic transducer, the ultrasonic transducer comprising a plurality of elements; A data acquisition system, the data acquisition system comprising analog electronics, an analog-to-digital converter, and a first CPU (central processing unit) or a first GPU (graphics processing unit); Display devices; keyboard; indicating equipment; a data acquisition device, the data acquisition device comprising an analog-to-digital converter (ADC); and a processing unit, the processing unit comprising a second CPU or a second GPU, The first CPU and the second CPU as well as the first GPU and the second GPU are adapted to: acquiring raw radio frequency data via the ultrasound transducer and the data acquisition system by transmitting a plurality of ultrasound beams in a manner such that a subsequent ultrasound beam is transmitted before a previous ultrasound beam is received by the ultrasound transducer; recording the raw radio frequency data with the data acquisition system; sending the raw radio frequency data to the processing unit; Demixing the raw radio frequency data into individual ultrasound beam data; processing the demixed ultrasound beam data and sending the demixed ultrasound beam data to a CPU memory or a GPU memory; performing beamforming on the demixed ultrasound beam data on the first CPU and the second CPU or the first GPU and the second GPU to obtain an ultrasound image; and The ultrasound image is processed and sent to the display device.
10. The system of claim 9, wherein the display device is remotely connected to the processing unit via an internet connection, a wireless connection, or a satellite connection.
11. The method of claim 9, wherein the ultrasonic transducer is a 1D linear array, a 1D curved array, a 1D phased array, or a 2D matrix array.
12. The system of claim 9, wherein the keyboard is a wireless keyboard or a software keyboard installed on the processing unit.
13. The system of claim 9, wherein the element is first used as a transmitter and subsequently as a receiver.
14. The system of claim 9, wherein the elements include a first row of elements and a second row of elements; the first row of elements are dedicated to either transmitting or receiving; and the second row of elements are dedicated to either receiving or transmitting.
15. The system of claim 9, wherein the elements include odd elements and even elements; the odd elements are dedicated to transmission or reception; and the even elements are dedicated to reception or transmission.
16. The system of claim 9, wherein the raw RF data is demixed include: (i) taking the ultrasound radio frequency data as input; (ii) extracting the original beam record by inverting the jitter time and applying receiver apodization to the ultrasound beam; (iii) sorting the raw beam records into a common receiver domain; (iv) performing de-spiking and random noise attenuation in the common receiver domain; as well as (v) reclassifying the processed data to obtain the individual ultrasound beam data. 17 . The system of claim 9 , wherein the data acquisition system and the processing unit share a same CPU or a same GPU, and the first CPU is the same as the second CPU or the first GPU is the same as the second GPU.
