Systems and methods for joint scan parameter selection

Abstract

The invention is entitled "Systems and methods for joint scan parameter selection." The invention provides methods and systems for improving image quality of ultrasound images by jointly selecting optimal scan parameter values. In one example, a method includes acquiring a plurality of ultrasound images of an anatomical region, each ultrasound image acquired with a different combination of parameter values for a first scan parameter and a second scan parameter; selecting a first parameter value for the first scan parameter and a second parameter value for the second scan parameter based on image quality of each image; and acquiring one or more additional ultrasound images with the first parameter value and the second parameter value.

Description

TECHNICAL FIELD

[0001] Embodiments of the subject matter disclosed herein relate to ultrasound imaging, and more specifically, to improving image quality for ultrasound imaging. BACKGROUND

[0002] Medical ultrasound is an imaging modality that employs ultrasound waves to probe the internal structures of a patient's body and produce corresponding images. For example, an ultrasound probe including a plurality of transducer elements emits ultrasound pulses that are reflected or backscattered by structures in the body, refracted, or absorbed. The ultrasound probe then receives the reflected echoes, which are processed into an image. The ultrasound image of the internal structures can be saved for later analysis by a clinician to aid in diagnosis and / or can be displayed on a display device in real-time or near real-time. SUMMARY

[0003] In one embodiment, a method includes acquiring a plurality of ultrasound images of an anatomical region, each ultrasound image acquired with a different combination of parameter values for a first scan parameter and a second scan parameter; selecting a first parameter value for the first scan parameter and a second parameter value for the second scan parameter based on image quality of each image; and acquiring one or more additional ultrasound images with the first parameter value and the second parameter value.

[0004] The above advantages of the present specification, and other advantages and features, will be made apparent from the following detailed description. It should be understood that the above summary is intended to introduce some of the concepts of the concepts described further in the detailed description below. It is not intended to identify key or essential features of the claimed subject matter, the scope of which is defined uniquely by the claims that follow the detailed description. Furthermore, the claimed subject matter is not limited to implementations that solve any disadvantages mentioned above or in any part of this disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0005] Various aspects of the disclosure can be better understood with reference to the following detailed description when considered in connection with the accompanying drawings, in which:

[0006] Figure 1 A block diagram illustrating an exemplary embodiment of an ultrasound system is shown;

[0007] Figure 2 is a schematic diagram illustrating a system for generating ultrasound images at optimized parameter settings according to an exemplary embodiment;

[0008] Figure 3 is a schematic diagram illustrating a process for selecting ultrasound scan parameters according to an exemplary embodiment;

[0009] Figure 4 An exemplary embodiment of a method for generating ultrasound images at optimized parameter settings is shown.Figure 3 exemplary ultrasound images and associated image quality metrics acquired by the process; and

[0010] Figure 5A and Figure 5B is a flowchart illustrating an exemplary method for selecting ultrasound scan parameters during ultrasound imaging, according to an exemplary embodiment.

[0011] Figure 6 shows two exemplary timings of image acquisition relative to the cardiac cycle, according to Figure 5A and Figure 5B is an exemplary table showing ultrasound scan parameters that would be applied to acquire ultrasound images that can be used to select ultrasound scan parameters, according to the method of

[0012] Figure 7 shows two exemplary timings of image acquisition relative to the cardiac cycle.

[0013] Figure 8 is a flowchart illustrating an exemplary method for selecting image acquisition post-processing parameters during ultrasound imaging.

[0014] Figure 9 is a chart showing a scan sequence for acquiring a single ultrasound image, according to an exemplary embodiment.

[0015] Figure 10 is a chart showing a scan sequence for acquiring a plurality of ultrasound images, according to an exemplary embodiment. DETAILED DESCRIPTION

[0016] Medical ultrasound imaging typically includes placing an ultrasound probe including one or more transducer elements onto an imaging subject, such as a patient, at the location of a target anatomical feature (e.g., abdomen, chest, etc.). Images are acquired by the ultrasound probe and displayed on a display device in real-time or near real-time (e.g., the images are displayed as soon as they are generated and without intentional delay). An operator of the ultrasound probe can view the images and adjust various acquisition parameters and / or the position of the ultrasound probe in order to obtain high quality images of the target anatomical feature (e.g., heart, liver, kidney, or another anatomical feature). Adjustable acquisition parameters include transmit frequency, transmit depth, gain, beam steering angle, beamforming strategy, and / or other parameters. Changing the acquisition parameters to acquire optimal images (e.g., images of desired quality) can be very challenging and based on user experience. The image quality variation with respect to the acquisition parameters is not a well-studied problem. Thus, the adjustments made by the operator to the acquisition parameters in order to acquire optimal images are typically subjective. For example, the operator can adjust various acquisition parameters until an image that appears optimal to the operator is acquired, and the process of adjusting the acquisition parameters can be undefined or repeated between exams. Additionally, various post-acquisition image parameters that can affect image quality, such as the bandwidth and center frequency of filtering of the received ultrasound data, can also be adjusted by the operator. This subjectivity and lack of deterministic process can result in non-reproducible results, and in many ultrasound exams, the highest quality images possible can not be acquired.

[0017] Accordingly, in accordance with the embodiments disclosed herein, the problem of image acquisition parameter optimization and / or image acquisition post-processing optimization is addressed via a feedback system based on an automated image quality measurement algorithm configured to automatically identify acquisition parameters that will generate the best possible image for the anatomical structure being imaged. The automated image quality measurement algorithm can include an artificial intelligence assisted feedback system to optimize acquisition parameters in a joint manner with a plurality of images each acquired with a different combination of scan parameter values to reach optimal acquisition parameter settings based on automatically identified image quality metrics. For example, a set of images and / or cine loops can be acquired each with a different possible combination of emission depth and emission frequency. The image or cine loop with the highest image quality (e.g., as detected by an artificial intelligence based system) can be identified and the optimal emission depth and optimal emission frequency can be set to the depth and frequency at which the identified highest quality image was acquired. In doing so, the optimal acquisition parameters (e.g., depth and frequency) for a given scan plane / anatomical feature can be identified in a reproducible manner, which can increase consistency in image quality across different ultrasound exams. Additionally, in some examples, different parameter values for one or more acquisition post-processing parameters can be applied to the images to generate a set of replica images each having a different parameter value for each acquisition post-processing parameter. Image quality can be determined for each replica image and the replica image with the highest image quality can be selected. The parameter value for that acquisition post-processing parameter can be set to the parameter value from the selected replica image and applied to subsequent images. Selecting optimal acquisition and / or acquisition post parameters can simplify the operator workflow, which can reduce exam time and can facilitate higher quality exams even for more novice operators.

[0018] Figure 1 An example ultrasound system is shown in FIG. 1, which includes an ultrasound probe, a display device, and an imaging processing system. Via the ultrasound probe, ultrasound images can be acquired and displayed on the display device. The images can be acquired using various scan parameters, such as frequency and depth, which have parameter values that can be selected to increase the image quality of the acquired images. To select the scan parameter values, a plurality of images acquired with different scan parameter values can be analyzed to determine which scan parameter values produce images with the highest image quality. As shown in FIG. 1, the image processing system includes one or more image quality models, such as a depth model and one or more frequency models, which can be deployed according to a joint process to determine the image quality of each of a plurality of images, as shown in FIG. 2, and select scan parameter values that will result in relatively higher image quality. In FIG. 2, the joint process is shown to include a depth model and a frequency model, which can be used to determine the image quality of each of a plurality of images, as shown in FIG. 3, and select scan parameter values that will result in relatively higher image quality. Figure 2 Figure 3 Figure 4 Figure 5A and Figure 5B ​​​A method for jointly selecting scan parameter values ​​during ultrasound imaging is shown in FIG. The method for jointly selecting scan parameter values ​​may include applying a table of scan parameter values, such as Figure 6 This may result in a different order (such as Figure 7 After a target scan parameter value has been selected, different post-acquisition processing parameter values ​​may be applied to an image to create a set of adjusted images, and an image quality model may be deployed to determine which image from the set of adjusted images has the highest image quality, and therefore determine which post-acquisition processing parameter value(s) should be applied to that image and / or subsequent images, as shown. Figure 8 The method is shown in the following. Figure 9 The scanning sequence shown and / or according to Figure 10 The scan sequence shown was used to acquire the images disclosed herein.

[0019] See also Figure 1 , shows a schematic diagram of an ultrasound imaging system 100 according to an embodiment of the present disclosure. The ultrasound imaging system 100 includes a transmit beamformer 101 and a transmitter 102, which drives elements (e.g., transducer elements) 104 within a transducer array (referred to herein as a probe 106) to transmit pulsed ultrasound signals (referred to herein as transmit pulses) into a body (not shown). According to one embodiment, the probe 106 can be a one-dimensional transducer array probe. However, in some embodiments, the probe 106 can be a two-dimensional matrix transducer array probe. As further explained below, the transducer elements 104 can be made of piezoelectric material. When a voltage is applied to a piezoelectric crystal, the crystal physically expands and contracts, thereby emitting an ultrasonic spherical wave. In this way, the transducer elements 104 can convert the electronic transmit signal into an acoustic transmit beam.

[0020] After the elements 104 of the probe 106 transmit pulsed ultrasound signals into the body (of the patient), the pulsed ultrasound signals are backscattered from structures inside the body (such as blood cells or muscle tissue) to produce echoes that return to the elements 104. The echoes are converted into electrical signals or ultrasound data by the elements 104, and the electrical signals are received by the receiver 108. The electrical signals representing the received echoes pass through the receive beamformer 110, which outputs the ultrasound data. In addition, the transducer elements 104 may generate one or more ultrasound pulses based on the received echoes to form one or more transmit beams.

[0021] According to some embodiments, the probe 106 can include electronic circuitry to perform all or part of transmit beamforming and / or receive beamforming. For example, all or a portion of the transmit beamformer 101, the transmitters 102, the receivers 108, and the receive beamformer 110 can be located within the probe 106. In this disclosure, the term "scan" or "in scan" can also be used to refer to the process of acquiring data by transmitting and receiving ultrasound signals. In this disclosure, the term "data" can be used to refer to one or more data sets acquired with the ultrasound imaging system. In one embodiment, a machine learning model can be trained using data acquired via the ultrasound system 100. The user interface 115 can be used to control the operation of the ultrasound imaging system 100, including for input of patient data (e.g., patient history), for changing scan or display parameters, for initiating a probe depolarization sequence, etc. The user interface 115 can include one or more of the following: a rotary element, a mouse, a keyboard, a trackball, hard keys linked to specific actions, soft keys that can be configured to control different functions, and a graphical user interface displayed on the display device 118.

[0022] The ultrasound imaging system 100 also includes a processor 116 to control the transmit beamformer 101, the transmitter 102, the receiver 108, and the receive beamformer 110. The processor 116 is in electronic communication (e.g., communicatively connected) with the probe 106. For the purposes of this disclosure, the term “electronic communication” can be defined to include both wired and wireless communication. The processor 116 can control the probe 106 to acquire data according to instructions stored on the processor’s memory, and / or the memory 120. The processor 116 controls which of the elements 104 are active and the shape of the beams transmitted from the probe 106. The processor 116 is also in electronic communication with the display device 118, and the processor 116 can process data (e.g., ultrasound data) into images for display on the display device 118. The processor 116 can include a central processing unit (CPU) according to one embodiment. According to other embodiments, the processor 116 can include other electronic components capable of performing processing functions, such as a digital signal processor, a field programmable gate array (FPGA), or a graphics board. According to other embodiments, the processor 116 can include multiple electronic components capable of performing processing functions. For example, the processor 116 can include two or more electronic components selected from the list of electronic components including: a central processing unit, a digital signal processor, a field programmable gate array, and a graphics board. According to another embodiment, the processor 116 can also include a complex demodulator (not shown) that demodulates RF data and generates raw data. In another embodiment, the demodulation can be performed earlier in the processing chain. The processor 116 is adapted to perform one or more processing operations according to a plurality of selectable ultrasound modalities on the data. In one example, the data can be processed in real-time during a scan session as the echo signals are received by the receiver 108 and transmitted to the processor 116. For the purposes of this disclosure, the term “real-time” is defined to include processes performed without any intentional delay. For example, embodiments can acquire images at a real-time rate of 7 to 20 frames / second. The ultrasound imaging system 100 can be capable of acquiring 2D data of one or more planes at a significantly faster rate. However, it should be understood that the real-time frame rate can depend on the length of time it takes to acquire each frame of data for display. Thus, the real-time frame rate can be slower when a relatively large amount of data is acquired. Thus, some embodiments can have a real-time frame rate that is significantly faster than 20 frames / second, while other embodiments can have a real-time frame rate that is lower than 7 frames / second. The data can be temporarily stored in a buffer (not shown) during a scan session and processed in a less real-time manner in real-time or offline operations. Some embodiments of the present invention can include multiple processors (not shown) to handle processing tasks handled by the processor 116 according to the example embodiments described above.For example, in the display of images, a first processor can be used to demodulate and decimate the RF signals, while a second processor can be used to further process the data (e.g., by augmenting the data as described further herein). It will be understood that other implementations can use different processor arrangements.

[0023] The ultrasound imaging system 100 can continuously acquire data at a frame rate of, for example, 10 to 30 Hz (e.g., 10 to 30 frames per second). Images generated from the data can be refreshed on the display device 118 at a similar frame rate. Other implementations can acquire and display data at different rates. For example, some implementations can acquire data at a frame rate of less than 10 Hz or greater than 30 Hz, depending on the size of the frames and the intended application. A memory 120 is included for storing frames of processed acquired data. In an exemplary implementation, the memory 120 has sufficient capacity to store at least several seconds of ultrasound data frames. The data frames are stored in a manner that facilitates retrieval according to their order of acquisition or time. The memory 120 can include any known data storage medium.

[0024] In various implementations of the present application, the processor 116 can process data through different mode dependent modules (e.g., B-mode, color Doppler, M-mode, color M-mode, spectral Doppler, elastography, TVI, strain, strain rate, etc.) to form 2D or 3D data. For example, one or more modules can generate B-mode, color Doppler, M-mode, color M-mode, spectral Doppler, elastography, TVI, strain, strain rate, and combinations thereof, etc. As one example, one or more modules can process color Doppler data, which can include traditional color blood flow Doppler, power Doppler, HD flow, etc. The image lines and / or frames are stored in memory and can include timing information indicative of the time at which the image lines and / or frames were stored in memory. These modules can include, for example, scan conversion modules for performing scan conversion operations to convert acquired images from beam space coordinates to display space coordinates. A video processor module can be provided that reads the acquired images from memory and displays the images in real time as a procedure (e.g., ultrasound imaging) is performed on a patient. The video processor module can include a separate image memory, and ultrasound images can be written to the image memory for reading and display by the display device 118.

[0025] In various embodiments of the present disclosure, one or more components of the ultrasound imaging system 100 can be included in a portable handheld ultrasound imaging device. For example, the display 118 and the user interface 115 can be integrated into an external surface of a handheld ultrasound imaging device, which can further include the processor 116 and the memory 120. The probe 106 can include a handheld probe in electronic communication with the handheld ultrasound imaging device to collect raw ultrasound data. The transmit beamformer 101, the transmitter 102, the receiver 108, and the receive beamformer 110 can be included in the same or different parts of the ultrasound imaging system 100. For example, the transmit beamformer 101, the transmitter 102, the receiver 108, and the receive beamformer 110 can be included in the handheld ultrasound imaging device, the probe, and combinations thereof.

[0026] After performing a two-dimensional ultrasound scan, a data block containing the scan lines and their samples is generated. After applying the back-end filters, a process called scan conversion is performed to transform the two-dimensional data block into a displayable bitmap image with additional scan information such as depth, angle of each scan line, etc. During scan conversion, an interpolation technique is applied to fill in missing holes (i.e., pixels) in the resulting image. These missing pixels occur because each element of the two-dimensional block should typically cover many pixels in the resulting image. For example, in current ultrasound imaging systems, a bi-cubic interpolation is applied, which utilizes neighboring elements of the two-dimensional block. Thus, if the two-dimensional block is relatively small compared to the size of the bitmap image, the scanned-converted image will include areas of poor resolution or low resolution, particularly for areas that are large in depth.

[0027] The ultrasound images acquired by the ultrasound imaging system 100 can be further processed. In some embodiments, the ultrasound images produced by the ultrasound imaging system 100 can be transmitted to an image processing system, where in some embodiments the ultrasound images can be analyzed by using one or more machine learning models trained using ultrasound images and corresponding ground truth outputs in order to assign scan parameter-specific image quality metrics to the ultrasound images. As used herein, ground truth output refers to an expected or “correct” output based on a given input into a machine learning model. For example, if a machine learning model is being trained to classify images of cats, the ground truth output for the model when fed an image of a cat is the label “cat.” As explained in more detail below, if a machine learning model is being trained to classify ultrasound images based on image quality factors associated with depth (e.g., visibility of certain anatomical features), the ground truth output for the model can be a label indicating a level of the image quality factor, e.g., on a scale of 1-5, where 1 is the lowest image quality level (e.g., reflecting insufficient or inadequate depth, least optimal depth) and 5 is the highest image quality level (e.g., reflecting sufficient depth, optimal depth). Similarly, if a machine learning model is being trained to classify ultrasound images based on image quality factors associated with frequency (e.g., speckle), the ground truth output for the model can be a label indicating a level of the image quality factor, e.g., on a scale of 1-5, where 1 is the lowest image quality level (e.g., reflecting high / non-smooth speckle, least optimal frequency) and 5 is the highest image quality level (e.g., reflecting low / smooth speckle, optimal frequency).

[0028] Although described herein as separate systems, it should be understood that in some embodiments the ultrasound imaging system 100 includes the image processing system. In other embodiments, the ultrasound imaging system 100 and the image processing system can comprise separate devices. In some embodiments, the images produced by the ultrasound imaging system 100 can be used as a training data set for training one or more machine learning models, where the machine learning models can be used to perform one or more steps of ultrasound image processing as described below.

[0029] Referring to Figure 2FIG. 2 illustrates an image processing system 202, in accordance with example embodiments. In some embodiments, the image processing system 202 is incorporated into the ultrasound imaging system 100. For example, the image processing system 202 can be disposed in the ultrasound imaging system 100 as the processor 116 and the memory 120. In some embodiments, at least a portion of the image processing system 202 is disposed at a device (e.g., an edge device, a server, etc.) that is communicatively coupled to the ultrasound imaging system via a wired connection and / or a wireless connection. In some embodiments, at least a portion of the image processing system 202 is disposed at a separate device (e.g., a workstation) that can receive images / mapping from the ultrasound imaging system or from a storage device that stores images / data generated by the ultrasound imaging system. The image processing system 202 can be operatively / communicatively coupled to a user input device 232 and a display device 234. At least in some examples, the user input device 232 can comprise the user interface 115 of the ultrasound imaging system 100, and the display device 234 can comprise the display device 118 of the ultrasound imaging system 100.

[0030] The image processing system 202 includes a processor 204 configured to execute machine readable instructions stored in a non-transitory memory 206. The processor 204 can be single- or multi-core and programs executing thereon can be configured for parallel or distributed processing. In some embodiments, the processor 204 can optionally include separate components distributed among two or more devices, which can be remotely located and / or configured for coordinated processing. In some embodiments, one or more aspects of the processor 204 can be virtualized and executed by remotely accessible networked computing devices configured in a cloud computing configuration.

[0031] The non-transitory memory 206 can store an image quality model 208, a training module 210, and ultrasound image data 212. The image quality module 208 can include one or more machine learning models, such as deep learning networks, including a plurality of weights and biases, activation functions, loss functions, gradient descent algorithms, and instructions for implementing the one or more deep neural networks to process input ultrasound images. For example, the image quality model 208 can store instructions for implementing a deep model 209 and / or one or more frequency models 211. The deep model 209 and the one or more frequency models 211 can each include one or more neural networks. The image quality module 208 can include trained and / or untrained neural networks and can also include training routines or parameters (e.g., weights and biases) associated with the one or more neural network models stored therein.

[0032] The depth model 209 can be a neural network (e.g., a convolutional neural network) trained to identify far-field structures in an ultrasound image and determine whether the far-field structures (e.g., structures beyond / below the focal point of the ultrasound beam relative to the transducer of the ultrasound probe) are at an expected depth. The trainable depth model 209 can be trained to identify far-field structures in a scan plane / view-specific manner. For example, the trainable depth model can be trained to identify far-field structures in a four-chamber view of a heart, but not in a parasternal long axis (PLAX) view of a heart. Thus, in some examples, the depth model 209 can actually include multiple depth models, each specific to a different scan plane or anatomical view. The trainable depth model 209 can be trained to output a first image quality metric that reflects the quality of the input ultrasound image as a function of the transmit acquisition depth. For example, the far-field structures identified by the depth model can change in appearance / visibility as the depth changes, and the first image quality metric output by the depth model can reflect the appearance / visibility of these structures as an indicator of whether the depth used to acquire the ultrasound image is an optimal depth.

[0033] The one or more frequency models 211 can include one or more neural networks or other machine learning models trained to output respective second image quality metrics that represent image quality factors that change as a function of the transmit frequency. The one or more frequency models 211 can include a first frequency model that evaluates speckle size (referred to as a speckle model), a second frequency model that evaluates key landmarks (referred to as a landmark detection model), and a third frequency model that evaluates global image quality relative to a library of a large population of ultrasound images (referred to as a global image quality model). The speckle model can be trained to output a speckle image quality metric that reflects the level of speckle smoothness in the input ultrasound image. The speckle image quality metric can increase as the frequency increases when the speckle smoothness increases as the frequency increases. The landmark detection model can be trained to output a landmark image quality metric that reflects the appearance / visibility of certain anatomical features (landmarks) in the input ultrasound image. For example, certain anatomical features, such as valves in a four-chamber view, can start to degrade image quality / appearance as the transmit frequency increases. Thus, the landmark detection model can identify key landmarks in the input ultrasound image and output a landmark image quality metric based on the image quality / visibility of the identified key landmarks. Because the key landmarks change as the scan plane / anatomical view changes, the landmark detection model can include multiple different landmark detection models, each specific to a different scan plane or anatomical view.

[0034] A trainable global image quality model can be trained to assess the overall image quality of an input ultrasound image relative to a large population library of ultrasound images. For example, the global image quality model can be trained on a plurality of ultrasound images of a plurality of different patients, where an expert (e.g., a cardiologist or other clinician) annotates or labels each training ultrasound image with an overall image quality score (e.g., on a scale of 1-5, where 1 is the lowest image quality and 5 is the highest image quality). After training / validation, the global image quality model can then generate an output reflecting a global image quality metric of an input ultrasound image relative to the overall image quality of the training ultrasound images. By including an overall image quality metric that is not particularly affected by depth or frequency, patient-specific image quality issues can be considered.

[0035] The non-transitory memory 206 can also include a training module 210 including instructions for training one or more machine learning models stored in the image quality model 208. In some embodiments, the training module 210 is not disposed at the image processing system 202. Thus, the image quality model 208 includes a trained and validated network.

[0036] The non-transitory memory 206 can also store ultrasound image data 212, such as ultrasound images captured by the ultrasound imaging system 100. Figure 1 For example, the ultrasound image data 212 can include ultrasound image data acquired by the ultrasound imaging system 100. The ultrasound images of the ultrasound image data 212 can include ultrasound images acquired by the ultrasound imaging system 100 at different parameter values of different scan parameters, such as different frequencies and / or different depths. Additionally, when the training module 210 is stored in the non-transitory memory 206, the ultrasound image data 212 can store ultrasound images, ground truth outputs, iterations of machine learning model outputs, and other types of ultrasound image data that can be used to train the image quality model 208. In some embodiments, the ultrasound image data 212 can store ultrasound images and ground truth outputs in an ordered format such that each ultrasound image is associated with one or more corresponding ground truth outputs. However, in examples where the training module 210 is not disposed at the image processing system 202, the image / ground truth outputs that can be used to train the image quality model 210 can be stored elsewhere.

[0037] In some embodiments, the non-transitory memory 206 can include components disposed on two or more devices that can be remotely located and / or configured for coordinated processing. In some embodiments, one or more aspects of the non-transitory memory 206 can include remotely accessible networked storage devices configured in a cloud computing configuration.

[0038] User input device 232 can include one or more of a touchscreen, a keyboard, a mouse, a trackpad, a motion-sensing camera, or other devices configured to enable a user to interact with and manipulate data within image processing system 202. In one example, user input device 232 can enable a user to select an ultrasound image for training a machine learning model, to indicate or label a location of an interventional device in ultrasound image data 212, or for further processing using a trained machine learning model.

[0039] Display device 234 can include one or more display devices using almost any type of technology. In some embodiments, display device 234 can include a computer monitor and can display ultrasound images. Display device 234 can be combined in a shared enclosure with processor 204, non-transitory memory 206, and / or user input device 232, or can be a peripheral display device and can include a monitor, a touchscreen, a projector, or other display devices known in the art that can enable a user to view ultrasound images produced by an ultrasound imaging system and / or interact with various data stored in non-transitory memory 206.

[0040] It should be understood that Figure 2 The illustrated image processing system 202 is for illustration and not limitation. Another suitable image processing system can include more, fewer, or different components.

[0041] Turning to Figure 3 which illustrates a process 300 for jointly selecting scan parameter values for two scan parameters (here, depth and frequency). The scan parameter values can be selected according to which scan parameter values produce the highest image quality, where the image quality is determined by a machine learning model. The method 300 can be performed by Figures 1-2components of the ultrasound system 102, e.g., the image acquisition can be via the ultrasound probe 106 and the parameter value selection can be performed by the image processing system 202. As shown, image acquisition can begin at 304 once an operator of the ultrasound system positions the ultrasound probe to image a target scan plane (shown at 302). The image acquisition can include acquiring a plurality of images, each acquired at a different combination of depth and frequency. For example, a first plurality of images can be acquired, where each of the first plurality of images is acquired at a single first frequency (e.g., 1.4 MHz) of a set of frequencies and a different depth value of a set of depth values (e.g., a first image at 30 cm, a second image at 17 cm, and a third image at 10 cm); a second plurality of images can be acquired, where each of the second plurality of images is acquired at a single second frequency (e.g., 1.7 MHz) of the set of frequencies and a different depth value of the set of depth values; and one or more additional pluralities of images can be acquired, where each of the additional pluralities of images is acquired at a next frequency of the set of frequencies (e.g., one image at 2 MHz and one image at 2.3 MHz) and a different depth value. In this way, images can be acquired for each possible combination of depth and frequency of the set of depths and the set of frequencies.

[0042] Each of the plurality of images is input to a plurality of models, including a depth model 306, which can be Figure 2 a non-limiting example of the depth model 209 of FIG. 2. The depth model 306 can be a machine learning model, such as a neural network, trained to determine a first image quality score that reflects a quality of the input image from the perspective of whether one or more key anatomical features in the target scan plane are sufficiently visible / of sufficiently high quality. The one or more key anatomical features can be features whose visibility / quality changes as a function of depth, and thus can be used as a marker whose depth value can result in a high quality image.

[0043] The one or more models can include a global image quality model 310, a landmark detection model 308, and a speckle model 312 (which can be the models described above with respect to FIG. 2). The global image quality model 310 can be trained to determine a second image quality score that reflects a quality of the input image from the perspective of whether the one or more key anatomical features in the target scan plane are sufficiently visible / of sufficiently high quality. The landmark detection model 308 can be trained to determine a third image quality score that reflects a quality of the input image from the perspective of whether the one or more key anatomical features in the target scan plane are sufficiently visible / of sufficiently high quality. The speckle model 312 can be trained to determine a fourth image quality score that reflects a quality of the input image from the perspective of whether the one or more key anatomical features in the target scan plane are sufficiently visible / of sufficiently high quality. Figure 2The global image quality model, landmark detection model, and speckle model are non-limiting examples of machine learning models that can be used to determine the image quality of each image. Each of the global image quality model 310, landmark detection model 308, and speckle size model 312 can be a machine learning model, such as a neural network. The global image quality model 310 can be trained to assign a score to each image (e.g., on a scale of 1-5) in a population-wide manner based on the image quality of the image. The landmark detection model 308 and the speckle size model 312 can each output a score that reflects how much the image of a given patient changes in quality with ultrasound transmit frequency, as speckle on the ultrasound image can smooth / change in size with increasing frequency, and certain key anatomical landmarks can become more or less visible with frequency changes. The scores output from each of the global image quality model, landmark detection model, and speckle size model can be combined to generate a cumulative second image quality score. The image quality metrics (e.g., from each of the models discussed herein) can be output at 314.

[0044] Thus, after each image is input into the models, two scores can be assigned to each image: a first image quality score, which is output from the depth model 306, and a second image quality score, which is a cumulative score calculated from the output of each of the global image quality model 310, landmark detection model 308, and speckle size model 312. The image from the plurality of images that has the highest combined image quality score (e.g., the first image quality score combined with the second image quality score) can be identified as the selected image, as shown at 316. If there are two or more images that have the same highest combined image quality score, the image acquired at the highest frequency can be selected. The depth used to acquire the selected image can be set to the selected depth value, and the frequency used to acquire the selected image can be set to the selected frequency value. Any additional images of the target scan plane or view that the operator desires and / or that the scan protocol indicates can be acquired at the selected depth value and the selected frequency value, as shown at 318. According to Figure 3 The joint process shown selects the depth value and the frequency value only after all of the images are acquired.

[0045] Figure 3 The joint process shown can result in a total of m x n images being acquired in order to select the depth value and the frequency value for subsequent acquisition. The number m of images can be based on how many depth values are available / selected for optimization. As noted above, there can be three possible depth values, but other numbers of depth values are possible, such as two or four or more. Likewise, the number n of images can be based on how many frequency values are available / selected for optimization. As noted above, there can be four possible frequency values, but other numbers of frequency values are possible, such as three, five, etc. Additionally, while the joint process shown involves acquiring a number of images at each of a number of depth values and a number of frequency values, other joint processes are possible. For example, the joint process can involve acquiring a number of images at each of a number of depth values at a single frequency value, or a number of images at a single depth value at a number of frequency values, or a number of images at a single depth value and a single frequency value. Figure 3A process for selecting values of two scan parameters is shown, but more scan parameters can be selected / optimized, such as gain, beam steering, beamforming strategy, filtering, etc. In such examples, the number of images acquired can be m x n x l, where the number l of images is based on the values of gain (or other parameter, such as beam steering) that can be used for optimization. In this way, images can be acquired for each possible combination of depth, frequency, and / or any other acquisition parameter value.

[0046] The joint process described above can result in more images being acquired compared to a sequential process in which m + n images are acquired and a first scan parameter (e.g., depth) is selected and then a second scan parameter (e.g., frequency) is selected, which can make the sequential process more practical and easier to implement than the joint process. However, for the joint process, the total time to select parameter values of the scan parameters can be shorter. Moreover, when the scan parameters are dependent on each other, the joint process of acquiring all possible combinations can reveal any unexpected combinations that result in high image quality, whereas the sequential process can omit possible combinations based on the assumption that, for example, frequency changes do not affect image quality based on depth.

[0047] Figure 4 A process is shown that can be performed to select values of two scan parameters, such as depth and frequency. The process can be performed to select values of more than two scan parameters, such as depth, frequency, and gain. Figure 3 A plurality of ultrasound images 400 acquired and analyzed during the process shown. The plurality of ultrasound images 400 includes a first plurality of images 401. Each image in the first plurality of images 401 is acquired at a different depth value (e.g., a first image 402 acquired at 30 cm, a second image 404 acquired at 17 cm, and a third image 406 acquired at 10 cm) and a same first frequency (e.g., 1.4 MHz). The plurality of ultrasound images 400 also includes a second plurality of images 410. Each image in the second plurality of images 410 is acquired at a different depth value and a same second frequency (e.g., a first image 412 acquired at 30 cm, a second image 404 acquired at 17 cm, and a third image 406 acquired at 10 cm, each of which is acquired at 1.7 MHz). The plurality of ultrasound images 400 also includes a third plurality of images 420. Each image in the third plurality of images 420 is acquired at a different depth value and a same third frequency (e.g., a first image 422 acquired at 30 cm, a second image 424 acquired at 17 cm, and a third image 426 acquired at 10 cm, each of which is acquired at 2.0 MHz). The plurality of ultrasound images 400 also includes a fourth plurality of images 430. Each image in the fourth plurality of images 430 is acquired at a different depth value and a same fourth frequency (e.g., a first image 432 acquired at 30 cm, a second image 434 acquired at 17 cm, and a third image 436 acquired at 10 cm, each of which is acquired at 2.3 MHz).

[0048] Figure 4 Each image shown includes two quality indicators, a predicted IQ rating and a cumulative score. The cumulative score can be based on outputs from one or more of the models shown, such as the depth model, landmark model, global image quality model, and speckle size model. The predicted IQ rating can be obtained from the cumulative score using an empirical threshold determined during a training process. The predicted IQ rating classifies the input ultrasound image or cine loop into three image quality levels, such as 1 for poor / low image quality, 2 for acceptable image quality, and 3 for good / high image quality, while the cumulative score is a continuous number used to select the best acquisition from a set of acquisitions. Figure 3

[0049] The image with the highest combined score (e.g., predicted IQ score and cumulative score) from the plurality of images can be selected, and the selected depth value and the selected frequency value can be the depth value and the frequency value used to acquire the selected image. For example, as shown, image 404 has a predicted IQ score of 3 and a cumulative score of 6, which is higher than the scores of the remaining images. Thus, the selected depth value can be 17 cm when image 404 was acquired at a depth value of 17 cm, and the selected frequency can be 1.4 MHz when image 404 was acquired at a frequency value of 1.4 MHz.

[0050] Figure 5A and Figure 5B A flowchart is shown that illustrates an exemplary method 500 for joint ultrasound imaging parameter selection according to one embodiment. Specifically, the method 500 involves acquiring a plurality of ultrasound images at different parameter values of different scan parameters, and then processing the acquired ultrasound images by a plurality of machine learning models to select the best scan parameter values, which can improve the image quality of the displayed ultrasound images. The method 500 is described with reference to the system and components of Figures 1-2 but it should be understood that the method 500 can be implemented with other systems and components without departing from the scope of the present disclosure. The method 500 can be performed according to instructions stored in a non-transitory memory of a computing device, such as the image processing system 202 of Figure 2

[0051] At 502, an ultrasound image is acquired and displayed on a display device. For example, the ultrasound image can be acquired by the ultrasound imaging system 200 and displayed on the display device 202. Figure 1 ​​The ultrasound probe 106 acquires ultrasound images and displays them to an operator via the display device 118. The images can be acquired and displayed in real-time or near real-time, and can be acquired with default or user-specified scan parameters (e.g., default depth, frequency, etc.). At 504, the method 500 determines whether an indication has been received that a target scan plane is being imaged. The target scan plane can be a scan plane that is specified by a scan protocol or designated by an operator as a target scan plane. For example, ultrasound images can be acquired as part of an ultrasound exam, where certain anatomical features are imaged in certain views / axes in order to diagnose a patient condition, measure aspects of the anatomical features, etc. For example, during a heart exam, one or more target scan planes (also referred to as views) of a patient’s heart can be imaged. The target scan planes can include a four-chamber view, a two-chamber view (which can also be referred to as a short-axis view), and a long-axis view (which can also be referred to as a PLAX view or a three-chamber view). One or more images can be acquired in each scan plane and saved as part of the exam for later analysis by a clinician, such as a cardiologist. When acquiring images for an exam, an ultrasound operator (e.g., an ultrasound exam physician) can move the ultrasound probe until the operator determines that a target scan plane is being imaged, and then the operator can input an input (e.g., via the user interface 115) indicating that the target scan plane is being imaged. In another example, the ultrasound system (e.g., via the image processing system 202) can automatically determine that a target scan plane is being imaged. For example, each acquired ultrasound image (or some frequency of acquired ultrasound images, such as every fifth image) can be input into a detection model configured to automatically detect a current scan plane. An indication can be generated if the current scan plane is a target scan plane.

[0052] If no indication is received that the target scan plane is being imaged, the method 500 returns to 502 to continue acquiring and displaying ultrasound images (e.g., with default or user-set scan parameters). If an indication is received that the target scan plane is being imaged, the method 500 proceeds to 505 to optionally select a parameter selection acquisition protocol in accordance with the target plane. When the parameter selection acquisition protocol is performed, a plurality of images are acquired (e.g., in a sequential manner) with two or more scan parameters (such as depth and frequency) being varied for each image such that images are acquired with all possible combinations of parameter values. In some examples, the relative order of image acquisition can vary based on the target scan plane. For example, images can be acquired in accordance with a first acquisition protocol in which the depth is held constant at a first value while a first set of images are each acquired with a different frequency value, then the depth is changed to a second value and held constant while a second set of images are each acquired with a different frequency value, then the depth is changed to a third value and held constant while a third set of images are each acquired with a different frequency value, and so on. In a second acquisition protocol, the frequency is held constant at a first value while a first set of images are each acquired with a different depth value, then the frequency is changed to a second value and held constant while a second set of images are each acquired with a different depth value, then the frequency is changed to a third value and held constant while a third set of images are each acquired with a different depth value, and so on. Other acquisition protocols are possible, such as a more random distribution in the variation of both depth and frequency, rather than the more ordered combinations described above with respect to the first and second acquisition protocols.

[0053] Which acquisition protocol to perform can be decided based on the target scan plane. For example, the first acquisition protocol can be performed when imaging a four-chamber view of the heart, while the second acquisition protocol can be performed when imaging a PLAX view of the heart. As will be described below with respect to FIGS. 6-8, the user can be provided with the option to select which acquisition protocol to perform, or the system can automatically select which acquisition protocol to perform based on the target scan plane. Figure 7Explained in more detail, the second acquisition protocol can result in all images with the first frequency value being acquired during the same relative phase of the cardiac cycle, all images of the second frequency value being acquired during different phases of the cardiac cycle compared to the first frequency, but all images acquired at the second frequency being acquired at the same relative phase of the cardiac cycle as each other, and so on. When comparing images acquired at different frequency values but at the same depth value, the fact that the images are also acquired at different phases of the cardiac cycle can cause some of the frequency-based changes to become confused or difficult to detect. Changes related to depth, however, are the opposite. By contrast, the first acquisition protocol can result in a first set of images acquired at the same relative cardiac phase with the same depth and different frequencies, a second set of images acquired at the same relative cardiac phase (albeit different from the first set) with the same next depth and different frequencies, and so on. When comparing images acquired at different frequency values but at the same depth value, the fact that the images are acquired at the same relative phase of the cardiac cycle can cause fewer of the frequency-based changes to become confused or difficult to detect. Changes related to depth, however, are the opposite. Thus, the decision of which acquisition protocol to select can be based on whether detecting changes based on depth or frequency is more important when selecting the parameter values, and the type of anatomy being scanned (e.g., whether the anatomy exhibits periodic motion). Once an acquisition protocol is selected, images can be acquired according to that protocol, as explained below. In some examples, however, only one protocol can be available, so there can be no selection process.

[0054] At 506, a first set of images is acquired, with each image acquired at a first parameter value of a first set of parameter values of a first scan parameter and a different parameter value of a second set of parameter values of a second scan parameter. For example, the first scan parameter can be depth and the first set of parameter values can be the different depth values described above (e.g., 10 cm, 17 cm, and 30 cm). Thus, the first set of images can be acquired at one of the depth values (e.g., 10 cm). The second scan parameter can be frequency and the second set of parameter values can be the different frequencies described above (e.g., 1.4 MHz, 1.7 MHz, 2 MHz, and 2.3 MHz). Thus, each of the first set of images can be acquired at a different parameter value of the second set of parameter values (e.g., one image at 1.4 MHz, one image at 1.7 MHz, one image at 2 MHz, and one image at 2.3 MHz). The first set of images can include four images or more than four images (if more than four frequency values are to be selected from). Any other scan parameters that can have been optimized during acquisition of the first set of images (e.g., gain, etc.) can remain constant during acquisition of the first set of images. Further, during acquisition of the first set of images, at least in some examples, the acquired images can be displayed on a display device at a frame rate at which the images are acquired. In some examples, a set of cine loops, or a mix of images and cine loops, can be acquired. When cine loops are acquired, different cine loops can be acquired, for example, at the same depth value and different frequency values as explained above.

[0055] At 508, a second set of images (and / or cine loops) is acquired, each image acquired at a second parameter value of the first set of parameter values of the first scan parameter and a different parameter value of the second set of parameter values of the second scan parameter. For example, each of the second set of images can be acquired at one of the depth values different from the first values described above (e.g., 17 cm). Each of the second set of images can be acquired at a different parameter value of the second set of parameter values (e.g., one image at 1.4 MHz, one image at 1.7 MHz, one image at 2 MHz, and one image at 2.3 MHz). The second set of images can include four images or more than four images (if more than four frequency values are to be selected from). Any other scan parameters that can have been optimized during acquisition of the second set of images (e.g., gain, etc.) can remain constant during acquisition of the second set of images. Further, during acquisition of the second set of images, at least in some examples, the acquired images can be displayed on a display device at a frame rate at which the images are acquired.

[0056] At 510, the method 500 includes determining whether a set of images has been acquired for each parameter value of the first set of parameter values. For example, the method can include determining, after acquiring the first set of images and / or the second set of images, how many parameter values of the first set of parameter values and whether a corresponding set of images has been acquired for each possible parameter value of the first set of parameter values (e.g., with one image acquired at each different parameter value of the second set of parameter values). If not, for example, if additional images are to be acquired to complete a parameter selection acquisition protocol in which images are acquired at each different possible combination of parameter values, the method 500 proceeds to 512 to acquire one or more respective sets of images (and / or cine loops) for each remaining parameter value of the first set of parameter values, and then the method 500 returns to 510 to determine whether a set of images has been acquired for each parameter value of the first set of parameter values (e.g., if the acquisition protocol is complete).

[0057] If it is determined that a set of images has been acquired for each parameter value of the first set of parameter values (e.g., the acquisition protocol is complete), the method 500 proceeds to 514 to determine a quality metric for each image of each set of images (e.g., each image acquired according to the parameter selection acquisition protocol) (and / or each cine loop). The quality metric can be determined according to a plurality of models, as indicated at 516. For example, as previously explained, a first quality metric can be determined by a depth model such as the depth model 209 of Figure 2 The first quality metric can be inputted as an input to the depth model, and the depth model can output a respective first quality metric for each inputted image. In some examples, the model can be selected based on the target scan plane. For example, a first depth model can be selected when the target scan plane is a four-chamber view, and a second depth model can be selected when the target scan plane is a two-chamber view. Further, a second quality metric can be determined according to one or more frequency-related models such as the one or more frequency models 211 of Figure 2 The second quality metric can be inputted as an input to the speckle model, the landmark detection model, and / or the global image quality model, and these models can output respective sub-metrics for each inputted image. The respective sub-metrics can be combined (e.g., added or averaged) to generate the second quality metric. In some examples, the model can be selected based on the target scan plane. For example, a first landmark model can be selected when the target scan plane is a four-chamber view, and a second landmark model can be selected when the target scan plane is a two-chamber view. The first and second metrics for a given image can be combined to result in a quality metric for that image.

[0058] At 518, the image with the highest quality metric is selected. For example, referring back to Figure 4 The second image 404 is assigned an image metric (e.g., a total score) of 9, which is according to the above with respect to Figure 4The highest quality metric of all images acquired by the acquisition protocol described. Therefore, the second image 404 may be selected. In examples where more than one image has the highest quality metric, the image acquired with the highest frequency may be selected. For example, Figure 4 In the example presented, if second image 414 had a quality metric of 9 (eg, instead of 8.9 as shown), second image 414 may be selected instead of second image 404 because second image 414 was acquired at 1.7 MHz instead of 1.4 MHz.

[0059] At 520, if Figure 5B As shown, a first parameter value for a first scan parameter at which the selected image was acquired is identified and set as the selected parameter value for the first scan parameter. For example, if the first scan parameter is depth and the selected image was acquired at a depth of 17 cm, a depth value of 17 cm may be identified as the selected parameter value. At 522, a second parameter value for a second scan parameter at which the selected image was acquired is identified and set as the selected parameter value for the second scan parameter. For example, if the second scan parameter is frequency and the selected image was acquired at a frequency of 1.4 MHz, a frequency value of 1.4 MHz may be identified as the selected parameter value. In some examples, selecting both the first parameter value and the second parameter value may be performed only after all images determined by the parameter selection acquisition protocol have been acquired.

[0060] At 523, method 500 optionally includes setting target acquisition post-processing parameters, as described below with respect to Figure 8 This is described in more detail. Briefly, once acquisition scan parameter values ​​have been selected as described above, target values ​​for one or more post-acquisition processing parameters can be selected based on the image quality of one or more sets of adjusted images. For example, an image can be replicated, with different post-acquisition processing parameter values ​​applied to each replicate to create a set of adjusted images. The image with the highest image quality from the set of adjusted images can be selected, and the parameter values ​​for that image can be applied to subsequent images. Exemplary post-acquisition processing parameters include filtering parameters (e.g., bandwidth, center frequency), image contrast, image gain, etc.

[0061] At 524, one or more ultrasound images are acquired with the selected value of the first scan parameter and the selected value of the second scan parameter. Thus, once the scan parameter values have been selected for the target scan plane based on the determined image quality metrics as described above, the selected scan parameter values can be set and any additional images acquired by the ultrasound probe can be acquired with the selected scan parameter values set. This can include setting the transmit depth of the ultrasound probe to the selected depth value and setting the transmit frequency of the ultrasound probe to the selected frequency. In some examples, the selected scan parameter values for the target scan plane can be saved in memory. Then, if the operator moves the ultrasound probe such that the target scan plane is not imaged, but then moves the ultrasound probe back such that the target scan plane is imaged again, the previously determined selected scan parameter values for that scan plane can be automatically applied. Additionally, if target post-acquisition processing parameters are set (e.g., in accordance with the method of Figure 8 ), the one or more ultrasound images acquired at 524 can be processed in accordance with the target post-acquisition processing parameters, e.g., the received ultrasound data can be filtered in accordance with the parameter values of the filter (e.g., bandwidth and center frequency) that were determined in accordance with the method of Figure 8 .

[0062] At 526, the method 500 determines whether the current exam includes more target planes. Whether the current exam includes more target planes can be determined based on user input. For example, the operator can input user input indicating that a new scan plane is being imaged, that a new scan plane will be imaged, or that the exam is complete. In other examples, determining whether the exam includes more target planes can be performed automatically based on the system determining that a different scan plane is being imaged or that the scan has terminated. If the exam does not include more target scan planes, e.g., if the current exam is complete and terminates imaging, the method 500 proceeds to 528 to display the acquired images, quality metrics, and selected parameter settings, and then the method 500 returns. It will be appreciated that the acquired images can be displayed at 524 and / or at other points during the method 500. Further, the selected parameter settings can be displayed at 524 to allow the operator to view and confirm the parameter settings. The quality metrics can also be displayed at other points in time, such as at 524. Additionally, the images acquired at 524 can be archived upon request by the operator.

[0063] If the examination does include more target scan planes, the method 500 proceeds to 530 to determine whether an indication has been received that the next target plane is being imaged, similar to the determination made at 504 and described above. If the indication has not been received, the method 500 proceeds to 532 to continue acquiring images with the selected values of the first and second scan parameters (e.g., as explained above with respect to 524), and then the method 500 returns to 530 to continue determining whether the indication has been received. If the indication has been received, the method 500 proceeds to 534 and optionally limits the parameter values of one or both of the first and second scan parameters. For example, as described above, the first scan parameter can have three possible parameter values, and the second scan parameter can have four possible scan values. However, once the selected parameter values have been determined for a given scan plane, those selected parameter values can be applied to the next target plane, limiting the available values that can be optimized. For example, if the first target plane is a four-chamber view, and the next target plane is a two-chamber view, one or both of the selected values can be used to acquire an image in the two-chamber view. If one of the selected values is used but not the other (e.g., a selected depth is used), the selection of the selected value of the other scan parameter (e.g., frequency) can be re-performed for the next target plane. However, when switching from a four-chamber view to a PLAX view, for example, both parameter values can be re-determined and thus 534 can not be performed.

[0064] At 536, 505-524 can be repeated for the next target plane. For example, a plurality of images of the next target scan plane can be acquired, each acquired with a different combination of parameter values of the first and second scan parameters, a quality metric can be determined for each image, and the image with the highest quality metric can be identified. The acquisition settings used to acquire the selected image (e.g., the depth value and the frequency value) can be set as the selected values of the first and second scan parameters, and then one or more additional images of the next target scan plane can be acquired with the selected values of the first and second scan parameters. This process can be repeated for all of the additional target scan planes until the examination is complete.

[0065] Although method 500 is described above with respect to jointly varying depth and frequency to determine target depth and frequency values ​​that will result in high quality images, other acquisition scan parameters may be varied according to the above methods without departing from the scope of the present disclosure. For example, the beamforming strategy and frequency may be varied jointly. The beamforming strategy may include the type of beamforming employed, e.g., the intensity / type of ACE processing. Exemplary beamforming strategies (which may be considered different "parameter values" for the beamforming strategy) may include delayed-sum, coherent plane wave recombination, and diverging beams. In order to select a target beamforming strategy and frequency, a set of images may be acquired, each image being acquired with a different combination of beamforming strategy and frequency (e.g., a first image being acquired with a first beamforming strategy and a first frequency, a second image being acquired with a second beamforming strategy and a first frequency, a third image being acquired with a first beamforming strategy and a second frequency, a fourth image being acquired with a second beamforming strategy and a second frequency, and so on). Each image in the set of images may be assigned a quality metric, as described above. For example, each image can be input to a speckle model, a landmark detection model, and / or a global image quality model, and these models can output a corresponding sub-metric for each input image. The corresponding sub-metrics can be combined (e.g., added or averaged) to generate a quality metric for each image. The image with the highest quality metric can be selected, and the beamforming strategy and frequency used to acquire the selected image can be set for subsequent image acquisitions. In examples where depth is not a scan parameter to be varied and selected, the depth model explained above can be omitted from the quality metric determination.

[0066] Figure 6 It is shown that it can be applied to acquire 12 images (such as Figure 4 Table 600 shows different combinations of scan parameter values ​​(e.g., depth and frequency) for a set of 12 images (shown in Figure 6). Table 600 is an example of an acquisition protocol that determines the order in which images are acquired (e.g., the order in which combinations of acquisition parameter settings will be used). The ultrasound system may have a frame rate of 10 Hz and, therefore, may acquire an image every tenth of a second. As can be understood from Table 600, images may be acquired such that a first set of images is acquired at the same first depth (from a set of three depths) and different frequencies (from a set of four frequencies), a second set of images is acquired at the same second depth (from the set of three depths) and different frequencies (from the set of four frequencies), and a third set of images is acquired at the same third depth (from the set of three depths) and different frequencies (from the set of four frequencies). Thus, for the first four images, the depth may remain constant while the frequency is varied. For the next four images, the depth may remain constant (at a different depth than the first four images) and the frequency may be varied, and for the final four images, the depth may remain constant at a different depth and the frequency may be varied.

[0067] As understood by table 600, an image will be acquired for each possible combination of parameter values for a first set of parameter values having three values and a second set of parameter values having four values. Table 600 can be stored in a memory (e.g., memory 120) of a computing device, and images can be acquired in accordance with table 600 during parameter selection as described above with respect to FIGS. 1-4. Figure 1 Figure 5A Figure 5B

[0068] Because motion of the imaged anatomical feature can contribute to fluctuations in image quality, it can be desirable to obtain images described herein (e.g., for determining optimal scan parameters) during periods in which no motion occurs, or during periods in which motion between images is comparable. When imaging the heart, it can be challenging to obtain images without motion or with comparable motion given the movement of the heart during the cardiac cycle. For example, for a patient with a heart rate of 60 beats per minute, the cardiac cycle can last one second, which is approximately the same amount of time for acquiring all 12 images in accordance with table 600. Thus, the decision of whether the frequency remains constant for a certain duration while the depth changes (as shown in FIG. 6A), or whether the depth remains constant while the frequency changes (as shown in FIG. 6B) can depend on the anatomical structure being imaged (e.g., whether the heart is being imaged, and if so, which view of the heart). Figure 6 Figure 4 Figure 6

[0069] Figure 7 Two exemplary plots of ultrasound frequency as a function of time during a parameter selection acquisition protocol in which multiple images are acquired are shown. Figure 7 Each plot in FIG. 6 also shows an approximation of the cardiac cycle time aligned with the frequency. The first plot 702 shows the frequency as a function of time relative to the cardiac cycle when, for each frequency, the frequency remains constant while the depth varies. This arrangement of scan parameter variability during acquisition can result in all images having a first frequency (fl) being obtained during a first phase of the cardiac cycle (e.g., systole onset), all images having a second frequency (f2) being obtained during a second phase of the cardiac cycle (e.g., systole end), all images having a third frequency (f3) being obtained during a third phase of the cardiac cycle (e.g., diastole onset), and all images having a fourth frequency (f4) being obtained during a fourth phase of the cardiac cycle (e.g., diastole end).

[0070] ​​​​​​In contrast, the second graph 704 illustrates the frequency over time with respect to the cardiac cycle when, for each depth, the depth is held constant while the frequency varies. This arrangement of scan parameter variability during acquisition can result in an image with frequency fl being obtained at each cardiac phase (e.g., one image during the beginning of systole, one image during the end of systole, one image during the beginning of diastole, and one image during the end of diastole). The remaining frequencies can follow a similar distribution (e.g., one image per frequency at each phase of the cardiac cycle). The distribution illustrated in graph 704 can be advantageous relative to the distribution illustrated in graph 702 when motion affects the frequency-based image quality detection to a higher degree than the depth-based image quality detection. When comparing images of different frequencies but the same depth to one another to determine which image has the highest image quality, a more reliable determination can be made when all of the images being compared are acquired in the same relative phase of the cardiac cycle. For example, as illustrated in graph 704, the first four frames of each image are each acquired at a different frequency, but occur during the same phase of the cardiac cycle.

[0071] Turning now to Figure 8 , a method 800 for determining post-acquisition processing parameters for ultrasound images is presented. The method 800 can be performed as part of the method 500, e.g., once the acquisition parameter values for the target plane have been determined. In other examples, the method 800 can be performed independently of the method 500, e.g., in response to a user request to set the post-acquisition processing parameter values. The method 800 can be performed in accordance with instructions stored in a non-transitory memory of a computing device, such as the image processing system 202 of FIG. 2. Figure 2

[0072] At 802, ultrasound information for a single image is obtained. The ultrasound information can be acquired by an ultrasound probe in response to performance of the method 800, or the ultrasound information can be retrieved from memory. In one non-limiting example, the ultrasound information can be sufficient to generate one image, and can be obtained by targeting acquisition scan parameters (e.g., at a target depth, a target frequency, etc.) as discussed above.

[0073] ​At 804, different parameter values of the first post-acquisition parameter are applied to the obtained ultrasound information to generate a first set of adjusted images. For example, the first post-acquisition parameter can be a filter center frequency, and the different parameter values can be different center frequencies (e.g., 3.2 MHz, 3.4 MHz, and 3.6 MHz, or different multiples of the transmit frequency, such as the transmit frequency, twice the transmit frequency, and three times the transmit frequency). In another example, the first post-acquisition parameter can be a filter bandwidth, and the different parameter values can be different bandwidths (e.g., 1 MHz, 1.2 MHz, and 1.4 MHz). Each different parameter value can be applied to the information to generate an image for each parameter value. For example, when the first post-acquisition parameter is a filter center frequency, the first set of adjusted images can include a first image generated with a center frequency of 3.2 MHz, a second image generated with a center frequency of 3.4 MHz, and a third image generated with a center frequency of 3.6 MHz. The same ultrasound information can be used to generate each image in the first set of adjusted images. Any other post-acquisition parameter can be held constant at a default value or a commanded value.

[0074] At 806, a quality metric is determined for each image in the first set of adjusted images. The quality metric for each image can be determined by inputting each image as input to one or more image quality models, as described above with respect to Figure 2 、 Figure 3 、 Figure 5A and Figure 5B For example, each image can be input as input to a global image quality model, a landmark detection model, and a speckle model, and each model can output a respective quality sub-metric that can be summed or averaged to arrive at an overall image quality metric for each image.

[0075] At 808, the image in the first set of adjusted images with the highest image quality metric is selected. If two or more images have the same highest image quality metric, additional metrics can be used to select from the two or more images, such as the global image quality model sub-metric. At 810, the first post-acquisition parameter is set to the parameter value of the selected image. For example, if the selected image is generated with a filter center frequency of 3.2 MHz, the filter center frequency can be set to 3.2 MHz.

[0076] At 812, the above process can be repeated for any additional post-acquisition parameters. For example, after selecting the first post-acquisition parameter value, the ultrasound information can be used again to generate duplicate images, each of which has a different parameter value for a second post-acquisition parameter (such as a filter bandwidth) to form a second set of adjusted images. When the first post-acquisition parameter has been set, the images in the second set of adjusted images can be generated using the set parameter value for the first post-acquisition parameter. An image quality metric can be determined for each image in the second set of adjusted images, and the image with the highest image quality metric can be selected. The parameter value for the second post-acquisition parameter of the image with the highest image quality metric can be set as the parameter value for the second post-acquisition parameter. At 814, the set parameter value for each post-acquisition parameter is applied to any subsequent images, such as the current view plane. Method 800 then ends.

[0077] While method 800 is described above as a sequential process including selecting parameter values ​​for two or more post-acquisition parameters, a combined process may alternatively be used. In the combined process, a set of replicated images may be generated, each image having a different combination of parameter values ​​for two or more post-acquisition parameters. The image quality of each image may be determined as described above, and the image with the highest image quality metric may be selected. Parameter values ​​for the two or more post-acquisition parameters used to generate the selected image may be selected and set as parameter values ​​for subsequent image processing.

[0078] The image acquisition process used to acquire the ultrasound images described herein may be performed according to a suitable scanning sequence. Figure 9 Graph 900 is shown, which illustrates a first exemplary scan sequence that can be performed to acquire ultrasound information that can be used to generate a single image. Graph 900 illustrates a sector scan, but other scan geometries are possible. Each line in graph 900 represents a transmit direction, and, for example, transmits are fired sequentially from left to right. When a transmit parameter (such as frequency) is changed, the transmits can all be fired at the same parameter value (e.g., at the same frequency P1) to generate a first image. The frequency can then be adjusted to a second frequency, and the transmits can all be fired sequentially at the second frequency to generate a second image.

[0079] When exploring target parameter values, Figure 9 The scanning sequence will result in subsequent images being acquired with different parameter values. However, subsequent shots within an image can be acquired with different parameters. The scanning process for acquisition will include firing N times in the same firing direction with N different parameter values ​​to record data in that direction, and then moving on to the next firing direction. Figure 10This scanning procedure is shown in FIG. 13, which shows a graph 1300 that illustrates a second example scan sequence that can be performed to acquire ultrasound information that can be used to generate a plurality of images. The graph 1300 shows a sector scan, but other scan geometries are possible. Each solid line in the graph 1300 represents a firing direction and a first parameter value, while each dashed line represents a different parameter value (firing in the firing direction to the left of the solid line), and the firings are fired, for example, sequentially from left to right. Instead of including only one firing parameter value, the graph 1300 includes four firing parameter values P1-P4. For example, the firing parameter can be frequency, and each parameter value can be a different frequency.

[0080] During image acquisition, the firings can be fired sequentially, but for each firing direction, a firing can be fired for each parameter value, after which the movement continues to the next firing direction. For example, for a first firing direction, a firing can be fired at P1, a firing can be fired at P2, a firing can be fired at P3, and a firing can be fired at P4 (although the different solid / dashed lines are placed next to each other for illustrative purposes, it should be understood that each firing of P1-P4 for the first firing direction would be fired in the same firing direction). The firing direction can be updated to a second firing direction, and a set of firings can be fired in the second firing direction, one for each parameter value. This process can be repeated until all firing directions have been fired with all parameter values. A first image can be generated from the information acquired while firing at the first parameter value, a second image can be generated from the information acquired while firing at the second parameter value, a third image can be generated from the information acquired while firing at the third parameter value, and a fourth image can be generated from the information acquired while firing at the fourth parameter value. This scan sequence for parameter exploration will fire several times in each direction, after which the movement continues to the next firing direction. This can result in a longer time needed to acquire all directions of the image, but will result in very low lag between the different parameters to be compared for image quality.

[0081] In at least some examples, as noted above with respect to Figure 5A and Figure 5B the scan sequence shown in FIG. 12 can be performed during image acquisition for parameter selection. In other examples, the scan sequence shown in FIG. 13 can be performed during image acquisition for parameter selection. Figure 10 Figure 9 A technical effect of the techniques described above with respect to

[0082] A technical effect of jointly selecting scan parameter values includes increased image quality and reduced operator workflow demands. Another technical effect is more consistent image quality across multiple exams.

[0083] ​In another representation, a system includes an ultrasound probe, a memory storing instructions, and a processor communicatively coupled to the memory and, when executing the instructions, configured to: process ultrasound information obtained by the ultrasound probe into a first set of replica images, each replica image processed according to a different acquisition post-processing parameter value of a plurality of acquisition post-processing parameter values of a first acquisition post-processing parameter; determine an image quality metric for each replica image of the first set of replica images; select a replica image having a highest image quality metric; and process additional acquired ultrasound information according to the acquisition post-processing parameter value used to process the ultrasound information into the selected replica image. In one example, each replica image of the first set of replica images is processed from the same ultrasound information, such that the replica images are identical except for the different acquisition post-processing parameter values used to create the replica images. In one example, the processor is configured, after selecting the replica image having the highest image quality metric, to: process the ultrasound information into a second set of replica images, each replica image of the second set of replica images processed according to a different acquisition post-processing parameter value of a plurality of acquisition post-processing parameter values of a second acquisition post-processing parameter; determine an image quality metric for each replica image of the second set of replica images; select a replica image of the second set of replica images having a highest image quality metric; and process additional acquired ultrasound information according to the acquisition post-processing parameter value of the second acquisition post-processing parameter used to process the ultrasound information into the selected replica image. In one example, each replica image of the first set of replica images is processed according to different parameter values of the second acquisition post-processing parameter, and the additional acquired ultrasound information is processed according to the parameter value of the second acquisition post-processing parameter used to process the ultrasound information into the selected replica image. In one example, the ultrasound information can be acquired by the ultrasound probe according to the first and second target scan parameter values selected by the joint process described above with respect to Figure 5A and Figure 5B the ultrasound information is acquired with the first and second target scan parameter values selected by the joint process.

[0084] When introducing elements of various embodiments of the present disclosure, the articles "a," "an," and "the" are intended to mean that there are one or more of the elements. The terms "first," "second," and the like do not denote any order, quantity, or importance, but are used to distinguish one element from another. The terms "comprises," "comprising," "includes," "including," and the like are intended to be inclusive and mean that there can be additional elements other than the listed elements. As used herein, the term "connected to" or "coupled to" means that one object (e.g., material, element, structure, member, etc.) can be directly or indirectly connected or coupled to another object, whether mechanically, electrically, or otherwise, and whether or not there is an intervening object or objects between the one object and another object. Also, it will be understood that a reference to "one embodiment" or "an embodiment" of the present disclosure is not intended to be interpreted as excluding the existence of additional embodiments that also incorporate the recited features.

[0085] In addition to any prior indicated modifications, many other variations and substitutions of the described arrangements can be designed by those skilled in the art without departing from the spirit and scope of the present description, and the appended claims are intended to cover such modifications and arrangements. Thus, although the information has been described above with reference to specific and particular aspects that are at present considered to be the most practical and preferred, it will be apparent to those skilled in the art that numerous modifications can be made without departing from the principles and concepts set forth herein, including, but not limited to, form, function, manner of operation, and use. Equally, as used herein, the examples and embodiments are intended to be illustrative in all respects, and are not intended to be limiting in any way.

Claims

1. A method for joint scan parameter selection, comprising: acquiring a plurality of ultrasound images of an anatomical region, each ultrasound image acquired at a different combination of parameter values of a first scan parameter and a second scan parameter, the different combinations achieved through a joint process that encompasses all possible combinations, wherein the acquisition of the different combinations of parameter values of the first scan parameter and the second scan parameter is achieved through an acquisition protocol based on a target scan plane; selecting a first parameter value of the first scan parameter and a second parameter value of the second scan parameter based on an image quality of each image; and acquiring one or more additional ultrasound images at the first parameter value and the second parameter value.

2. The method of claim 1, wherein the parameter values comprise a first set of parameter values for the first scan parameter and a second set of parameter values for the second scan parameter, and wherein acquiring the plurality of images comprises: acquiring a first set of images, each of the first set of images acquired at a different parameter value of the first set of parameter values and the same first parameter value of the second set of parameter values; and acquiring a second set of images, each of the second set of images acquired at a different parameter value of the first set of parameter values and the same second parameter value of the second set of parameter values.

3. The method of claim 2, wherein acquiring the plurality of images further comprises acquiring one or more additional sets of images, each image in each respective additional set of images acquired at a different parameter value of the first set of parameter values and the same respective subsequent parameter value of the second set of parameter values.

4. The method of claim 1, wherein selecting the first parameter value comprises: determining a respective image quality metric for each image using a plurality of models; selecting an image of the plurality of images having a highest image quality metric; and setting the first parameter value to the parameter value of the first scan parameter at which the selected image was acquired.

5. The method of claim 4, wherein selecting the first parameter value comprises setting the second parameter value to the parameter value of the second scan parameter at which the selected image was acquired.

6. The method of claim 1, wherein the first scan parameter comprises depth and the second scan parameter comprises frequency.

7. The method of claim 1, wherein the first scan parameter comprises frequency and the second scan parameter comprises a beamforming strategy.

8. A method for an ultrasound system, comprising: in response to determining that a target scan plane of an anatomical region is currently being imaged by the ultrasound system, jointly selecting a target value of a first scan parameter and a target value of a second scan parameter based on a respective image quality metric for each image of a plurality of images of the anatomical region in the target scan plane, wherein the jointly selecting comprises acquiring each image of the plurality of images according to an acquisition protocol, the acquisition protocol determined based on the target scan plane, the acquisition protocol causing each image of the plurality of images to be acquired at a different respective combination of a value of the first scan parameter and a value of the second scan parameter; and applying the selected target value of the first scan parameter and the selected target value of the second scan parameter to one or more additional images of the anatomical region, wherein the selection of the target value of the first scan parameter and the target value of the second scan parameter is only performed after all images determined by the acquisition protocol have been acquired.

9. The method of claim 8, wherein jointly selecting the target value of the first scan parameter and the target value of the second scan parameter comprises acquiring each of the plurality of images with different combinations of values of the first scan parameter and the second scan parameter, each value of the first scan parameter selected from a set of values of the first scan parameter and each value of the second scan parameter selected from a set of values of the second scan parameter, and wherein applying the selected target value of the first scan parameter and the selected target value of the second scan parameter to one or more additional images of the anatomical region comprises acquiring the one or more additional images of the anatomical region with the selected target value of the first scan parameter and the selected target value of the second scan parameter.

10. The method of claim 9, wherein acquiring each of the plurality of images with different combinations of values of the first scan parameter and the second scan parameter comprises: acquiring a set of first images, each first image acquired with a same first value of the first scan parameter and a different value of the second scan parameter, and acquiring a set of second images, each second image acquired with a same second value of the first scan parameter and a different value of the second scan parameter.

11. The method of claim 10, wherein jointly selecting the target value of the first scan parameter and the target value of the second scan parameter based on the respective image quality metric of each of the plurality of images further comprises: selecting an image with a highest image quality metric from the plurality of images, setting the target value of the first scan parameter to the value of the first scan parameter with which the selected image was acquired, and setting the target value of the second scan parameter to the value of the second scan parameter with which the selected image was acquired.

12. The method of claim 11, wherein determining the respective image quality metric of each image comprises determining a respective image quality metric of each image via one or more image quality models.

13. The method of claim 12, wherein determining the respective image quality metric of each image via one or more image quality models comprises, for each image: determining a first sub-metric via a global image quality model; determining a second sub-metric via a landmark model; determining a third sub-metric via a speckle size model; and generating the image quality metric of the image by summing the first sub-metric, the second sub-metric, and the third sub-metric.

14. The method of claim 8, wherein the first scan parameter comprises a beamforming strategy and the second scan parameter comprises a frequency. ​ 15. The method of claim 8, wherein the first scan parameter is a first post-acquisition processing parameter and the second scan parameter is a second post-acquisition processing parameter, wherein the method further comprises generating the plurality of images by processing ultrasound information to generate a set of replicated images, each replicated image of the set of replicated images being processed according to a different combination of parameter values of the first post-acquisition processing parameter and the second post-acquisition processing parameter; and wherein applying the selected target value of the first scan parameter and the selected target value of the second scan parameter to one or more additional images of the anatomical region comprises processing subsequently acquired ultrasound image information according to the selected target value of the first post-acquisition processing parameter and the selected target value of the second post-acquisition processing parameter.

16. The method of claim 8, wherein the first scan parameter comprises a depth and the second scan parameter comprises a frequency, wherein the target scan plane is a first target scan plane, the target value of the first scan parameter is a first depth value, the target value of the second scan parameter is a first frequency value, and further comprising, in response to determining that a second target scan plane of the anatomical region is currently being imaged by the ultrasound system: selecting a second depth value and a second frequency value based on a respective image quality metric of each image of a second plurality of sequentially acquired images of the anatomical region in the second target scan plane and / or based on the first depth value and the first frequency value; and acquiring one or more additional images of the anatomical region at the second depth value and the second frequency value.

17. A system for joint scan parameter selection, comprising: an ultrasound probe; a memory, the memory storing instructions; and a processor, the processor communicatively coupled to the memory and, when executing the instructions, configured to: in response to determining that a target scan plane of an anatomical region is currently being imaged by the ultrasound probe, jointly select a target value of a first scan parameter and a target value of a second scan parameter based on a respective image quality metric of each image of a plurality of images of the anatomical region in the target scan plane, wherein the jointly selecting comprises acquiring each image of the plurality of images according to an acquisition protocol, the acquisition protocol being determined based on the target scan plane, the acquisition protocol causing each image of the plurality of images to be acquired at a different respective combination of values of the first scan parameter and the second scan parameter; and apply the selected target value of the first scan parameter and the selected target value of the second scan parameter to one or more additional images of the anatomical region, wherein the selection of the target value of the first scan parameter and the target value of the second scan parameter is performed only after all images determined by the acquisition protocol have been acquired.

18. The system of claim 17, wherein the memory stores one or more neural networks, and wherein the processor, when executing the instructions, is configured to input each image to the one or more neural networks to determine the respective image quality metric for each image.

19. The system of claim 17, wherein the first scan parameter comprises a beamforming strategy or a depth and the second scan parameter comprises a frequency.

20. The system of claim 17, wherein the first scan parameter comprises an acquisition post-filter bandwidth and the second scan parameter comprises an acquisition post-filter center frequency.

Images