Ultrasound diagnostic equipment and ultrasound diagnostic method

Abstract

To reduce the time required for scanning lesions and the skill level required to perform the scan. [Solution] The ultrasound diagnostic apparatus according to the embodiment comprises an acquisition unit, a generation unit, and a determination unit. The acquisition unit receives ultrasound data from an ultrasound transducer that moves across the area of ​​a patient or other subject in order to identify the location of a target feature. The generation unit generates a sequence of multiple images from the ultrasound data, each image showing a view of the subject at each position of the ultrasound transducer. The determination unit determines the differences between the images in the sequence and uses the differences to determine the portion of the sequence corresponding to the sweep motion by the ultrasound transducer and the portion of the sequence corresponding to the measurement motion by the ultrasound transducer.

Description

Technical Field

[0001] The embodiments described in this specification relate to an ultrasonic diagnostic apparatus and an ultrasonic diagnostic method.

Background Art

[0002] Liver lesion scanning may be performed by an ultrasound operator sweeping the abdomen of a subject to indicate the size and location of the lesion. To determine the location of the lesion, the operator searches for and identifies relevant surrounding anatomical landmarks of the liver, which requires considerable skill and time.

[0003] Liver lesion scanning is performed to detect metastases not previously known and / or to track metastases previously found as part of the treatment. Liver lesion scanning is typically performed visually by an ultrasound (UL) operator using 2D imaging technology (B-mode) from left to right or right to left across the abdomen of the subject using an ultrasound transducer.

[0004] FIG. 1 is a diagram showing an example of the hepatic segments of the Couinaud classification. The location where a lesion is detected during scanning is typically measured using an elliptical model that requires two measurements per lesion, and the location is recorded. The location includes one of the eight possible hepatic segments defined by the Couinaud classification. These segments are defined as areas of the liver surrounded by the major ducts of the liver. FIG. 1 shows three major ducts of the liver (right hepatic vein 12, middle hepatic vein 14, left hepatic vein 16).

[0005] Figure 2 is a schematic diagram illustrating an example of steps performed by a human operator in a clinical setting to scan or sweep the liver and determine the presence and location of liver lesions, according to known techniques. In step 202, the ultrasound operator linearly scans the subject's abdomen with the ultrasound transducer 214. Once the operator visually identifies a lesion, in step 204, the human operator manually positions the ultrasound transducer 214 at the center of the lesion. In step 206, the human operator rotates the ultrasound transducer 214 and may measure the dimensions of the lesion, for example, by measuring the maximum dimension of the lesion that becomes visible during the rotation of the ultrasound transducer. The human operator visually determines and records the size of the lesion. Then, in step 208, the operator searches for anatomical structures near the lesion to identify the liver segment containing the lesion. This can be a particularly time-consuming task and requires skill and knowledge from the human operator. Once sufficient anatomical structures have been identified by the human operator, the liver segment is assigned to the lesion, and then in step 210, the transducer is returned to its final position during the sweep phase. This may be the location where the lesion was first detected. Then, in step 212, the transducer sweep is resumed.

[0006] The process described above requires skilled manipulation of the position and orientation of the ultrasound transducer 214 to scan the vicinity of the lesion. These actions, in addition to sweeping the transducer during scanning, are time-consuming, physically stressful, and potentially cause musculoskeletal damage to the human operator. The process also requires cognitive skills to identify specific areas containing liver segments and lesions during the scan.

[0007] As mentioned above, identifying the correct liver segment, returning to the lesion site, and continuing the sweep requires a tremendous amount of time, effort, and skill from the human operator. Very long diagnostic processes reduce the throughput potential of the hospital or facility. Generally, the human skill required to identify liver segments is higher than the skill required to identify and measure lesions. [Prior art documents] [Patent Documents]

[0008] [Patent Document 1] U.S. Patent Application Publication No. 2020 / 0051257 [Overview of the project] [Problems that the invention aims to solve]

[0009] One of the problems that the embodiments disclosed herein and in the drawings aim to solve is to reduce the time required for scanning lesions and the skill level required to perform such scanning. However, the problems that the embodiments disclosed herein and in the drawings aim to solve are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]

[0010] The ultrasound diagnostic apparatus according to the embodiment comprises an acquisition unit, a generation unit, and a determination unit. The acquisition unit receives ultrasound data from an ultrasound transducer that moves across the area of ​​a patient or other subject in order to identify the location of a target feature. The generation unit generates a sequence of multiple images from the ultrasound data, each image representing a view of the subject at each position of the ultrasound transducer. The determination unit determines the differences between the images in the sequence and uses these differences to determine which parts of the sequence correspond to the sweep motion by the ultrasound transducer and which parts of the sequence correspond to the measurement motion by the ultrasound transducer.

[0011] Herein, embodiments and examples are described as non-limiting illustrations and are shown in the following figures. [Brief explanation of the drawing]

[0012] [Figure 1]Figure 1 shows an example of liver segments according to the Couinaud classification. [Figure 2] Figure 2 is a schematic diagram illustrating an example of the steps taken to scan the liver in a clinical setting. [Figure 3] Figure 3 is a schematic diagram showing an example of an apparatus according to the embodiment. [Figure 4] Figure 4 is a flowchart showing an example of a method for processing ultrasound images according to the embodiment. [Figure 5] Figure 5 shows five diagrams illustrating an example of the image obtained from a scan according to the embodiment and the effect of the transducer operation. [Figure 6] Figure 6 is a flowchart showing an example of an ultrasonic image processing method according to the embodiment. [Modes for carrying out the invention]

[0013] According to one embodiment, an ultrasound diagnostic apparatus is provided comprising an acquisition unit, a generation unit, and a determination unit. The acquisition unit is To locate target features, ultrasound data is received from an ultrasound transducer that moves across the area of ​​the patient or other subject. The generating unit is From the ultrasound data, a sequence of multiple images is generated in which each image shows a view of the subject at each position of the transducer. The aforementioned determination unit, The difference between images in the sequence is determined, and the difference is used to determine the portion of the sequence corresponding to the sweep motion by the transducer and the portion of the sequence corresponding to the measurement motion by the transducer. The measurement motion may be for determining the location of the target feature and / or measuring the target feature.

[0014] According to one embodiment, an ultrasound diagnostic method is provided. The ultrasound diagnostic method is To identify the location of a target feature, receiving ultrasonic data from an ultrasonic transducer that is moving across or has moved across an area of a patient or other subject, generating, from the ultrasonic data, a sequence of a plurality of images in which each image shows a view of the subject at each position of the transducer, determining a difference between the images in the sequence, using the difference to determine a portion of the sequence corresponding to a sweep motion by the transducer and a portion of the sequence corresponding to a measurement motion by the transducer, and including. The measurement motion may be for determining the location and / or measuring the target feature.

[0015] An example of a data processing device 20 according to an embodiment is schematically shown in FIG. 3. The data processing device 20 is an example of an ultrasonic diagnostic device. In the present embodiment, the data processing device 20 is configured to process medical image data. In other embodiments, the data processing device 20 may be configured to process any other suitable image data.

[0016] The data processing device 20 includes a computing device 22 that is, in this example, a personal computer (PC) or a workstation. The computing device 22 is connected to a display screen 26, or other display device, and one or more input devices 28 such as a computer keyboard and mouse.

[0017] The computing device 22 is configured to obtain a data set from the data storage unit 30. At least a part of the data obtained from the data storage unit includes medical imaging data such as data obtained using, for example, the ultrasonic scanner 24. The medical image data may include two-dimensional, three-dimensional, or four-dimensional ultrasonic data. The scanner 24 includes an ultrasonic scanner system including an ultrasonic transducer 25. It includes an ultrasonic transducer that can obtain ultrasonic data used to generate an image sequence by moving across an area of a patient or other subject, and any suitable scanner system may be used as long as each image represents a view of the subject at each position of the transducer. Suitable ultrasonic scanner systems include, but are not limited to, Canon Aplio i700, i800 and i900, Canon Aplio Flex, and Canon Aplio Go.

[0018] The computing device 22 may receive data from one or more additional data storage units (not shown) instead of, or in addition to, the data storage unit 30. For example, the computing device 22 may receive medical image data from one or more remote data storage units (not shown) that may form part of a Picture Archiving and Communication System (PACS) or other information system.

[0019] The computing device 22 provides processing resources for automatically or semi-automatically processing data. The computing device 22 includes a processing device 32. The processing device 32 includes an optional model training circuit 34 configured to train one or more models, a data processing circuit 36 configured to apply the trained model(s) to perform other processes according to the embodiments, and an interface circuit 38 configured to obtain user or other inputs and / or output the results of data processing.

[0020] The data processing circuit may apply one or more algorithms, such as a sweep position algorithm, a positioning algorithm, a liver surface detection algorithm, a measurement operation algorithm, or a liver or other target position algorithm.

[0021] In some embodiments, any suitable pre-trained model may be used, such as a convolutional neural network (CNN), another neural network, or a transformer architecture. In the case of a neural network, any suitable number of nodes in the input, output, and hidden layers, and any suitable number and arrangement of layers may be used. Any suitable loss function, activation function, pooling or unpooling layers, or other model features may be used. The pre-trained models may be used alone or in combination to provide the functionality of any of the algorithms mentioned in the specification. The models may be trained using supervised, semi-supervised, or unsupervised methods. For example, the models may be trained with annotated ultrasound image data including labels for lesions, liver surfaces, and / or Quinaud or other regions. The annotations may be provided, for example, by a human expert.

[0022] In alternative embodiments, a trained machine learning model is not used; instead, the algorithm and / or other processes are implemented using any suitable known non-machine learning computer programming technique.

[0023] In this embodiment, circuits 34, 36, and 38 are each implemented in the computing device 22 by a computer program having instructions readable by a computer executable to perform the method of the embodiment. However, in other embodiments, various circuits may be implemented as one or more application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs). In other embodiments, distributed processing may be provided for at least a portion of the processing that is provided in different locations or on different devices, such as in a network computing system or a cloud computing system.

[0024] Furthermore, the computing device 22 includes a hard drive, an operating system including RAM, ROM, a data bus, various device drivers, and other PC components including hardware devices such as a graphics card. Such components are not shown in Figure 3 for clarity.

[0025] The data processing device 20 in Figure 3 is configured to perform the methods shown and / or described below.

[0026] Figure 4 is a schematic diagram showing an example of a method 400 for processing ultrasound images and identifying and measuring lesions or other target features to determine the location of lesions or other target features within the liver, according to an embodiment.

[0027] In step 402, the operator initiates an ultrasound scan by sweeping the ultrasound transducer 25 across the abdomen of the patient or other subject and collecting ultrasound data received from the transducer. The transducer is considered to be performing a sweep motion in this part of the method. A sweep motion may be performed for various reasons, but is usually done to search for a target feature of interest, such as a lesion or other pathology.

[0028] The sweep motion includes the motion of the ultrasonic transducer 25 on the surface of the subject. During this motion, the ultrasonic transducer is usually in contact with the surface of the subject. Typically, the sweep motion follows a line or curve corresponding to the surface of the subject. In some examples, the sweep motion may include a curved motion that follows, for example, the contour of the patient's surface. When the transducer 25 follows a straight line, the transducer does not tilt. However, when it follows a curved portion of the surface of the subject, the axis of the transducer will tilt with respect to the Cartesian coordinate system. For example, when considering the Cartesian coordinate system, it can be seen that when the transducer sweeps across a curved surface of the subject, the axis of the transducer perpendicular to the measurement surface is tilted, and when the transducer sweeps across a plane of the subject, the axis is constant. This is distinguished from the measurement motion, which may include, for example, significant rotational motion of the transducer, as described later. For example, the essence of rotational motion during measurement, such as the amount or direction of rotation or other properties of rotational motion, can be easily distinguished from the gentle inclination that may occur when the transducer sweeps against the curved surface of the subject.

[0029] Generally speaking, since ultrasound acquires images from within the subject, the transducer will typically remain in contact with the subject's surface. Thus, unless the user requests repositioning of the transducer during the procedure (for example, repositioning to return to the sweep position after placing it in the liver region of a discovered lesion), the transducer will not move towards or away from the subject. Generally, though not necessarily, the user maintains contact with the subject throughout the entire procedure. Thus, the sweep motion is performed in contact with the subject's surface and typically follows a simple linear or curved path. The images observed during the sweep change in response to changes in the anatomical structure beneath the transducer as it follows the path. The translation of the transducer to the lesion location can be thought of as motion involving the task of centering, for example, a currently visible feature in a 2D view, or otherwise positioning it. Transducer translation, such as centering a feature of interest in a view, is also performed while the transducer remains in contact with the subject's surface. Typically, only the measurement motion, described later in relation to step 408, has a significant rotational component. Sweep motion typically does not include significant rotational motion. Often, translation motion, such as centering a feature of interest, such as a target, on an image displayed to the user, and translation motion preceding the measurement motion, also does not include significant rotation. Once the start of the measurement motion is determined, the previous image may be analyzed to identify where the translation motion began, for example, where there was a transition from sweep motion to translation motion to center the feature of interest.

[0030] The scan may be a subcostal scan. Contact of the ultrasonic transducer with the subject includes situations in which a gel or other material is placed between the transducer and the surface of the subject, according to known ultrasonic techniques. In other embodiments, the subject may be scanned using various other operations of the ultrasonic transducer 25.

[0031] The ultrasonic transducer 25 scans the subject in a specified direction and acquires ultrasonic data. This ultrasonic data is received by a receiving circuit (not shown) of the processing unit 32, and an ultrasonic image sequence is generated by an image generation circuit (not shown). In this case, the receiving circuit is an example of an acquisition unit, and the image generation circuit is an example of an image generation unit. The ultrasonic image acquired from the transducer at the selected time under consideration may be referred to as the current image. The ultrasonic image acquired from the transducer during the sweep operation is called the sweep image and can be distinguished from the measurement image described later.

[0032] The interface circuit 38, the data processing circuit 36, and the display device 26 cooperate to display the operator interface to the operator on the display device 26. The data processing circuit 36 ​​is an example of a determination unit and a control unit. Any suitable operator interface may be provided. For example, it may include a window that displays the current image generated during scanning. It may include the ability to change the appearance of a specific feature of interest or a specific image, such as annotations, highlights, zoom in, or other specific features of interest, such as target features. The operator interface may also be used to display control parameters or measurement parameters and to give workflow commands or descriptions. Any suitable format and functionality of the operator interface may be provided. For example, it may include any suitable known operator interface features and functionality.

[0033] The operator may select a workflow that defines the sweep direction as left to right or right to left. In other embodiments, the sweep direction may be defined automatically. Ultrasound images acquired during the sweep motion are labeled to indicate that they were acquired during the sweep motion.

[0034] A key feature of the embodiments is that the system can process ultrasound image sequences acquired using a transducer to determine differences between images in the sequence, and use these differences to determine the parts of the sequence corresponding to different types of motion, such as sweep motion or measurement motion for measuring target features. Furthermore, some embodiments feature that the system can automatically determine the transducer's position by, for example, aligning the ultrasound image to a reference image, and can assign locations (such as liver segments where target features are determined to exist) to target features that are the subject of the measurement motion. This reduces the skill level required to perform scans and assign liver segments or other locations to lesions or other target features, while simultaneously reducing the operator's workload.

[0035] Further details regarding image processing for determining different types of motion and assigning locations to lesions will be described using the system shown in Figure 2, according to the method illustrated in Figure 4.

[0036] In step 404, the position of the transducer is determined for each image in the sequence, or for at least some of the images in the sequence. Any suitable coordinate system may be used.

[0037] The transducer's position may be determined by matching the images in the sequence to at least one atlas, reference image, or other reference dataset. The matching may include, for example, any suitable rigid or non-rigid alignment procedure.

[0038] At least one of the images in the sequence may be processed by the processing device 32 to detect anatomical landmarks or features within the field of view of the transducer. In this way, the transducer can be assigned a relative position to the location of the anatomical landmark. The anatomical landmarks or features may include any features that are clearly visible in the ultrasound image, such as the blood vessels and contour or part of the contour of the liver.

[0039] A transducer may be assigned a relative position to one or more anatomical landmarks. In some embodiments, an anatomical landmark may be assigned a relative position to other anatomical landmarks.

[0040] To detect the relative motion of the transducer between images, for example, two or more images from an image sequence may be compared with each other. The difference between the captured images may be detected as a result of the transducer's motion, and this difference may be used to detect the transducer's displacement and / or velocity. By aligning at least a portion of the images with at least one atlas, reference image, or other reference image dataset to set the transducer's relative position to anatomical features in at least a portion of the images, the transducer's position in the other images can also be determined based on the determined relative motion.

[0041] The transducer position may be automatically updated using the relative position of the transducer determined with respect to at least one anatomical landmark, combined with measured displacement and / or velocity. The transducer position thus determined may be used to correct or adjust the transducer position determined using a comparison of the ultrasound image with a reference image. In another embodiment, the transducer position may be determined as a relative position to at least one anatomical landmark, where at least a portion of the acquired ultrasound image is compared with one or more liver images or liver surfaces.

[0042] A liver plane or liver image may be used as a reference image of the liver and surrounding anatomical structures. In ultrasound processing, a plane may be considered to include or represent an image of a specific anatomical structure acquired from a specific imaging direction. A liver plane may be considered to include or represent an image of at least a portion of the liver viewed from a specific direction. In some examples, a liver plane set is a set of images of at least a portion of the subject's liver acquired by an ultrasound transducer or other imaging device during a sweep motion over the subject's abdominal surface. The images may include images of the liver itself, as well as liver-related blood vessels and branching points of liver-related ducts. The images may be acquired at different transducer angles relative to the subject surface. The images may be acquired from multiple subjects to illustrate anatomical variations in different subjects.

[0043] A liver plane set may be a sequence of images of the liver and liver-related tubules. The order of the images in the sequence may be based on the operating direction of the ultrasound transducer used to obtain the liver plane. If an image obtained during a liver scan is identified as matching a known liver plane, the location of the ultrasound transducer 25 at the position where the image was acquired using that liver plane can be determined.

[0044] Anatomical variations between subjects mean that there are variations in the order of the planes, and not all subjects will have all the planes. However, upon completion of the sweep, each subject will have a sufficient number and order of planes identified, allowing for the identification of liver segments at any point during the sweep. Each liver plane image is annotated with the liver segments visible in that image. In this way, each liver plane can be used as an atlas for defining liver segments. Once a matching liver plane is identified, the processing unit 32 records the current position of the ultrasound transducer 25.

[0045] The liver plane set may be saved in memory such as the data storage unit 30 and provided to the sweep position algorithm during scanning. In the current embodiment, the liver plane is defined as an anatomical structure that can be identified in the left-to-right and right-to-left sweep directions relative to the liver. The liver plane images are used as an atlas image such that any position in a matched image corresponds to an equivalent position in the "current" liver plane image. Once the position of an anatomical feature is identified both on the acquired image and the reference image, an equivalent position on the reference image may be assigned to each position on the acquired image. If some planes have anatomical structures similar to others but occur in different parts of the plane identification sequence, the sweep position algorithm can use that plane identification sequence to refine its analysis.

[0046] Successfully identifying the liver plane and associating it with the acquired ultrasound image makes it possible to determine not only the relative position of the transducer to the landmark, but also the equivalent position of the transducer on the reference image. Subsequently, the transducer's subsequent operation can be used to determine a new relative position of the transducer to the landmark, either in addition to or independently of identifying the liver plane at that new position. There may be one liver plane associated with each ultrasound image, or there may be a subset of acquired ultrasound images with one associated liver plane. Two or more ultrasound images may be associated with the same liver plane. In this way, each anatomical feature detected and matching the reference image will have an assigned position.

[0047] The set of surfaces used by the algorithm is sufficient to locate ultrasound transducers within regions defined by the Quinaud classification. For example, this requirement is met by a set of surfaces that include major vascular landmarks and / or their branches in space. The set of surfaces may also include the shape of the liver margins and the origins and endpoints of tubules associated with the liver. Each liver surface is annotated with the liver regions it comprises. The liver surface images are used as an atlas, where any location in a matched image corresponds to an equivalent location in the identified liver surface images. Liver surface detection in known sweep orders, such as left-to-right or right-to-left, may contribute to the process of identifying liver regions.

[0048] Anatomical variations between subjects can be interpreted as variations in the order of the planes, meaning that not all subjects necessarily possess all planes.

[0049] Before comparing the ultrasound image with the liver surface, the Couinard segmentation may be non-rigidly deformed to match the anatomical structure depicted in the acquired ultrasound image. To determine the next liver surface in the liver surface sequence, the order of the liver surfaces identified according to the acquired ultrasound image of the subject can be used. If the acquired ultrasound image matches the liver surface, the ultrasound image can be used to determine whether the sweep is from left to right or right to left.

[0050] Step 404 continues until a candidate liver lesion or other target feature is observed. The operator may input into the system that a candidate lesion or other target feature has been observed.

[0051] Step 406 is invoked when the operator observes a liver lesion in the acquired ultrasound image. If no lesion is detected by further investigation by the operator, the sweep motion of step 402 continues until the operator inputs. If the operator observes a candidate lesion during the sweep, the operator must confirm or deny the presence of the lesion. Confirmation of a candidate lesion may require additional operation of the ultrasound transducer. If a lesion is confirmed, the operator typically needs to perform further transducer operation to measure the lesion. A candidate lesion may be a false positive, and the operator may dismiss it as a lesion at any point after the initial observation.

[0052] If a lesion is detected in step 406, the transducer enters a measurement mode in step 408, which includes a measurement motion of the transducer. The measurement motion may include, or can be performed immediately after, a translation, in which the transducer moves to a location assigned to an anatomical feature and positions the anatomical feature in the center of the ultrasound image generated by the transducer at the location associated with the anatomical feature, or at any other desired location. The measurement motion may further include rotating the ultrasound transducer to rotate the imaging plane. Rotation may be performed to measure the maximum and minimum dimensions of the lesion, or to perform any other desired measurements. Rotation motion following translation motion during the measurement phase may include rotating the transducer while maintaining the lesion in the imaging plane of the transducer. Identifying the start location of translation motion preceding measurement motion can be difficult in pure time-advancing analysis. However, measurement rotation is easily detectable. Once rotation is observed, the previous image can be backtracked from that point to identify the point / image where the translation motion began (for example, the point at which the substitution from sweep motion to translation motion occurred in order to center on the feature of interest).

[0053] During measurement mode, the processing unit 32 may identify the Quinaud region containing the lesion and store one or more associated ultrasound images and / or reference images. The region can be automatically identified by the processing circuit as the relative location of the transducer to a patient anatomical structure, such as a liver segment. This may be automatically determined by processing the ultrasound image using the processing of stage 402. The processing unit 32 may provide the identified liver segment to the operator when the operator wishes to confirm the identified liver segment or input the correct liver segment containing the lesion. There may be a set of eligibility conditions that require operator confirmation, such as when a threshold is used and / or when the reliability of the resulting results is low. In some examples, the liver plane that traces the liver plane containing the lesion in the sweep direction of the first scan may be used as the first liver segment. The ultrasound transducer may then sweep in the opposite direction to the sweep direction of the first scan. The liver segment classification using the two different scan directions may then be compared. If the two classifications do not correspond, the processing unit may consider the reliability of the liver segment identification to be low and request operator confirmation. Associated ultrasound images and / or reference images may be annotated to identify anatomical features and liver segments according to the Quinault classification. Images of liver segments and / or visual and / or auditory indications, or any other suitable indications, may be provided to the operator using the interface circuit 38.

[0054] If the operator measures a lesion or denies it as a false positive, in step 410, the operator returns the transducer 25 to the sweep position where the candidate lesion was first observed before the measurement motion began. Images recorded during the measurement phase are labeled as having been acquired during the measurement motion. The operator may be automatically given a command via the operator interface to return the transducer to the sweep position immediately before the measurement motion began, and the sweep may be resumed from the interrupted position. This command may be generated by the system based on the transducer position determined by processing the ultrasound images.

[0055] In the processing shown in Figure 4, a suitable alignment algorithm is used to match the image data to an atlas, reference image(s), or other reference data. In other embodiments, the data processing circuit 36 ​​may use one or more suitable pre-trained models applied to the ultrasound data and / or images to match the image data to an atlas, reference image(s), or other reference data, and / or to determine the location(s), and / or to determine different types of motion. The pre-trained machine learning model may process the acquired image sequence to determine the relative position of the transducer to one or more anatomical landmarks in the transducer's imaging plane or to reference points selected by the operator. In another embodiment, the data processing circuit 20 may include a machine learning model trained on a dataset including one or more liver images or liver planes, which processes the acquired image sequence to determine the relative position of the transducer to one or more anatomical landmarks in the transducer's imaging plane or to reference points selected by the operator. In some embodiments, the pre-trained machine learning model(s) may be trained using a discretionary model training circuit 34, or may be trained by another device and downloaded to the device 20. In other embodiments, the trained machine learning model(s) may be hosted on a remote server, and the processing circuit may send the data to the trained machine learning model(s) for processing via, for example, a network or other suitable connection, and receive the results of the processing.

[0056] Returning to Figure 5, further details of the measurement of lesions or other target features in several embodiments are described.

[0057] When measuring a lesion, the lesion is centered, and then the transducer is rotated to determine and record the maximum diameter of the lesion. Typically, the user then determines and records the maximum diameter perpendicular to the first measured diameter, and these measurements give an elliptical model of the lesion. The probe is then returned to the sweep position.

[0058] Figure 5 shows five images illustrating an example of the image acquired from the scan and the effect of transducer movement on the contours of anatomical landmarks visible within the scan. In this embodiment, vascular landmarks are described, but other landmarks may be used.

[0059] Figure 5a shows the sweep plane and liver 62 with the illustrated sweep image plane 64. The sweep image plane is perpendicular to the axis of the transducer view or the sweep axis. The image plane is then rotated counterclockwise to obtain the rotated image plane 66. The future image plane 68 shows the image plane when the sweep is continued without interruption and without rotation of the image plane.

[0060] Figures 5b-5e illustrate how the operation of the transducer affects the visual characteristics of the ultrasound image obtained from the transducer. Figure 5b shows the field of view 72 of the imaging device and the blood vessel 70 when the transducer is stationary. The blood vessel 70 is visible in this image because it shares, or nearly shares, the sweep axis perpendicular to the sweep plane. Other blood vessels in the field of view are not visible because they are not aligned with the sweep axis.

[0061] Figure 5c shows the field of view when the transducer completes the translation motion. The translation motion may include, and may be referred to as, image panning. When the transducer completes the translation motion, new anatomical features will enter the image from the edges, and existing anatomical features will move out of the image edges. Anatomical features already in the center of the image may be moved up or down, but their relative positions should be maintained. There are few perceptible changes in the shape and size of blood vessels and the field of view during the translation motion.

[0062] Figure 5d shows the field of view when the transducer completes a typical sweep motion. During a sweep, new objects can enter the image from anywhere (similarly, old objects can leave the image from anywhere). Also, the relative position and shape of items in the image will change during a sweep. Small fluctuations are visible in the size and / or contour of vessel 70 during the sweep motion. Changes in the visible liver margin are slow during the sweep motion. The size and position of the field of view are similarly subject to small fluctuations. As the sweep progresses, other vessels may temporarily align with the sweep axis and temporarily appear in the image sequence. In Figure 5d, one of these vessels is labeled as the second vessel 74. Figure 5e shows the field of view when the transducer completes a rotation motion. Vessel 70 is shown significantly distorted and may disappear from the view. This may be due to misalignment between the vessel axis during rotation and the sweep axis. Also, the size and position of the visible liver portion are significantly affected in Figure 5e.

[0063] The transducer's initial centering motion is more focused on translation within the current image plane than on movement along the left-right liver axis, followed by a new center point / rotation near the center point / small shift.

[0064] Figure 6 is a schematic diagram showing an example of an ultrasound imaging method 500 according to an embodiment, which identifies and measures lesions and determines the location of lesions within the liver.

[0065] In step 502, the transducer 25 begins a linear sweep motion on the subject's liver. In step 504, the processing unit 32 determines, image by image, whether the image matches the liver surface or a previous sweep image.

[0066] All images except those generated during the measurement process are sweep images. The processing unit may use a trained machine learning model to determine if there is a match.

[0067] For example, if there is a match between the “current image” considered in relation to another embodiment and one or more liver surfaces, in step 508, the one or more liver surfaces are labeled as “current liver surface,” and all ongoing measurement operations are terminated. As a result, in step 510, the transducer 25 returns to sweep operation.

[0068] If the liver plane has not yet been defined, the process will be at the start of the sweep, and there will be no atlas. In this initial stage, the regions can be determined using known anatomical information, for example, by dividing them into Quinaud's regions 6 and 7, or regions 2 and 3, using information about the liver margin, for example. Alternatively, or additionally, once a first liver plane is reached after lesion detection, the location of the previously detected lesion can be defined using the location of that first liver plane.

[0069] If, in step 506, there is no liver surface or previous sweep image that matches the ultrasound image, then in steps 512 and 514, it is determined whether the processing of one or more ultrasound images using the processing device 32 exhibits translation motion, rotational motion, or a combination of translation motion and rotational motion.

[0070] If the processing of the ultrasound image sequence determines that only translation motion can be identified in the image, the sweep motion is continued in step 510.

[0071] If rotation is detected, or if a translation followed by rotation is detected, in step 516, transducer 25 enters a measurement motion. In step 516, the situation is such that the trend is defined as a measurement. The trend can be backtracked to the starting point, and these images in the backtracked image are defined as measurement motions rather than sweep motions.

[0072] You may move transducer 25 to a position where the lesion is centered in the view of transducer 25.

[0073] In step 518, the operator moves the transducer to position it over the lesion and the region of the liver containing the lesion. If the lesion is labeled as a false positive, or if the lesion is identified and measured, the transducer returns to its final position in the sweep motion before the measurement motion began, and the sweep continues in step 510.

[0074] The process in step 518 may be performed after the measurement rotation has started. It may also be performed retrospectively with the next liver surface detection (in the case of the first region, or in the case of locating in the region using both the previous and subsequent liver surfaces). In some embodiments, a less reliable flag may be used if the location determined using the previous and subsequent liver surfaces does not match.

[0075] For each detected liver lesion, the sweep position can be determined from the point where the transducer transitions from sweep motion to translation motion or measurement motion. The processing unit can determine the location of the lesion from the translation used to recenter the ultrasound image to the lesion during the measurement motion phase. This provides the location information necessary to identify the liver segment of the lesion, for example, using an annotated reference image.

[0076] Next, further embodiments, or additional alternative features usable in conjunction with the embodiments described above, will be described. A set of settings that the operator may select / input to initiate a liver sweep using the imaging technique according to the embodiments will be described below. The device may communicate with the operator using interface circuit 38.

[0077] 1) Define the current phase: "Sweep" or "Measure" The phase may be a phase for confirming / measuring lesions while the scan is in progress. In some embodiments, the confirmation / measurement phase may be entered automatically, and in other embodiments, it may be started manually.

[0078] 2) Set the current phase direction: The user sets the direction to either "left to right" or "right to left".

[0079] 3) Start image recording: Start the ultrasound hardware and begin imaging from the ultrasound transducer.

[0080] 4) Start the liver surface detection algorithm: The liver surface detection algorithm compares the current image with a set of potential liver surfaces to find matches. The liver surface algorithm may use a trained machine learning model to compare the image with the liver surfaces.

[0081] 5) Start the measurement-action algorithm: The measurement-action algorithm recognizes the operation of the ultrasound transducer to measure lesions, which is different from the operation of the ultrasound transducer in an uninterrupted liver sweep. The transducer position is used to assign a location to the lesion and to determine the equivalent location on an annotated reference image.

[0082] 6) Flag the current image with the current phase: The phase may include "sweep" and "measure".

[0083] In another embodiment, the face detection algorithm may modify the a priori probability of a potential face using data from previous face identification, sweep direction, and / or sweep phase information. The face detection algorithm may also modify probabilities during post-processing, using the output probabilities of one or more potential faces to identify the most likely one or more faces. The algorithm may include a machine learning model that performs some or all of the processing. Such a method may include the following steps and criteria:

[0084] Step 1: Enter the current image.

[0085] Step 2: Enter one or more of the following: sweep direction, transducer phase, and previous liver surface identification.

[0086] Criterion 1: When the transducer is in the sweep phase in a known direction, the expected value can be used to modify the a priori probability of the next face.

[0087] Criterion 2: When the transducer transitions from the measurement motion phase to the sweep motion phase, it does not need to return to the exact previous sweep position. This is because the transducer may be misaligned due to user error, mechanical factors, or other reasons. In such cases, the desired sweep position is likely to be the previously identified face, or otherwise, the next most likely face in the sequence after the last identified face.

[0088] In another embodiment, the measurement-action identification algorithm recognizes the action of the ultrasound transducer to measure a lesion, which is different from the action of the ultrasound transducer in an uninterrupted liver sweep. When confirming / measuring a lesion, it centers the candidate lesion and then rotates the ultrasound transducer to measure the lesion. Once the lesion is measured, or discarded as a false positive, the transducer is returned to the sweep position. The measurement-action algorithm may include a machine learning model that performs some or all of the actions performed.

[0089] The algorithm may compare the current ultrasound image with a previous ultrasound image or reference image. When comparing the current image with a previous image, the algorithm considers two types of operation: The previous image may be one of a set of liver face images; The previous image may be an image acquired earlier in the same scan. The method for detecting measured motion may include the following steps.

[0090] Case 1: In-plane translation

[0091] Step 1: Compare the images, assuming the current image is a 2D translation of the previous image. Calculate the translation required to obtain the best image match, outputting the goodness-of-fit metric (e.g., DICE) and the best translation shift.

[0092] Step 2: If the translation fit metric is greater than the set threshold, flag the transducer as likely to have been translated.

[0093] Case 2: Rotation

[0094] Since the liver is not a sphere, the rotation of the transducer will change the point where the transducer surface intersects with the liver volume. This significantly affects the size and position of the liver margins in the view. Note that the liver margin passing through the rotation point is nominally fixed, so the changes mainly occur at the left / right liver margins of the transducer. This is in contrast to a sweep motion along a sweep axis, where the liver margins change slowly in the expected manner.

[0095] Step 1: Extract the visible liver periphery(s) from the image.

[0096] Step 2: As the simplest metric, calculate the length of the visible liver margin in each image. Compare the change in liver margin length against a threshold to determine if the current image is rotated. The threshold used should preferably be set according to the detected current liver surface.

[0097] Step 3: The rotation probability can be output as a metric based on the magnitude of the length change obtained by dividing by the "rotation threshold".

[0098] Step 4: If the rotation metric is greater than the set threshold, set a flag in the transducer to indicate that rotation has occurred.

[0099] Based on previously identified liver surfaces, thresholds applied to different metrics may be adjusted. Multiple previous images may be compared to the current image. By using two or more previous images, the system can identify trends in transducer behavior, with tolerance for small, insignificant changes rather than flashy, instantaneous changes. Translation metrics calculated by the behavior-measurement algorithm may be averaged over a buffer of previous images. Similarly, rotation metrics calculated by the behavior-measurement algorithm may be averaged over a buffer of previous images. If the average translation fit metric is greater than a set threshold, a flag is set on the transducer indicating that translation is likely to have occurred. If the average rotation metric is greater than a set threshold, a flag is set on the transducer indicating that rotation is likely to have occurred.

[0100] If the system shows only in-plane translation followed by rotation, or significant rotation (relative to the lesion on the sweep axis), then the system must be in the measurement-operation phase.

[0101] If the user adds a new lesion measurement, the system knows it must be in the measurement-operation phase. The algorithm will then determine that the transducer is in the measurement phase, even if the likelihood of translation and rotation is low.

[0102] If the system detects that it has transitioned from the sweep phase to the measurement-action phase, the algorithm can backtrack to the earliest image matching the translation or rotation trend to determine where the previous sweep phase ended. This image may be labeled as a candidate lesion location.

[0103] The following criteria may be used to terminate the measurement operation.

[0104] Case 1: Sweep Phase The liver surface is detected by a surface detection algorithm. If a liver surface is detected, the sweep position is reset to the detected liver surface or the "sweep position".

[0105] Case 2: Compare the current image with an image recorded prior to the measurement motion phase. Comparison: Reuse the translation method and accept translation tolerances, as the transducer may not return to the exact same spot. If the current image and the previous sweep image match well, accept the translation tolerance between the images, then end the measurement-motion phase and restart the sweep phase. Set the sweep position of the matched image to the position of the matched previous image.

[0106] In some embodiments, more complex metrics can be used to identify the in-plane translation and rotation state of the transducer or imaging device. The two-dimensional image from the transducer is formed from axes perpendicular to two sweep axes. One of these axes may be referred to as "up / down" and represent the length of the patient's body. The other may be referred to as "internal / external" and represent the height / depth of the transducer. The internal / external axis corresponds to "near / far". The left / right axis is from the viewpoint of the image and is the up / down axis. Rotation modifies the distance to the liver periphery along the left-right axis of the transducer more than the distance to the near-far liver periphery in front of the transducer. Similarly, the current projection shape of the liver periphery also changes, so alternative metrics such as image compactness (periphery divided by area) can be used. Nonlinear averaging can be used for trends to avoid sporadic noise. In some examples, an arithmetic mean may be used. In other embodiments, a local median may be used to allow for transducer oscillation. Each of these metrics may be used in the embodiment to determine whether the transducer is in the sweep phase or the measurement phase.

[0107] In some embodiments, the positioning algorithm determines the location of a lesion in the acquired image and identifies the liver segment containing the lesion. The positioning algorithm may include the use of a trained machine learning model.

[0108] The sweep position can be considered as the position resulting from the movement of the transducer along the sweep axis. As the sweep continues, liver planes are detected, and the sweep position may be defined by the last detected liver plane; for example, the sweep position may be "on liver plane XYZ" or "after liver plane XYZ". If retrospective positioning is also performed, for example before detecting the first liver plane, or for reliability evaluation, the position may also be defined as "before liver plane ABC" and "after plane XYZ". The processing for "after" or "before" may be substantially the same, both functioning to identify the sweep axis position where the transducer is currently located. The liver plane that coincides with the scan image at the location of the lesion and may be called the "current liver plane" provides an atlas image that can be used to identify liver segments in the coverage area including the sweep plane.

[0109] Step 1: If the operator observes a lesion during the sweep, the sweep motion is stopped at that point. The sweep position is the currently / last viewed plane. If a first liver plane has not yet been detected, the next detected liver plane may be used to define the sweep position for lesion detection. The sweep position provides an appropriate atlas image for the liver region.

[0110] Step 2: The operator centers the lesion in the current view by panning the view of the ultrasound transducer. After measuring the lesion, this translation action defines the position of the transducer within the atlas of the sweep location.

[0111] Step 3: To measure the long and short axes of the lesion, the operator rotates the transducer over the lesion and records the appropriate measurements.

[0112] Step 4: Once the user has completed the lesion measurement, the lesion region is defined using the atlas from Step 1 and the positional information from Step 2. If the first liver plane has not yet been identified, simpler methods may be used to define the liver region, such as the upper or lower half of the image or similar variations thereof. In some cases, the determination of the liver segment can be delayed until the first liver plane is identified, and the atlas image is provided using the liver plane identified at that point.

[0113] To obtain uniaxial measurements of lesion locations within the liver, a sweep plane position can be used. The matched liver plane provides the necessary atlas image to identify liver segments in this portion of the sweep.

[0114] The necessary translation from the sweep position to the measurement position yields two other coordinates. These are later compared to the current liver plane to identify the liver segment of the lesion.

[0115] The current liver plane may be used as an atlas image containing all liver segments visible on the annotated liver plane image. Annotated liver segments on the liver surface: The liver surface image will be used later as an atlas image so that any location within the matching image corresponds to the equivalent point in the identified liver surface image. Translation required from the center of the sweep phase to the center of the lesion. ○ When the user initiates image rotation instead of translation, the system identifies the lesion center as the center of the image. output: Liver segment for a given lesion

[0116] Lesions may be observed between consecutive surface detections. Each liver lesion may be surrounded by two liver surface detections. Lesions observed in only one surface may be analyzed as described in the embodiments.

[0117] If two interfaces are detected for a given lesion, the position algorithm can be run on both liver surfaces and the results can be compared. If the results do not match, this may suggest that operator review may be necessary. The review can be performed when identifying the lesion during the scan, leaving manual operator intervention in the local context.

[0118] In some embodiments, the sweep detection algorithm compares the detected liver surface sequences to determine whether the sequence contains a left-to-right scan sweep or a right-to-left scan sweep, or whether it cannot be derived from the input data. The algorithm may include a trained machine learning model.

[0119] Automating liver segment identification allows operators with skills limited to lesion identification and measurement to perform sweeps and pinpoint the location of lesions in addition to measuring them.

[0120] In various embodiments, the identification of liver segments for each lesion is automated by (a) and (b) below. (a) Identifying sequences of known locations within the liver, and (b) Identify which type of ultrasound transducer operation the operator is currently performing. The transducer may perform either a sweep operation or a lesion measurement operation.

[0121] By combining knowledge of sweep locations through sequences of known locations with actions identified as lesion measurements, it becomes possible to automatically identify liver segments. This automation reduces the time required for sweeping and the skill level needed to perform it.

[0122] Methods according to various embodiments may offer numerous benefits to the efficiency of processing for subjects, operators, and the clinical environment. Unless excluded, the method would reduce the complex series of transducer operations typically used for liver segment identification. The method could increase scan speed, benefiting both subjects and operators, as well as the system's output in the clinical environment. The method would simplify the operator's processing workflow, assisting less experienced operators with the technically challenging and time-consuming task of liver segment identification. It would also reduce the potential for musculoskeletal damage to operators due to stressful operations.

[0123] According to at least some embodiments, an ultrasonic diagnostic apparatus is provided that includes a processing circuit. The processing circuit is To locate target features, ultrasound data is received from an ultrasound transducer that moves across the area of ​​the patient or other subject. From the ultrasound data, a sequence of multiple images is generated in which each image shows a view of the subject at each position of the transducer. The difference between images in the sequence is determined, and the portion of the sequence corresponding to the sweep motion by the transducer and the portion of the sequence corresponding to the measurement motion by the transducer are determined using the difference.

[0124] The processing circuit may use the difference between the images to determine the portion of the sequence corresponding to the translation motion that centers or otherwise positions the target feature on one of the plurality of images prior to the measurement motion.

[0125] The processing circuit may assign locations to the target features based on the position of the transducer in the portion of the sequence corresponding to the measurement motion.

[0126] The processing circuit may determine the position of the transducer relative to the patient or other subject by processing the sequence of ultrasound images and identifying at least one anatomical feature in at least a portion of the plurality of images.

[0127] The aforementioned at least one anatomical feature may include at least one of the following: the margin of the liver, blood vessels, vascular structures, branching points of blood vessels or other ductal structures, a part of the liver, or a liver segment.

[0128] Identifying the at least one anatomical feature may include matching at least one of the multiple images in the sequence to at least one atlas, reference image, or other reference dataset.

[0129] The aforementioned at least one atlas, reference image, or other reference dataset may include at least one liver surface.

[0130] The processing circuit may match the plurality of images to a sequence of plurality of liver surfaces.

[0131] The measurement motion may include at least rotational motion.

[0132] The processing circuit may determine at least one of the rotational motion, translational motion, or matching motion based on the difference between the images in the sequence.

[0133] The difference between images in the sequence may include a difference in alignment, and the alignment is performed between the plurality of images in the sequence and one or more reference images.

[0134] The determination of at least one of the rotational motion, translational motion, or matching motion may include comparing the measurement of the difference between the images with a threshold.

[0135] The sweep motion may include linear or curved motion in a direction different from the linear or curved motion that may be included in the measurement motion.

[0136] The processing circuit may determine the direction of motion of the transducer using the difference between the images.

[0137] The difference between images may include at least one of the following: a difference in the shape, size, or orientation of at least one feature within the image, or a change in the spacing, relative size, or orientation of multiple features within the image.

[0138] The aforementioned ultrasound diagnostic device is i) The target features include lesions or other pathologies. ii) The measurement motion includes motion associated with a measurement process for measuring the size or other properties of the target feature. iii) The ultrasound diagnostic apparatus further includes the transducer, It may be at least one of the following.

[0139] Assigning a location to the target feature may include assigning the target feature to one of several liver segments arranged according to the Quinaud section or other sectioning scheme.

[0140] The processing circuit may apply a trained model to the ultrasonic data or the plurality of images to determine the difference between the images, and / or determine the portion of the sequence corresponding to the sweep motion by the transducer and the portion of the sequence corresponding to the measurement motion.

[0141] The ultrasound diagnostic apparatus may further include a user input device, and the processing circuit, based on the user input received via the user input device, i) Recording the start or end of a sweep motion or measurement motion, ii) Confirming or rejecting the location to which the target feature is assigned, iii) Approving or rejecting candidate target features, iv) Record the measurement data, or v) Enabling the resumption of the sweep motion or measurement motion by providing an output indicating that the transducer has returned to the previous sweep position or measurement position, or an output providing guidance on how to return to the previous sweep position or measurement position. Perform at least one of the following:

[0142] According to at least some embodiments, an ultrasound diagnostic method is provided. The ultrasound diagnostic method is To locate target features, ultrasound data is received from an ultrasound transducer that is moving across or has moved across the area of ​​the patient or other subject. From the ultrasound data, a sequence of multiple images is generated in which each image shows a view of the subject at each position of the transducer. To determine the difference between images in the aforementioned sequence, Using the difference, the portion of the sequence corresponding to the sweep motion by the transducer and the portion of the sequence corresponding to the measurement motion by the transducer are determined. Includes.

[0143] A given embodiment provides a method for automatically identifying the location of a lesion observed from a sequence of 2D ultrasound images from an ultrasound scan of the liver, as defined according to the Quinaud classification. The method may include at least some or all of the following steps. The objective is to identify images in a sequence that match a known liver plane from a set of known liver planes, where the liver plane is a specific view of identifiable anatomical structures within the liver, and may be defined from the viewpoint of a liver sweep obtained in a left-right (right-left) sweep direction, and the liver plane may be sufficient to identify all major vascular landmarks and their branches (including anatomical variations) within the liver. The ultrasonic transducer i) When moving along the left-right axis of the liver to perform a standard sweep of the liver, ii) When moving to view and measure lesions within the liver, Identifying from the image sequence, This involves identifying translations in the image plane rather than movement along the left-right axis of the liver, where the lesion is moved to the center of the view (which can be zero if the lesion is already in the center), When a centrally located lesion is viewed from multiple angles for measurement, it is necessary to identify the rotation of the imaging plane, Identifying the "last sweep image" from the image sequence, it was determined that the ultrasound transducer was moving along the left-right axis of the liver immediately before it was identified that the transducer was moving for viewing and measuring the lesion. The translation action after the last sweep image of the UL transducer, which is to calculate the translation action to the location identified as rotating to measure a lesion within the liver, By identifying the liver surface and translating the last sweep image to the lesion measurement location, the location within the liver is determined. Includes.

[0144] Each liver surface detection image may have regions that match an annotated Quinaud classification. Subsequently, the previous liver surface may be used as an atlas, and the Quinaud classification may be identified using translation coordinates determined by the translation required to recenter the lesion image from the last sweep.

[0145] The Quinaud segment atlas of the liver surface may be first subjected to non-rigid deformation before determining the Quinaud segment to obtain the best agreement with the visible anatomical structures of the liver (e.g., the location of the liver margin and major ducts) observed in the last image detected as an agreement with the liver surface.

[0146] Translations may be identified by comparing images based on the assumption of a previous 2D translation. The translation required to obtain the best image match may be calculated while outputting the goodness-of-fit metric and the best translation shift. If the goodness-of-fit exceeds a threshold, the image may be flagged as a potential translation action.

[0147] A series of consecutive images may be flagged as a potential translation operation. All images from the first such flagged images may be identified as being the result of an operation for viewing and / or measuring lesions within the liver.

[0148] The images may be analyzed assuming the possibility of the ultrasound transducer rotating. The current extent of the imaging plane of the liver due to the transducer position may be measured using one or more suitable metrics. If the change in the metric exceeds a threshold, the image may be flagged as a potential rotational movement.

[0149] A series of consecutive images may be flagged as representing a potential rotational movement. All images from the first such flagged images may be identified as being due to a movement for viewing and / or measuring a lesion within the liver.

[0150] A threshold may be set to identify potential translations based on the previously detected liver surface.

[0151] A threshold may be set to identify potential rotation based on the previously detected liver surface.

[0152] The ultrasound operator may be required to confirm the Quinault classification or manually override it.

[0153] The following liver surface may be used in the same way as the previous surface. If the two resulting Quinaud divisions are identical, this can be considered a reliable result. If they differ, the less reliable result may be returned.

[0154] You may only ask the ultrasound operator to verify the Quinaud classification if unreliable results are returned.

[0155] In some embodiments, features other than lesions within the liver may be the features of interest.

[0156] The user interface of the device may display the available current Quinaud segments given in the previous liver surface detection. This may be done, for example, by any one or more of the following: text annotations for each potential Quinaud segment, a color specific to each potential Quinaud segment, an area in the liver schematic diagram representing each potential Quinaud segment, color highlighting of the current ultrasound image representing the potential Quinaud segment, or a combination of these features.

[0157] The user interface of the ultrasound diagnostic device may indicate a Quinaud segment, which is identified when the current ultrasound transducer location is considered to be at the center of a lesion. This may be done by any one or more of the following: a text annotation for the potential Quinaud segment, a color specific to the potential Quinaud segment, an area in a liver schematic diagram indicating the potential Quinaud segment, color highlighting of the current ultrasound image indicating the potential Quinaud segment, or a combination of these features.

[0158] While specific circuits are described herein, in alternative embodiments, one or more functions of these circuits may be provided by a single processing resource or other component, or a function provided by a single circuit may be provided by a combination of two or more processing resources or other components. A reference to a single circuit encompasses multiple components that provide the functionality of that circuit, regardless of whether such components are separated from each other. A reference to multiple circuits encompasses a single component that provides the functionality of those circuits.

[0159] Although embodiments of the present invention have been described above, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments and their variations can be implemented in a variety of other forms, and various omissions, substitutions, changes, and combinations can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims of the invention and its equivalents. [Explanation of Symbols]

[0160] 20 Data Processing Devices 22 Computing devices 24 Scanners 25 transducers 26 Display Screens 28 Input devices 30 Data storage unit 32 Processing Unit 34 Model Training Circuits 36 Data Processing Circuits 38 Interface Circuits

Claims

1. An acquisition unit that receives ultrasound data from an ultrasound transducer that moves across the area of ​​a patient or other subject in order to identify the location of a target feature, A generation unit generates a sequence of multiple images from the ultrasound data, each image representing a view of the subject at each position of the ultrasound transducer. A determination unit that determines the difference between images in the sequence and uses the difference to determine the portion of the sequence corresponding to the sweep motion by the ultrasonic transducer and the portion of the sequence corresponding to the measurement motion by the ultrasonic transducer, An ultrasound diagnostic device equipped with the following features.

2. The determination unit uses the difference between the images to determine the portion of the sequence corresponding to the translation motion that centers or otherwise positions the target feature on one of the plurality of images prior to the measurement motion. The ultrasound diagnostic apparatus according to claim 1.

3. The determination unit assigns a location to the target feature based on the position of the ultrasonic transducer in the portion of the sequence corresponding to the measurement motion. The ultrasound diagnostic apparatus according to claim 1.

4. The determination unit processes the sequence of ultrasound images and identifies at least one anatomical feature in at least a portion of the plurality of images to determine the position of the transducer relative to the patient or other subject. The ultrasound diagnostic apparatus according to claim 1.

5. The aforementioned at least one anatomical feature includes at least one of the following: the margin of the liver, a blood vessel, a vascular structure, a branching point of a blood vessel or other vascular structure, a part of the liver, or a liver segment. The ultrasound diagnostic apparatus according to claim 4.

6. Identifying the at least one anatomical feature includes matching at least one of the multiple images in the sequence to at least one atlas, reference image, or other reference dataset. The ultrasound diagnostic apparatus according to claim 5.

7. The above at least one refers to at least one atlas, reference image, or other reference dataset, which includes at least one liver surface. The ultrasound diagnostic apparatus according to claim 6.

8. The determination unit matches the plurality of images to a sequence of plurality of liver surfaces, The ultrasound diagnostic apparatus according to claim 7.

9. The measurement motion includes at least rotational motion. The ultrasound diagnostic apparatus according to claim 2.

10. The determination unit determines at least one of the rotation motion, the translation motion, or the matching motion based on the difference between the images in the sequence. The ultrasound diagnostic apparatus according to claim 9.

11. The difference between images in the sequence includes a difference in alignment, and the alignment is performed between the plurality of images in the sequence and one or more reference images. The ultrasound diagnostic apparatus according to claim 10.

12. The determination of at least one of the rotational motion, translational motion, or matching motion includes comparing the measurement of the difference between the images with a threshold. The ultrasound diagnostic apparatus according to claim 10.

13. The sweep motion includes linear or curved motion in a direction different from the linear or curved motion that may be included in the measurement motion. The ultrasound diagnostic apparatus according to claim 9.

14. The determination unit uses the difference between the images to determine the direction of motion of the ultrasonic transducer. The ultrasound diagnostic apparatus according to claim 1.

15. The difference between images includes at least one of the following: a difference in the shape, size, or orientation of at least one feature within the image, or a change in the spacing, relative size, or orientation of multiple features within the image. The ultrasound diagnostic apparatus according to claim 1.

16. i) The target feature includes a lesion or other pathology, ii) The measurement motion includes motion associated with a measurement process for measuring the size or other properties of the target feature. iii) The ultrasound diagnostic apparatus further includes the ultrasound transducer, At least one of the following applies: The ultrasound diagnostic apparatus according to claim 1.

17. Assigning a location to the target feature includes assigning the target feature to one of several liver segments arranged according to the Quinard segmentation or other segmentation scheme. The ultrasound diagnostic apparatus according to claim 3.

18. The determination unit determines the difference between images by applying the trained model to the ultrasound data or the plurality of images, and / or determines the portion of the sequence corresponding to the sweep motion by the ultrasound transducer and the portion of the sequence corresponding to the measurement motion. The ultrasound diagnostic apparatus according to claim 1.

19. Further equipped with a user input device, Based on the user input received via the user input device, i) Record the start or end of the sweep motion or measurement motion. ii) Confirming or rejecting the location to which the target feature is assigned, iii) Approving or rejecting candidate target features, iv) Recording measurement data, or v) Enabling the resumption of the sweep motion or measurement motion by providing an output indicating that the ultrasonic transducer has returned to the previous sweep position or measurement position, or an output providing guidance on how to return to the previous sweep position or measurement position. The control unit further comprises performing at least one of the following: The ultrasound diagnostic apparatus according to claim 1.

20. An ultrasound diagnostic method using an ultrasound diagnostic device, To locate target features, ultrasound data is received from an ultrasound transducer that is moving across or has moved across the area of ​​the patient or other subject. From the ultrasound data, a sequence of multiple images is generated in which each image shows a view of the subject at each position of the ultrasound transducer. Determine the difference between images in the aforementioned sequence. Using the difference, the portion of the sequence corresponding to the sweep motion by the ultrasonic transducer and the portion of the sequence corresponding to the measurement motion by the ultrasonic transducer are determined. Ultrasound diagnostic methods.

Images