A heart scan positioning method, device, storage medium and electronic equipment

Abstract

The embodiment of the specification discloses a kind of heart scan positioning method, device, medium and equipment, with initial positioning scan sequence and anatomical positioning scan sequence, method includes: performing initial positioning scan sequence for positioning scanning, generate the initial scan positioning parameter of anatomical positioning scan sequence;Perform anatomical positioning scan sequence for acquisition reconstruction, obtain the three-dimensional body data of heart part to carry out analysis processing, generate the scan positioning space parameter of multiple standard heart anatomical section, to call scan positioning space parameter when executing heart target diagnosis scan sequence Generation of magnetic resonance scan instruction is scanned.Above technical scheme, by the initial scan positioning parameter generated by initial positioning scan sequence, to guide anatomical positioning scan sequence to obtain high-resolution three-dimensional body data, and then generate the scan positioning space parameter of standard heart anatomical section, eliminate the subjective bias of manual operation, improve the automation level and consistency of scan positioning.

Description

Technical Field

[0001] This specification relates to the field of scanning and positioning technology, and in particular to a cardiac scanning and positioning method, device, storage medium, and electronic device. Background Technology

[0002] Cardiac magnetic resonance imaging (MRI) has become the "gold standard" for assessing cardiac structure, function, and myocardial tissue characteristics due to its advantages such as being radiation-free and having high soft tissue contrast. However, current cardiac MRI scanning procedures heavily rely on the professional experience and subjective judgment of technicians. Significant differences exist among operators, leading to subjective biases in the identification and definition of cardiac anatomy, which can easily result in errors in scan plane localization. Furthermore, the cumbersome and time-consuming procedures severely restrict clinical throughput, posing a particular challenge to patient groups with poor cooperation, such as those with heart failure, the elderly, and children. Summary of the Invention

[0003] This specification provides a cardiac scanning localization method, apparatus, storage medium, and electronic device, the technical solutions of which are as follows: In a first aspect, embodiments of this specification provide a cardiac scanning localization method, configured with an initial localization scanning sequence and an anatomical localization scanning sequence, the method comprising: An initial localization scan sequence is executed to perform a localization scan to obtain a triorthogonal anatomical image of the heart, and spatial region of interest is analyzed on the triorthogonal anatomical image to generate initial scan localization parameters for the anatomical localization scan sequence. Based on the initial scanning and positioning parameters, the anatomical positioning scanning sequence is executed to acquire and reconstruct high-resolution three-dimensional volume data of the heart region. Based on the three-dimensional volume data, multi-scan plane analysis processing is performed to generate multiple standard cardiac anatomical sections and scanning positioning spatial parameters. These scanning positioning spatial parameters are then used to generate magnetic resonance scanning commands for scanning control processing when executing cardiac target diagnostic scanning sequences.

[0004] In one feasible implementation, the step of generating initial scanning localization parameters for the anatomical localization scanning sequence by spatial region of interest parsing of the triorthogonal anatomical images includes: Target recognition is performed on the cardiac region in the triorthogonal anatomical images to obtain the spatial location information of the cardiac region after filtering out background chest tissue; The three-dimensional target bounding box of the heart region in three-dimensional space is determined based on spatial location information; Initial scan positioning parameters for the anatomical positioning scan sequence are generated based on the spatial geometric features of the three-dimensional target bounding box.

[0005] In one feasible implementation, generating initial scan localization parameters for the anatomical localization scan sequence based on the spatial geometric features of the three-dimensional target bounding box includes: Extract the three-dimensional coordinates of the center point of the three-dimensional target bounding box; The three-dimensional target bounding box is calculated in three orthogonal directions in the human anatomical coordinate system, including the front-back direction, the head-to-toe direction, and the left-right direction. A spatial reference coordinate system is constructed based on the three-dimensional coordinates of the center point and the three orthogonal directions. Scanning field parameters and positioning line parameters are generated based on the spatial reference coordinate system. Initial scanning positioning parameters for the anatomical positioning scanning sequence are obtained based on the scanning field parameters and positioning line parameters.

[0006] In one feasible implementation, the step of performing multi-scan plane analytical processing based on the three-dimensional volume data to generate scanning positioning spatial parameters for multiple standard cardiac anatomical sections includes: The high-resolution type of 3D volume data is input into a pre-trained deep learning model; The deep learning model performs multiplanar spatial regression calculations on each standard cardiac anatomical section in three-dimensional space to obtain the scanning positioning spatial parameters of each standard cardiac anatomical section.

[0007] In one feasible implementation, the standard cardiac anatomical sections include axial sections, coronal sections, sagittal sections, left ventricular short-axis sections, two-chamber sections, three-chamber sections, four-chamber sections, and right ventricular outflow tract sections; the scanning positioning spatial parameters include the plane normal vector and plane center coordinates of each of the standard cardiac anatomical sections in three-dimensional space.

[0008] In one feasible implementation, after obtaining the scanning positioning spatial parameters of each of the standard cardiac anatomical sections, the method further includes: Extract the planar normal vectors of each of the aforementioned standard cardiac anatomical sections; A preset multi-plane reconstruction algorithm is invoked to perform three-dimensional interpolation and resampling operations on the three-dimensional volume data along the normal vectors of each plane, so as to reconstruct a two-dimensional planar preview image corresponding to each standard cardiac anatomical section. The two-dimensional plane preview image is displayed in the user interface.

[0009] In one feasible implementation, the step of calling the scan positioning spatial parameters to generate magnetic resonance scan commands for scan control processing during the execution of a cardiac target diagnostic scan sequence includes: Save the scanning positioning spatial parameters of multiple standard cardiac anatomical sections; When a command to start the cardiac target diagnostic scanning sequence is received, the scanning positioning spatial parameters are read as target section spatial parameters according to the protocol tag of the cardiac target diagnostic scanning sequence. Based on the target section spatial parameters and the cardiac target diagnostic scanning sequence, instruction parameters are configured to generate magnetic resonance scanning instructions for driving the underlying scan.

[0010] In one feasible implementation, after performing multi-scan plane analysis processing based on the three-dimensional volume data, the method further includes: When the multi-scan plane parsing process fails, the automatic positioning process is suspended and the manual positioning scanning process is triggered. In response to manual positioning line adjustment operations performed by the user through graphical control software, the reference anatomical plane parameters determined manually are obtained; The reference anatomical plane parameters are updated and saved so that when performing cardiac target diagnostic scanning sequences, the scan positioning spatial parameters are called to generate magnetic resonance scanning commands for scan control processing.

[0011] Secondly, embodiments of this specification provide a cardiac scanning positioning device, the device comprising: The scanning module is used to perform an initial localization scanning sequence to obtain a triorthogonal anatomical image of the heart, and to perform spatial region of interest parsing on the triorthogonal anatomical image to generate initial scanning localization parameters for the anatomical localization scanning sequence. The acquisition module is used to perform the anatomical positioning scan sequence based on the initial scan positioning parameters to acquire and reconstruct high-resolution three-dimensional volume data of the heart region. The parsing module is used to perform multi-scan plane parsing processing based on the three-dimensional volume data to generate scanning positioning spatial parameters for multiple standard cardiac anatomical sections. These scanning positioning spatial parameters are then used to generate magnetic resonance scanning commands for scanning control processing when executing cardiac target diagnostic scanning sequences.

[0012] Thirdly, embodiments of this specification provide a computer storage medium storing a plurality of instructions adapted for loading by a processor and executing the above-described method steps.

[0013] Fourthly, embodiments of this specification provide an electronic device that may include: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to execute the above-described method steps.

[0014] The beneficial effects of the technical solutions provided in some embodiments of this specification include at least the following: In one or more embodiments of this specification, an initial positioning scan sequence is executed to perform a positioning scan, generating initial scanning positioning parameters for the anatomical positioning scan sequence. The anatomical positioning scan sequence is then executed to acquire and reconstruct three-dimensional volume data of the heart region for analysis and processing, generating scanning positioning spatial parameters for multiple standard cardiac anatomical sections. These scanning positioning spatial parameters are then invoked to generate magnetic resonance imaging (MRI) scan commands during the execution of the cardiac target diagnostic scan sequence. By employing the above technical solution, the initial scanning positioning parameters generated by the initial positioning scan sequence guide the anatomical positioning scan sequence to acquire high-resolution three-dimensional volume data, thereby generating scanning positioning spatial parameters for standard cardiac anatomical sections. This eliminates subjective biases from manual operation and improves the automation level and consistency of scanning positioning. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic flowchart of a cardiac scanning localization method provided in the embodiments of this specification; Figure 2 This is a schematic diagram of a scenario for a cardiac scanning localization method provided in the embodiments of this specification; Figure 3 This is a schematic flowchart of another cardiac scanning localization method provided in the embodiments of this specification; Figure 4 This is a schematic flowchart of another cardiac scanning localization method provided in the embodiments of this specification; Figure 5 This is a schematic flowchart of another cardiac scanning localization method provided in the embodiments of this specification; Figure 6 This is a schematic diagram of the structure of a cardiac scanning and positioning device provided in the embodiments of this specification; Figure 7 This is a schematic diagram of the structure of an electronic device provided in the embodiments of this specification. Detailed Implementation

[0017] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0018] In the description of this specification, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. In the description of this specification, it should be noted that, unless otherwise expressly specified and limited, "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices. Those skilled in the art can understand the specific meaning of the above terms in this specification based on the specific circumstances. Furthermore, in the description of this specification, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0019] The present specification will now be described in detail with reference to specific embodiments.

[0020] In one embodiment, such as Figure 1 As shown, a cardiac scanning localization method is proposed. This method can be implemented using a computer program and can run on a cardiac scanning localization device based on the von Neumann architecture. The computer program can be integrated into an application or run as a standalone utility application. The cardiac scanning localization device can be a terminal device.

[0021] Specifically, the cardiac scanning localization method includes: S102: Perform an initial localization scan sequence to obtain a triorthogonal anatomical image of the heart, and perform spatial region of interest parsing on the triorthogonal anatomical image to generate initial scan localization parameters for the anatomical localization scan sequence. In a fully automated MRI cardiac scanning and positioning system, the initial positioning scan sequence, also known as the Cardiac_Scout sequence, is an ultra-fast positioning sequence where physical and spatial geometric parameters are statically set before the scan begins via the imaging protocol manager. This sequence acquires triorthogonal anatomical images of the heart within the thoracic cavity in the shortest possible time, establishing a static, coarse anatomical reference coordinate system for subsequent high-precision scans. The initial positioning scan sequence pre-sets a low acquisition matrix and a relatively thick slice thickness to compress the scan time to within a few seconds of a single breath-hold or free breathing. In terms of spatial geometry, it pre-sets the scanning field of view to be large enough to cover the entire adult thoracic cavity, and pre-configures the slice orientation to be completely parallel to the standard anatomical axes of the human body in three orthogonal directions (i.e., axial, coronal, and sagittal). Simultaneously, the scanning geometric center is locked at the magnet center by default or initialized by a surface positioning point.

[0022] The anatomical localization scanning sequence, also known as the Cardiac_SmartScout sequence, is a high-resolution 3D volumetric data acquisition sequence that is statically preset at the physical parameter level but dynamically calculated and filled in at the spatial geometric parameter level during runtime. It is used to acquire high-resolution 3D volumetric data containing rich anatomical details based on dynamically calculated precise spatial guidance, serving as the core data foundation for subsequent multiplanar regression to generate standard cardiac anatomical section scanning localization spatial parameters. In terms of acquisition parameters, it pre-sets a high acquisition matrix and sets slice thickness and interslice spacing to isotropic or near-isotropic. For contrast, it pre-configures fat suppression pulses or blood flow contrast enhancement technology to maximize the image contrast between the myocardium and the blood pool. The key difference between the anatomical positioning scan sequence and the initial positioning scan sequence is that the spatial geometric parameters (including the scan center coordinates, scan block boundaries, and three-dimensional rotation angles) of the anatomical positioning scan sequence are left as null values ​​at the beginning of the workflow. Only after the initial positioning scan sequence is completed and a three-dimensional target bounding box and spatial reference coordinate system are generated in the background will the system's underlying resolver automatically pull the origin of the spatial reference coordinate system and the three anatomical axis offset values, convert them into a spatial orientation matrix, and automatically fill the calculated geometric values ​​into the scan control field within milliseconds before the anatomical positioning scan sequence is officially triggered.

[0023] Triorthogonal anatomical images refer to three mutually perpendicular anatomical plane images obtained through the initial positioning scan sequence, corresponding to the transverse, sagittal, and coronal planes in human anatomy, respectively. The three images together provide rough location information of the heart in three-dimensional space, which is the basic data for subsequent spatial positioning and bounding box calculations.

[0024] Spatial region of interest resolution refers to the process of automatically identifying and spatially analyzing the heart region in triorthogonal anatomical images. Specifically, it includes filtering out chest background tissue through target detection or image segmentation algorithms, locating the two-dimensional contour of the heart region in each image, and then fusing information from the three orthogonal directions to infer the actual location and extent of the heart in three-dimensional space.

[0025] Initial scan positioning parameters refer to the spatial configuration parameters used to guide the data acquisition of anatomical positioning scan sequences. In specific implementations, these parameters include scan field parameters and positioning line parameters. Their function is to ensure that the high-resolution sequence can accurately cover the cardiac region and avoid missed scans or off-center scans.

[0026] For example, a rough panoramic image of the heart in three orthogonal directions is quickly acquired by executing a preset initial positioning scan sequence, and an algorithm is used to automatically perform spatial analysis of the heart region in the image, thereby replacing manual judgment. The blurry anatomical image is transformed into precise three-dimensional spatial positioning parameters, establishing a spatial reference for subsequent high-resolution precision scanning, ensuring that the device can automatically and accurately focus on the heart itself, laying the foundation for fully automated scanning.

[0027] S104: Based on the initial scanning positioning parameters, the anatomical positioning scanning sequence is executed to acquire and reconstruct high-resolution three-dimensional volume data of the heart region; Acquisition and reconstruction refers to the high-precision data acquisition based on dynamically filled spatial geometric parameters and statically preset physical parameters during the execution of the anatomical positioning scanning sequence. The original signal is then processed by a three-dimensional reconstruction algorithm, including spatial encoding and decoding, Fourier transform, and image reconstruction, to convert the original data domain signal into a three-dimensional volumetric data image in the spatial domain.

[0028] High-resolution 3D volumetric data refers to 3D image data with isotropic or near-isotropic spatial resolution obtained after acquisition and reconstruction. Its voxel size is basically equal in the three orthogonal directions, and it can be reconstructed from any angle in multiple planes without obvious image blurring or structural distortion.

[0029] For example, the initial localization scan sequence itself acquires low-resolution, thick-slice triorthogonal images, the image quality of which is insufficient to support subsequent high-precision multiplanar regression analysis of standard cardiac anatomical sections. If these low-resolution images are used directly for multiplanar analysis, due to the partial volume effect and insufficient spatial resolution caused by the thick slices, the deep learning model will be unable to accurately identify the subtle anatomical structures of the heart, and the accuracy of the regressed scan plane parameters will be difficult to guarantee. Therefore, it is necessary to perform high-resolution anatomical localization scan sequences for acquisition and reconstruction. The purpose is to obtain high-quality three-dimensional volume data with isotropic or near-isotropic spatial resolution, which can be clearly reconstructed in multiple planes from any angle, providing a reliable data foundation containing rich anatomical details for the accurate regression of subsequent deep learning models.

[0030] S106: Based on the three-dimensional volume data, perform multi-scan plane analysis processing to generate multiple standard cardiac anatomical sections and scan positioning spatial parameters. When executing the cardiac target diagnostic scan sequence, call the scan positioning spatial parameters to generate magnetic resonance scan commands for scan control processing.

[0031] Standard cardiac anatomical views refer to the standardized imaging planes used in cardiac magnetic resonance imaging (MRI) to assess cardiac structure, function, and myocardial tissue characteristics. Standard cardiac anatomical views include axial, coronal, sagittal, left ventricular short-axis, two-chamber, three-chamber, four-chamber, and right ventricular outflow tract views. The left ventricular short-axis view is used to assess left ventricular volume and ejection fraction; the four-chamber view is used to assess the size and function of the four chambers; the two-chamber and three-chamber views are used to assess ventricular wall motion in specific myocardial segments; and the right ventricular outflow tract view is used to assess the morphology of the right ventricular outflow tract.

[0032] Scanning and positioning spatial parameters refer to the set of parameters describing the position and orientation of each standard cardiac anatomical section in three-dimensional space. The parameters for each section include the plane normal vector and the coordinates of the plane's center point. The plane normal vector is used to determine the spatial orientation of the section (i.e., which direction the section is perpendicular to), and the coordinates of the plane's center point are used to determine the specific position of the section in three-dimensional space.

[0033] Cardiac target diagnostic scanning sequence refers to a functional scanning sequence used to acquire diagnostic images after localization. It can be a cardiac target diagnostic scanning sequence for any standard cardiac anatomical section. It is the sequence that obtains the final diagnostic image in cardiac magnetic resonance imaging, and its image quality directly depends on the accuracy of the scanning localization.

[0034] For example, deep learning models can be used to perform multi-plane regression analysis on 3D volume data, generating spatial positioning parameters for multiple standard cardiac anatomical sections simultaneously. Standardized section positioning information is intelligently extracted from the 3D volume data, providing directly callable, precise positioning parameters for subsequent target diagnostic sequences. This fundamentally avoids the amplification effect of key point detection errors, ensuring the standardization and repeatability of section positioning.

[0035] Please see Figure 2 , Figure 2 This is a schematic diagram of a cardiac scanning localization method proposed in this manual. The radiologist, as the operator of the system, issues operating commands and receives localization results and image preview information from the system through the host computer software MRConsole (magnetic resonance imaging control console, i.e., graphical control software running on a computer) located in the first-level user interaction layer. MRConsole can initiate calls to the second-layer business logic and intelligent processing layer, triggering the CardiacScoutPlanner module. This module performs spatial region of interest analysis based on the triorthogonal anatomical images acquired from the initial localization scan sequence (i.e., the Cardiac_Scout sequence). It identifies the heart region through object detection or segmentation algorithms, calculates bounding boxes in three-dimensional space, and generates initial scan localization parameters for the anatomical localization scan sequence (i.e., the Cardiac_SmartScout sequence). These localization parameters are stored in the temporary data cache module of the third-layer data and infrastructure layer for subsequent optimization of cached localization lines. Here, scan_plane refers to the human heart scan plane. Subsequently, the CardiacSmartScoutPlanner module is triggered. This module, combined with the deep learning inference module, performs multi-plane regression analysis based on the high-resolution three-dimensional volume data acquired and reconstructed from the anatomical localization scan sequence, generating scan localization spatial parameters for multiple standard cardiac anatomical sections. These scan localization spatial parameters are fed back to the user interaction layer for visualization on the MRConsole interface for physician preview and confirmation, and are also stored in the third-layer results database and temporarily cached. In addition, during the processing, the business logic layer can also load deep learning models and inference parameters from the model and configuration module of the third layer, and read or cache positioning parameters and intermediate results from the temporary data cache and result database.

[0036] On the other hand, MRConsole interacts directly with the MRI scanner in the fourth-layer external system and hardware layer, issuing scan control protocols and receiving acquired image signals. The acquired raw image data is stored in the third-layer medical image storage module (Digital Imaging and Communications in Medicine, DICOM). Furthermore, the fourth layer also includes the Hospital Information System (HIS) and the Picture Archiving and Communication System (PACS). The system supports data interaction with HIS / PACS, transmitting acquired DICOM images and generated scan positioning parameters back to these systems, enabling the storage, retrieval, archiving management, and clinical use of image data.

[0037] In the embodiments of this specification, an initial positioning scan sequence is executed to perform a positioning scan, generating initial scanning positioning parameters for the anatomical positioning scan sequence. The anatomical positioning scan sequence is then executed to acquire and reconstruct three-dimensional volume data of the heart region for analysis and processing, generating scanning positioning spatial parameters for multiple standard cardiac anatomical sections. These scanning positioning spatial parameters are then invoked to generate magnetic resonance imaging (MRI) scan commands during the execution of the cardiac target diagnostic scan sequence. By employing the above technical solution, the initial scanning positioning parameters generated by the initial positioning scan sequence guide the anatomical positioning scan sequence to acquire high-resolution three-dimensional volume data, thereby generating scanning positioning spatial parameters for standard cardiac anatomical sections. This eliminates subjective biases from manual operation and improves the automation level and consistency of scan positioning.

[0038] Please see Figure 3 , Figure 3 This is a schematic flowchart illustrating another embodiment of a cardiac scanning localization method proposed in this specification. Specifically: S202: Perform the initial localization scan sequence to obtain triorthogonal anatomical images of the cardiac region; For details, please refer to the method steps in other embodiments of this specification, which will not be repeated here.

[0039] S204: Perform target recognition on the heart region in the triorthogonal anatomical image to obtain the spatial location information of the heart region after filtering out the chest background tissue; Target recognition refers to the process of automatically identifying and locating the heart region from triorthogonal anatomical images using image processing algorithms (such as target detection, image segmentation, or deep learning models). Specifically, it includes distinguishing the heart tissue from the surrounding chest background tissue (such as lungs, bones, muscles, etc.) and outputting the contour or boundary information of the heart in each two-dimensional image.

[0040] Spatial location information refers to the set of data obtained after target recognition to describe the position of the heart in three-dimensional space. It includes the two-dimensional coordinate range of the heart on the respective planes of the three orthogonal anatomical images (axial, sagittal, and coronal), as well as the approximate central position and spatial distribution range of the heart in three-dimensional space derived by fusing information from the three orthogonal directions.

[0041] For example, triorthogonal anatomical images contain a large amount of background chest tissue information, such as the sternum, ribs, lung fields, and mediastinal fat. These background tissues overlap with the heart region in grayscale or have blurred boundaries. If an image containing the entire chest background is directly used for subsequent bounding box calculations, the interference from the background tissues will make it difficult for the algorithm to accurately define the true boundaries of the heart, and may even misidentify non-heart tissues as heart regions, resulting in incorrect bounding box sizes and positions. Therefore, it is necessary to accurately separate the heart region from the complex chest background to obtain pure spatial location information of the heart region, providing a reliable data foundation for the accurate calculation of subsequent 3D target bounding boxes.

[0042] S206: Determine the three-dimensional target bounding box of the heart region in three-dimensional space based on spatial location information; A 3D target bounding box refers to the smallest hexahedron that completely contains the target object (i.e., the heart region) in 3D space. It is typically an axis-aligned bounding box aligned with the anatomical coordinate system axes or a rotatable directed bounding box. In this scheme, the three dimensions of the bounding box correspond to the front-back direction, the head-to-toe direction, and the left-right direction, respectively. Its geometric features include the coordinates of the bounding box's center point and the side lengths in the three directions.

[0043] For example, the 3D target bounding box transforms the complex and irregular heart contour into standardized geometric parameters consisting of the center point coordinates and the side lengths in three directions by completely enclosing the heart region with a regular minimum hexahedron. This regularized representation not only facilitates spatial calculations by the system (such as calculating the scan center and determining the scan range), but also provides a unified geometric benchmark for the subsequent construction of a spatial reference coordinate system.

[0044] S208: Generate initial scanning positioning parameters for the anatomical positioning scanning sequence based on the spatial geometric features of the three-dimensional target bounding box; For example, the bounding box itself cannot be directly used as a scan control parameter that an MRI scanner can recognize and execute. When performing high-resolution scan sequences, the MRI scanner needs to specify the following spatial geometric parameters: which spatial coordinates the scan center should be aligned with (corresponding to the center point of the bounding box), the size of the scan field of view to cover (corresponding to the dimensions of the bounding box in each direction), and how the orientation of the scan plane should be set (corresponding to the three orthogonal directions of the bounding box). If these parameters are missing, the high-resolution scan sequence cannot be executed automatically and still requires manual input or adjustment by the technician. Therefore, it is necessary to convert the spatial geometric features of the three-dimensional target bounding box into initial scan positioning parameters that the MRI scanner can directly read and execute. In addition, these initial scan positioning parameters need to be temporarily cached for rapid reuse during multi-sequence scans.

[0045] In one feasible implementation, step S208 includes the following steps: S12: Extract the three-dimensional coordinates of the center point of the three-dimensional target bounding box; The three-dimensional coordinates of the center point refer to the geometric center of the smallest hexahedron that completely contains the heart region in three-dimensional space. This center point represents the overall geometric center of the heart in three-dimensional space and serves as the origin reference for subsequently constructing the spatial reference coordinate system. For example, in specific calculations, the center point coordinates are usually obtained by averaging the minimum and maximum boundary values ​​of the bounding box in the front-back direction, head-to-toe direction, and left-to-right direction.

[0046] For example, without extracting the center point coordinates separately, the system cannot determine which spatial location should be the origin of the spatial reference coordinate system, nor can it generate subsequent scan positioning parameters accordingly. Furthermore, in clinical practice, the heart is not always located at the physical isocenter of the MRI machine. The heart position varies among patients due to body size, respiratory status, or pathological changes (such as cardiac displacement or enlargement). Directly using the machine's default isocenter as the scan center may cause the heart region to deviate from the center of the scan field of view. Therefore, it is necessary to explicitly extract the three-dimensional coordinates of the center point from the bounding box as the precise origin of the spatial reference coordinate system.

[0047] S14: Solve the three orthogonal directions of the three-dimensional target bounding box in the human anatomical coordinate system, the three orthogonal directions including the front-back direction, the head-to-toe direction and the left-right direction; Solving refers to determining the three mutually perpendicular directional vectors of a three-dimensional target from its spatial shape and orientation in the sense of human anatomy. This process may involve coordinate transformation, principal component analysis, or geometric calculations based on image spatial location information.

[0048] The human anatomical coordinate system refers to a three-dimensional spatial coordinate system established with reference to the human body's own anatomical structure. Its three coordinate axes correspond to the anterior-posterior direction, the head-to-toe direction, and the left-to-right direction of the human body, respectively. This coordinate system changes with the body's position and differs from the inherent coordinate system of magnetic resonance imaging equipment (established based on the geometry of the scanning bed and magnet).

[0049] For example, in magnetic resonance imaging (MRI) scans, the angles of the scanning plane and the directions of the positioning lines need to be defined based on anatomical orientations. For instance, the short-axis section of the heart needs to be perpendicular to the long axis of the heart, which runs from the apex to the base. This long axis is typically close to the head-to-foot direction in the human anatomical coordinate system but not perfectly parallel. If the device's inherent coordinate system is used directly as the scanning reference, changes in patient position (such as raising or lowering the arm) or the natural tilt of the heart within the chest cavity can cause the scanning plane to mismatch with the actual anatomical structure. Therefore, it is necessary to calculate the three orthogonal directions of the heart in the human anatomical coordinate system from the spatial geometry of the bounding box. These directions are then used as the axes of the spatial reference coordinate system, ensuring that the generation of all subsequent scanning positioning parameters is based on the patient's own anatomical orientation, rather than the device's fixed geometric orientation.

[0050] S16: Construct a spatial reference coordinate system based on the three-dimensional coordinates of the center point and the three orthogonal directions, generate scanning field parameters and positioning line parameters based on the spatial reference coordinate system, and obtain initial scanning positioning parameters for the anatomical positioning scanning sequence based on the scanning field parameters and positioning line parameters.

[0051] A spatial reference coordinate system refers to a three-dimensional rectangular coordinate system established with the center point as the origin and the three calculated orthogonal directions as the coordinate axes. This coordinate system uses the geometric center of the patient's heart as the origin and the anatomical orientation of the heart as the coordinate axis directions, providing a unified spatial reference for the generation of all subsequent scan positioning parameters. Unlike the inherent coordinate system of the MRI equipment, this coordinate system adaptively adjusts according to changes in the patient's cardiac anatomy.

[0052] Scanning field of view parameters refer to the parameters that determine the acquisition range of an MRI scan, and are usually expressed as the size of the field of view. These parameters are calculated based on a spatial reference coordinate system and the side dimensions of the bounding box in each direction, ensuring that the scan range completely covers the cardiac region without including excessive irrelevant background tissue.

[0053] Positioning line parameters are parameters that determine the position and orientation of the scanning plane in space, including the center point of the scanning plane and the rotation angle of the scanning plane relative to each coordinate axis of the spatial reference coordinate system. These positioning line parameters are generated based on the origin and axis directions of the spatial reference coordinate system and are used to guide the positioning of the scanning center and the setting of slice angles in anatomical positioning scanning sequences.

[0054] For example, the original 3D target bounding box represents the rough spatial extent of the heart, while MRI equipment requires specific scan field values ​​and positioning line spatial positions during scanning. There are differences in format and semantics between the two. Therefore, it is necessary to first construct a spatial reference coordinate system based on the center point coordinates and three orthogonal directions. Then, within this coordinate system framework, the spatial geometric features of the bounding box are transformed into scan field parameters and positioning line parameters that the equipment can recognize, ultimately forming a complete initial scan positioning parameter set to guide the automatic execution of anatomical positioning scan sequences.

[0055] S210: Based on the initial scanning positioning parameters, the anatomical positioning scanning sequence is executed to acquire and reconstruct high-resolution three-dimensional volume data of the heart region; For details, please refer to the method steps in other embodiments of this specification, which will not be repeated here.

[0056] S212: Based on the three-dimensional volume data, perform multi-scan plane analysis processing to generate multiple standard cardiac anatomical sections and scan positioning spatial parameters. When executing cardiac target diagnostic scan sequences, call the scan positioning spatial parameters to generate magnetic resonance scan commands for scan control processing.

[0057] For details, please refer to the method steps in other embodiments of this specification, which will not be repeated here.

[0058] In the embodiments of this specification, an initial positioning scan sequence is executed to perform a positioning scan, generating initial scanning positioning parameters for the anatomical positioning scan sequence. The anatomical positioning scan sequence is then executed to acquire and reconstruct three-dimensional volume data of the heart region for analysis and processing, generating scanning positioning spatial parameters for multiple standard cardiac anatomical sections. These scanning positioning spatial parameters are then invoked to generate magnetic resonance imaging (MRI) scan commands during the execution of the cardiac target diagnostic scan sequence. By employing the above technical solution, the initial scanning positioning parameters generated by the initial positioning scan sequence guide the anatomical positioning scan sequence to acquire high-resolution three-dimensional volume data, thereby generating scanning positioning spatial parameters for standard cardiac anatomical sections. This eliminates subjective biases from manual operation and improves the automation level and consistency of scan positioning.

[0059] Please see Figure 4 , Figure 4 This is a schematic flowchart illustrating another embodiment of a cardiac scanning localization method proposed in this specification. Specifically: S302: Perform an initial localization scan sequence to obtain a triorthogonal anatomical image of the heart, and perform spatial region of interest parsing on the triorthogonal anatomical image to generate initial scan localization parameters for the anatomical localization scan sequence; For details, please refer to the method steps in other embodiments of this specification, which will not be repeated here.

[0060] S304: Based on the initial scanning positioning parameters, the anatomical positioning scanning sequence is executed to acquire and reconstruct high-resolution three-dimensional volume data of the heart region; For details, please refer to the method steps in other embodiments of this specification, which will not be repeated here.

[0061] S306: Input the high-resolution type of three-dimensional volume data into a pre-trained deep learning model, and use the deep learning model to perform multi-plane spatial regression operations on each standard cardiac anatomical section in three-dimensional space to obtain the scanning positioning spatial parameters of each standard cardiac anatomical section, so as to call the scanning positioning spatial parameters to generate magnetic resonance scanning instructions for scanning control processing when executing cardiac target diagnostic scanning sequence; A deep learning model refers to an artificial neural network model pre-trained on a large-scale labeled dataset. This model has learned the mapping relationship from 3D volume data to spatial parameters of standard cardiac anatomical sections. In this approach, the model takes 3D volume data as input and automatically extracts multi-scale spatial features from the volume data through a multi-layer neural network, outputting the spatial localization parameters of various standard cardiac anatomical sections. Pre-training means that the model has already completed the training process before being deployed to this system, and does not require retraining in actual use, only inference operations.

[0062] Multiplane spatial regression refers to the computational process by which a deep learning model directly predicts the position and orientation of various standard cardiac anatomical sections in three-dimensional space. Unlike traditional classification tasks, regression outputs continuous numerical parameters (rather than discrete class labels), specifically including the plane normal vector (describing the spatial orientation of the section) and the coordinates of the plane's center point (describing the position of the section in three-dimensional space) for each section.

[0063] For example, traditional procedures require technicians to manually locate multiple standard sections, such as the left ventricular short axis and four-chamber view, one by one. This is not only time-consuming and labor-intensive, but also subject to subjective differences in judgment among different technicians, directly affecting the accuracy and consistency of subsequent cardiac function measurements. Therefore, it is necessary to use deep learning models to perform automated multi-plane spatial regression calculations on three-dimensional volume data, simultaneously predicting the planar normal vectors and center point coordinates of all standard cardiac anatomical sections. This provides directly callable, precise positioning parameters for subsequent cardiac target diagnostic scanning sequences, achieving fully automated standardized positioning.

[0064] The following explains the model training process for deep learning models: Model creation: An initial deep learning model is created based on a machine learning model for a multiplanar spatial regression scenario of standard cardiac anatomical sections; Sample Data Acquisition: Acquire a large amount of sample data, which consists of high-resolution 3D volumetric data of the heart region collected historically. This 3D volumetric data has isotropic or near-isotropic spatial resolution characteristics, enabling multi-planar reconstruction from any angle, and contains complete information on cardiac anatomy, such as myocardial boundaries, cardiac chamber morphology, valve structure, and papillary muscles.

[0065] Sample Data Labeling: Based on the needs of multi-planar spatial regression scenarios using standard cardiac anatomical sections, an expert-based service is introduced to manually label the sample data with corresponding sample tags. These tags include standard spatial parameters for multiple standard cardiac anatomical sections for each 3D volume. The standard spatial parameters for each standard cardiac anatomical section include the standard plane normal vector and the coordinates of the standard plane center point in 3D space. The standard cardiac anatomical sections include axial, coronal, sagittal, left ventricular short-axis, two-chamber, three-chamber, four-chamber, and right ventricular outflow tract sections. To ensure labeling consistency, a cross-validation mechanism is introduced, where multiple experts independently label the data, and the consensus result is used as the final sample label.

[0066] Model training process: Sample data is input into the initial deep learning model for at least one round of training to obtain predicted spatial localization parameters, namely, the predicted plane normal vector and predicted plane center point coordinates output by the model for each standard cardiac anatomical section. Based on the predicted spatial localization parameters and the standard spatial parameters in the sample labels, a model loss function is used to determine the model loss value. This loss function can include the mean squared error loss function, the mean absolute error loss function, or a composite loss function combining angle error and distance error, used to quantify the angular deviation between the predicted plane normal vector and the standard plane normal vector, as well as the Euclidean distance deviation between the predicted plane center point coordinates and the standard plane center point coordinates. Based on this model loss value, the model parameters of the initial deep learning model are adjusted (e.g., updating network weights through backpropagation) until the model training termination condition is met, resulting in a trained deep learning model.

[0067] Optionally, the model's training termination conditions may include, for example, the loss function value being less than or equal to a preset loss function threshold, or the number of iterations reaching a preset threshold. Specific training termination conditions can be determined based on actual circumstances and are not specifically limited here.

[0068] It should be noted that the machine learning models involved in one or more embodiments of this specification include, but are not limited to, fitting of one or more of the following machine learning models: Convolutional Neural Network (CNN) model, Deep Neural Network (DNN) model, Recurrent Neural Networks (RNN) model, Embedding model, Gradient Boosting Decision Tree (GBDT) model, Logistic Regression (LR) model, etc.

[0069] In one feasible implementation, step S306 includes the following steps: S22: Saves the scanning positioning spatial parameters of multiple standard cardiac anatomical sections; For example, a single cardiac MRI scan typically requires multiple diagnostic sequences, each needing localization based on a corresponding anatomical section. Without saving these parameters, each diagnostic sequence requires rerunning the deep learning model for regression calculations, resulting in significant redundant computation and wasted time. Furthermore, minor fluctuations in model inference can lead to inconsistencies in section localization between different sequences. In addition, for patients requiring regular follow-up examinations, saved parameters can be directly reused during these visits, ensuring high consistency in section positions between examinations, thereby improving efficiency and guaranteeing localization consistency.

[0070] S24: When an instruction to start the cardiac target diagnostic scanning sequence is received, the scanning positioning spatial parameters are read as target section spatial parameters according to the protocol tag of the cardiac target diagnostic scanning sequence; Instructions for cardiac target diagnostic scanning sequences refer to the operational commands issued by radiologists through a user interface to initiate a specific diagnostic sequence scan. These commands include information about the type of sequence to be executed, such as initiating a four-chamber view scan or initiating a left ventricular short-axis view scan.

[0071] Protocol tags are identification information associated with each cardiac target diagnostic scan sequence, indicating the type of target anatomical section required for that sequence. For example, the protocol tag for a four-chamber view sequence corresponds to a four-chamber view. Protocol tags can be preset during sequence configuration or specified by the physician when creating the scan plan.

[0072] The target section spatial parameters refer to the scanning positioning spatial parameters of a specific standard cardiac anatomical section corresponding to the cardiac target diagnostic scanning sequence, which are read from the protocol tag of the sequence. These parameters may include the plane normal vector and the coordinates of the plane's center point. This parameter will serve as the spatial reference for subsequently generating magnetic resonance imaging (MRI) scan commands.

[0073] For example, different cardiac diagnostic scanning sequences require localization based on different anatomical sections. Although all parameters of multiple standard sections have been saved, the system does not know which section should be used for the currently initiated diagnostic sequence. Without a mapping mechanism between protocol tags and section parameters, the system cannot automatically select the correct parameters upon receiving the start command; either manual specification is required (increasing operational burden and prone to errors), or the default section can only be used (potentially leading to localization errors). Therefore, establishing an automatic mapping relationship between sequences and sections through protocol tags enables the system to accurately read the matching target section spatial parameters based on the protocol tag of the current sequence, achieving intelligent matching.

[0074] S26: Configure instruction parameters based on the target section spatial parameters and the cardiac target diagnostic scan sequence to generate magnetic resonance scan instructions for driving the underlying scan.

[0075] The bottom-level scan driver refers to the lowest-level control system in a magnetic resonance imaging (MRI) scanner. It is responsible for receiving scan commands and directly controlling hardware such as radio frequency coils and gradient coils to perform specific signal excitation and acquisition operations.

[0076] For example, the spatial parameters of the target section exist only in an abstract geometric form, namely, an infinitely extending ideal plane defined by the plane's normal vector and the coordinates of its center point. The underlying scan driver of an MRI scanner cannot directly understand this abstract geometric expression; it requires specific, executable positioning line parameters. For instance, the spatial parameters of a four-chamber view only define an infinitely extending two-dimensional plane, but in actual scanning, multiple layers of images need to be acquired near this plane location, such as a slice thickness of six millimeters, a slice spacing of four millimeters, and a total of ten slices. Furthermore, specific parameters such as the scanning field of view, repetition time, and echo time need to be determined based on the sequence type. Therefore, by fusing the abstract target section spatial parameters with the protocol tags of the diagnostic sequence, standard MRI scan instructions that the underlying driver can directly execute are generated, ensuring that the device can accurately perform data acquisition.

[0077] S308: Extract the planar normal vectors of each of the standard cardiac anatomical sections; call a preset multi-plane reconstruction algorithm to perform three-dimensional interpolation and resampling operations on the three-dimensional volume data along each of the planar normal vectors to reconstruct a two-dimensional planar preview image corresponding to each of the standard cardiac anatomical sections; display the two-dimensional planar preview image in the user interface.

[0078] The preset multi-plane reconstruction algorithm is a pre-configured computation method in the system, used to reconstruct two-dimensional planar images in arbitrary directions from three-dimensional volume data. This algorithm takes three-dimensional volume data as input, and calculates the signal intensity value of each pixel on the plane based on the plane normal vector and center point coordinates of the target plane through three-dimensional interpolation and resampling operations, thereby generating the corresponding two-dimensional image. Three-dimensional interpolation uses the signal values ​​of surrounding voxels to estimate the signal intensity at the target location, and resampling discretizes continuous spatial points into a regular pixel grid, which can be directly invoked without user adjustment during runtime.

[0079] For example, the spatial parameters output by deep learning models for scanning localization are merely a set of abstract mathematical values. Radiologists cannot intuitively judge from these values ​​whether the automatic localization results are accurate or whether the sections include key anatomical structures, such as whether the four-chamber view simultaneously shows the left and right ventricles and atria, or whether the left ventricular short-axis view is perpendicular to the heart's long axis and completely covers the left ventricle from the apex to the base. Without a preview image display and confirmation mechanism, physicians will have to directly trust the automatic localization results and initiate diagnostic sequence scanning. If the localization is inaccurate, the entire examination may fail and require a rescan. Therefore, multiplanar reconstruction technology transforms abstract planar parameters into visualized two-dimensional images, providing physicians with an intuitive means to verify the localization results.

[0080] In one feasible implementation, the standard cardiac anatomical sections include axial sections, coronal sections, sagittal sections, left ventricular short-axis sections, two-chamber sections, three-chamber sections, four-chamber sections, and right ventricular outflow tract sections; the scanning positioning spatial parameters include the plane normal vector and plane center coordinates of each of the standard cardiac anatomical sections in three-dimensional space.

[0081] For example, standard cardiac anatomical sections constitute a complete standard viewing system for clinical cardiac magnetic resonance imaging (MRI), used for global spatial localization and preliminary anatomical observation. The left ventricular short-axis section is the core section for assessing left ventricular volume and ejection fraction; the four-chamber section is used to assess the size and function of the four chambers; the two-chamber and three-chamber sections are used to assess ventricular wall motion in specific myocardial segments; and the right ventricular outflow tract section is used to assess the morphology of the right ventricular outflow tract. The plane normal vector and the coordinates of the plane center position are the simplest set of parameters required to uniquely determine a plane in three-dimensional space. This parameterized expression avoids redundant data storage and can be directly parsed and used by subsequent multi-plane reconstruction algorithms and scan command generation modules, achieving an efficient mapping from anatomical definition to mathematical expression.

[0082] In one feasible implementation, when the multi-scan plane resolution processing fails, the automatic positioning process is suspended and the manual positioning scanning process is triggered; in response to the manual positioning line adjustment operation performed by the user through the graphical control software, the manually determined reference anatomical plane parameters are obtained; the reference anatomical plane parameters are updated and saved so that when performing cardiac target diagnostic scanning sequences, the reference anatomical plane parameters are called to generate magnetic resonance scanning instructions for scanning control processing.

[0083] The reference anatomical plane parameters refer to the spatial parameters of the anatomical plane determined by the radiologist through graphical control software after manually adjusting the positioning lines when the automatic positioning process fails. These parameters include the plane normal vector and the coordinates of the plane center point, and their mathematical form is consistent with the automatically generated scanning positioning spatial parameters.

[0084] For example, deep learning models may encounter extreme situations in real-world clinical applications where the training data is not fully covered. For instance, a patient with a severe congenital heart defect may prevent the model from reliably predicting the spatial parameters of standard cardiac anatomical sections. Without a fallback mechanism, the system would be unable to complete the examination, causing scan interruption. Therefore, this step provides a manual fallback solution for automated localization failures: allowing radiologists to manually adjust the localization lines using graphical control software to determine the baseline anatomical plane parameters, and saving these manually determined parameters for subsequent diagnostic sequences. This mechanism leverages the efficiency advantages of automation while retaining the flexibility and robustness of manual intervention, ensuring the system can complete examinations in various clinical scenarios.

[0085] Please see Figure 5 , Figure 5 This is a flowchart illustrating another embodiment of the cardiac scanning localization method proposed in this specification. Specifically: The Cardiac_Scout sequence is loaded, and it is determined whether assisted localization is enabled. If assisted localization is enabled and the call is successful, the Cardiac_SmartScout sequence is automatically scanned and invoked. If assisted localization is not enabled or the call fails, the manual localization process begins. During the Cardiac_SmartScout sequence stage, if the automatic invocation is successful, multi-plane regression analysis is performed to generate scan localization spatial parameters, and the results are reused. If the invocation fails, the manual localization process is triggered again, and the user manually adjusts the localization lines through graphical control software to determine the reference anatomical plane parameters. These parameters are then updated and saved to replace the automatically generated scan localization spatial parameters for subsequent invocation and reuse of cardiac target diagnostic scan sequences, thus ensuring both automation efficiency and robust manual fallback.

[0086] In the embodiments of this specification, an initial positioning scan sequence is executed to perform a positioning scan, generating initial scanning positioning parameters for the anatomical positioning scan sequence. The anatomical positioning scan sequence is then executed to acquire and reconstruct three-dimensional volume data of the heart region for analysis and processing, generating scanning positioning spatial parameters for multiple standard cardiac anatomical sections. These scanning positioning spatial parameters are then invoked to generate magnetic resonance imaging (MRI) scan commands during the execution of the cardiac target diagnostic scan sequence. By employing the above technical solution, the initial scanning positioning parameters generated by the initial positioning scan sequence guide the anatomical positioning scan sequence to acquire high-resolution three-dimensional volume data, thereby generating scanning positioning spatial parameters for standard cardiac anatomical sections. This eliminates subjective biases from manual operation and improves the automation level and consistency of scan positioning.

[0087] The following will combine Figure 6 This specification provides a detailed description of the cardiac scanning and positioning device provided in the embodiments. It should be noted that... Figure 6 The cardiac scanning and positioning device shown is used to perform the functions described in this manual. Figures 1-5 The methods shown in the embodiments are illustrated for ease of explanation, showing only the parts related to the embodiments of this specification. For specific technical details not disclosed, please refer to this specification. Figures 1-5 The example shown.

[0088] Please see Figure 6 This diagram illustrates the structure of a cardiac scanning and positioning device according to an embodiment of this specification. The cardiac scanning and positioning device 1 can be implemented as all or part of a user terminal through software, hardware, or a combination of both. According to some embodiments, the cardiac scanning and positioning device 1 includes a scanning module 11, an acquisition module 12, and a parsing module 13. The scanning module 11 is used to perform an initial localization scanning sequence to perform localization scanning, so as to obtain a triorthogonal anatomical image of the heart, and to perform spatial region of interest parsing on the triorthogonal anatomical image to generate initial scanning localization parameters for the anatomical localization scanning sequence. The acquisition module 12 is used to perform the anatomical positioning scan sequence based on the initial scan positioning parameters to acquire and reconstruct high-resolution three-dimensional volume data of the heart region. The parsing module 13 is used to perform multi-scan plane parsing processing based on the three-dimensional volume data to generate scanning positioning spatial parameters for multiple standard cardiac anatomical sections. These scanning positioning spatial parameters are then used to generate magnetic resonance scanning commands for scanning control processing when executing cardiac target diagnostic scanning sequences.

[0089] In one feasible implementation, the step of generating initial scanning localization parameters for the anatomical localization scanning sequence by spatial region of interest parsing of the triorthogonal anatomical images includes: Target recognition is performed on the cardiac region in the triorthogonal anatomical images to obtain the spatial location information of the cardiac region after filtering out background chest tissue; The three-dimensional target bounding box of the heart region in three-dimensional space is determined based on spatial location information; Initial scan positioning parameters for the anatomical positioning scan sequence are generated based on the spatial geometric features of the three-dimensional target bounding box.

[0090] In one feasible implementation, generating initial scan localization parameters for the anatomical localization scan sequence based on the spatial geometric features of the three-dimensional target bounding box includes: Extract the three-dimensional coordinates of the center point of the three-dimensional target bounding box; The three-dimensional target bounding box is calculated in three orthogonal directions in the human anatomical coordinate system, including the front-back direction, the head-to-toe direction, and the left-right direction. A spatial reference coordinate system is constructed based on the three-dimensional coordinates of the center point and the three orthogonal directions. Scanning field parameters and positioning line parameters are generated based on the spatial reference coordinate system. Initial scanning positioning parameters for the anatomical positioning scanning sequence are obtained based on the scanning field parameters and positioning line parameters.

[0091] In one feasible implementation, the step of performing multi-scan plane analytical processing based on the three-dimensional volume data to generate scanning positioning spatial parameters for multiple standard cardiac anatomical sections includes: The high-resolution type of 3D volume data is input into a pre-trained deep learning model; The deep learning model performs multiplanar spatial regression calculations on each standard cardiac anatomical section in three-dimensional space to obtain the scanning positioning spatial parameters of each standard cardiac anatomical section.

[0092] In one feasible implementation, the standard cardiac anatomical sections include axial sections, coronal sections, sagittal sections, left ventricular short-axis sections, two-chamber sections, three-chamber sections, four-chamber sections, and right ventricular outflow tract sections; the scanning positioning spatial parameters include the plane normal vector and plane center coordinates of each of the standard cardiac anatomical sections in three-dimensional space.

[0093] In one feasible implementation, after obtaining the scanning positioning spatial parameters of each of the standard cardiac anatomical sections, the method further includes: Extract the planar normal vectors of each of the aforementioned standard cardiac anatomical sections; A preset multi-plane reconstruction algorithm is invoked to perform three-dimensional interpolation and resampling operations on the three-dimensional volume data along the normal vectors of each plane, so as to reconstruct a two-dimensional planar preview image corresponding to each standard cardiac anatomical section. The two-dimensional plane preview image is displayed in the user interface.

[0094] In one feasible implementation, the step of calling the scan positioning spatial parameters to generate magnetic resonance scan commands for scan control processing during the execution of a cardiac target diagnostic scan sequence includes: Save the scanning positioning spatial parameters of multiple standard cardiac anatomical sections; When a command to start the cardiac target diagnostic scanning sequence is received, the scanning positioning spatial parameters are read as target section spatial parameters according to the protocol tag of the cardiac target diagnostic scanning sequence. Based on the target section spatial parameters and the cardiac target diagnostic scanning sequence, instruction parameters are configured to generate magnetic resonance scanning instructions for driving the underlying scan.

[0095] In one feasible implementation, after performing multi-scan plane analysis processing based on the three-dimensional volume data, the method further includes: When the multi-scan plane parsing process fails, the automatic positioning process is suspended and the manual positioning scanning process is triggered. In response to manual positioning line adjustment operations performed by the user through graphical control software, the reference anatomical plane parameters determined manually are obtained; The reference anatomical plane parameters are updated and saved so that when performing cardiac target diagnostic scanning sequences, the scan positioning spatial parameters are called to generate magnetic resonance scanning commands for scan control processing.

[0096] It should be noted that the cardiac scanning positioning device provided in the above embodiments is only illustrated by the division of the above functional modules when performing the cardiac scanning positioning method. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the cardiac scanning positioning device and the cardiac scanning positioning method embodiments provided in the above embodiments belong to the same concept, and the implementation process is detailed in the method embodiments, which will not be repeated here.

[0097] The example numbers in this specification are for descriptive purposes only and do not represent the superiority or inferiority of the examples.

[0098] This specification also provides a computer storage medium that can store multiple instructions adapted to be loaded and executed by a processor as described above. Figures 1-5 The cardiac scanning localization method described in the illustrated embodiment can be found in the following documentation for its specific execution process. Figures 1-5 The specific details of the illustrated embodiments will not be elaborated here.

[0099] This specification also provides a computer program product that stores at least one instruction, said at least one instruction being loaded and executed by the processor as described above. Figures 1-5 The cardiac scanning localization method described in the illustrated embodiment can be found in the following documentation for its specific execution process. Figures 1-5 The specific details of the illustrated embodiments will not be elaborated here.

[0100] Please refer to Figure 7 This is a structural block diagram of an electronic device provided in an embodiment of this specification. The electronic device in this specification may include one or more of the following components: a processor 1010, a memory 1020, an input device 1030, an output device 1040, and a bus 1050. The processor 1010, memory 1020, input device 1030, and output device 1040 may be connected to each other via the bus 1050.

[0101] Processor 1010 may include one or more processing cores. Processor 1010 connects to various parts of the electronic device using various interfaces and lines, and performs various functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 1020, and by calling data stored in memory 1020. Optionally, processor 1010 may be implemented using at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). Processor 1010 may integrate one or more of a central processing unit (CPU), graphics processing unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the displayed content; and the modem handles wireless communication. It is understood that the modem may also not be integrated into processor 1010 and may be implemented separately through a communication chip.

[0102] The memory 1020 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 1020 may include non-transitory computer-readable storage medium. The memory 1020 may be used to store instructions, programs, code, code sets, or instruction sets.

[0103] The input device 1030 is used to receive input instructions or data, and includes, but is not limited to, a keyboard, mouse, camera, microphone, or touch device. The output device 1040 is used to output instructions or data, and includes, but is not limited to, a display device and a speaker. In this embodiment, the input device 1030 can be a temperature sensor for acquiring the operating temperature of the electronic device. The output device 1040 can be a speaker for outputting audio signals.

[0104] In addition, those skilled in the art will understand that the structure of the electronic device shown in the above figures does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements. For example, the electronic device may also include radio frequency circuits, input units, sensors, audio circuits, wireless fidelity (WIFI) modules, power supplies, Bluetooth modules, etc., which will not be described in detail here.

[0105] In the embodiments of this specification, the entity executing each step may be the electronic device described above. Optionally, the entity executing each step may be the operating system of the electronic device.

[0106] exist Figure 7 In the electronic device, the processor 1010 can be used to call a program stored in the memory 1020 and execute it to implement the cardiac scanning localization method as described in the various method embodiments of this specification.

[0107] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory, or random access memory, etc.

[0108] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals involved in the embodiments of this specification are all authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the object characteristics, interactive behavior characteristics, and user information involved in this specification were all obtained under full authorization.

[0109] The above-disclosed embodiments are merely preferred embodiments of this specification and should not be construed as limiting the scope of this specification. Therefore, any equivalent variations made in accordance with the claims of this specification shall still fall within the scope of this specification.

Claims

1. A cardiac scanning localization method, characterized in that, The method, which includes an initial localization scan sequence and an anatomical localization scan sequence, comprises: An initial localization scan sequence is executed to perform a localization scan to obtain a triorthogonal anatomical image of the heart, and spatial region of interest is analyzed on the triorthogonal anatomical image to generate initial scan localization parameters for the anatomical localization scan sequence. Based on the initial scanning and positioning parameters, the anatomical positioning scanning sequence is executed to acquire and reconstruct high-resolution three-dimensional volume data of the heart region. Based on the three-dimensional volume data, multi-scan plane analysis processing is performed to generate multiple standard cardiac anatomical sections and scanning positioning spatial parameters. These scanning positioning spatial parameters are then used to generate magnetic resonance scanning commands for scanning control processing when executing cardiac target diagnostic scanning sequences.

2. The method according to claim 1, characterized in that, The step of generating initial scanning localization parameters for the anatomical localization scanning sequence by spatial region of interest parsing of the triorthogonal anatomical images includes: Target recognition is performed on the cardiac region in the triorthogonal anatomical images to obtain the spatial location information of the cardiac region after filtering out background chest tissue; The three-dimensional target bounding box of the heart region in three-dimensional space is determined based on spatial location information; Initial scan positioning parameters for the anatomical positioning scan sequence are generated based on the spatial geometric features of the three-dimensional target bounding box.

3. The method according to claim 2, characterized in that, The generation of initial scan localization parameters for the anatomical localization scan sequence based on the spatial geometric features of the three-dimensional target bounding box includes: Extract the three-dimensional coordinates of the center point of the three-dimensional target bounding box; The three-dimensional target bounding box is calculated in three orthogonal directions in the human anatomical coordinate system, including the front-back direction, the head-to-toe direction, and the left-right direction. A spatial reference coordinate system is constructed based on the three-dimensional coordinates of the center point and the three orthogonal directions. Scanning field parameters and positioning line parameters are generated based on the spatial reference coordinate system. Initial scanning positioning parameters for the anatomical positioning scanning sequence are obtained based on the scanning field parameters and positioning line parameters.

4. The method according to claim 1, characterized in that, The process of performing multi-scan plane analysis based on the three-dimensional volume data to generate scanning positioning spatial parameters for multiple standard cardiac anatomical sections includes: The high-resolution type of 3D volume data is input into a pre-trained deep learning model; The deep learning model performs multiplanar spatial regression calculations on each standard cardiac anatomical section in three-dimensional space to obtain the scanning positioning spatial parameters of each standard cardiac anatomical section.

5. The method according to claim 4, characterized in that, The standard cardiac anatomical sections include axial sections, coronal sections, sagittal sections, left ventricular short-axis sections, two-chamber sections, three-chamber sections, four-chamber sections, and right ventricular outflow tract sections; the scanning positioning spatial parameters include the plane normal vector and plane center coordinates of each of the standard cardiac anatomical sections in three-dimensional space.

6. The method according to claim 4, characterized in that, After obtaining the scanning positioning spatial parameters of each of the standard cardiac anatomical sections, the method further includes: Extract the planar normal vectors of each of the aforementioned standard cardiac anatomical sections; A preset multi-plane reconstruction algorithm is invoked to perform three-dimensional interpolation and resampling operations on the three-dimensional volume data along the normal vectors of each plane, so as to reconstruct a two-dimensional planar preview image corresponding to each standard cardiac anatomical section. The two-dimensional plane preview image is displayed in the user interface.

7. The method according to claim 1, characterized in that, The step of generating magnetic resonance scanning instructions by calling the scan positioning spatial parameters during the execution of a cardiac target diagnostic scanning sequence for scan control processing includes: Save the scanning positioning spatial parameters of multiple standard cardiac anatomical sections; When a command to start the cardiac target diagnostic scanning sequence is received, the scanning positioning spatial parameters are read as target section spatial parameters according to the protocol tag of the cardiac target diagnostic scanning sequence. Based on the target section spatial parameters and the cardiac target diagnostic scanning sequence, instruction parameters are configured to generate magnetic resonance scanning instructions for driving the underlying scan.

8. The method according to claim 1, characterized in that, After performing multi-scan plane analysis processing based on the three-dimensional volume data, the process further includes: When the multi-scan plane parsing process fails, the automatic positioning process is suspended and the manual positioning scanning process is triggered. In response to manual positioning line adjustment operations performed by the user through graphical control software, the reference anatomical plane parameters determined manually are obtained; The reference anatomical plane parameters are updated and saved so that when performing cardiac target diagnostic scanning sequences, the scan positioning spatial parameters are called to generate magnetic resonance scanning commands for scan control processing.

9. A cardiac scanning and positioning device, characterized in that, The device, equipped with an initial localization scan sequence and an anatomical localization scan sequence, includes: The scanning module is used to perform an initial localization scanning sequence to obtain a triorthogonal anatomical image of the heart, and to perform spatial region of interest parsing on the triorthogonal anatomical image to generate initial scanning localization parameters for the anatomical localization scanning sequence. The acquisition module is used to perform the anatomical positioning scan sequence based on the initial scan positioning parameters to acquire and reconstruct high-resolution three-dimensional volume data of the heart region. The parsing module is used to perform multi-scan plane parsing processing based on the three-dimensional volume data to generate scanning positioning spatial parameters for multiple standard cardiac anatomical sections. These scanning positioning spatial parameters are then used to generate magnetic resonance scanning commands for scanning control processing when executing cardiac target diagnostic scanning sequences.

10. A computer storage medium, characterized in that, The computer storage medium stores a plurality of instructions, which are adapted to be loaded by a processor and executed as method steps as claimed in any one of claims 1 to 8.

11. An electronic device, characterized in that, include: A processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and executed the method steps as claimed in any one of claims 1 to 8.

Images