Ultrasonic image and mode pattern space mapping method and system based on key point constraint

Abstract

The invention relates to an ultrasonic image and mode pattern space mapping method and system based on key point constraint, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a to-be-mapped ultrasonic image and a corresponding anatomical structure segmentation result; determining a first group of key points in the ultrasonic image coordinate system; acquiring a preset mode pattern and a second group of key points, and establishing a geometric transformation model from the ultrasonic image coordinate system to the mode pattern coordinate system according to the first group of key points and the second group of key points; and mapping the anatomical structure contour, the lesion contour and the pixel grid to the mode pattern coordinate system by using the geometric transformation model. The problems that in the prior art, it is difficult to keep the pattern pattern consistent with the real uterus shape of a patient, and the focus position is inaccurate are solved, accurate and automatic mapping of the anatomical structure and the focus contour is achieved, the quality evaluation and robust processing capacity is achieved, and a standardized and quantifiable comparison ultrasonic structure report is generated.

Description

Technical Field

[0001] This application relates to the fields of medical image processing and medical visualization technology, and in particular to a method, system, storage medium and computer program product for spatial mapping of ultrasound images and pattern diagrams based on key point constraints. Background Technology

[0002] Schematic diagrams are commonly used in gynecological ultrasound reports to illustrate the relative positions and spatial relationships of structures such as the uterus, endometrium, and fibroids. Compared to directly annotating on the ultrasound grayscale image, schematic diagrams have advantages such as simple structure, uniform scale, and strong information comparability, making it easier for doctors to quickly understand the location of lesions and conduct follow-up comparisons.

[0003] However, existing schematic diagrams mostly rely on manual drawing or simple scaling / translation of fixed templates, making it difficult to maintain consistency with the actual uterine morphology of patients. On the one hand, the uterine contour exhibits significant individual differences and is affected by factors such as probe posture, section offset, and bladder fullness. On the other hand, the representation of fibroid / polyp locations requires establishing a stable reference with the uterine cavity (the central axis determined by the endometrium). Traditional rigid or weak constraint alignment often leads to systematic shifts in lesion locations on the schematic diagram, affecting clinical use. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, system, computer device, storage medium, and computer program product for spatial mapping of ultrasound images and pattern diagrams based on key point constraints to address the above-mentioned technical problems.

[0005] In a first aspect, this application provides a spatial mapping method for ultrasound images and pattern diagrams based on key point constraints, the method comprising: Obtain the ultrasound image to be mapped and the segmentation results of the corresponding anatomical structures in the ultrasound image; In the ultrasound image coordinate system, identify the first set of key points used to describe the anatomical morphology; Obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics; A geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system is established based on the first set of key points and the second set of key points. Using the geometric transformation model, the anatomical structure outline, lesion outline, and pixel grid are mapped to the pattern diagram coordinate system to generate a standardized structural representation.

[0006] In one embodiment, the first set of key points includes at least: the cephalic endpoint, the pedunculated endpoint, the midpoint of the anterior wall, and the midpoint of the posterior wall of the uterine contour; and / or, the cephalic endpoint, the pedunculated endpoint, the midpoint of the anterior wall, and the midpoint of the posterior wall of the endometrial contour. The midpoints of the anterior and posterior walls are determined based on the division of the intima's central axis.

[0007] In one implementation, before establishing the geometric transformation model, the first set of key points are normalized to transform their coordinates to a preset normalized coordinate system.

[0008] In one embodiment, the geometric transformation model is a linear transformation model, which is obtained by minimizing the error between all key point pairs; the linear transformation model includes an affine transformation model or a perspective transformation model.

[0009] In one embodiment, the geometric transformation model is a thin-plate spline transformation model, which minimizes the overall bending energy while accurately matching key points.

[0010] In one implementation, when solving the linear system of the thin plate spline transformation model, a regularization parameter is introduced to suppress overfitting and balance the key point matching accuracy with deformation smoothness.

[0011] In one embodiment, the method further includes: Calculate the quality assessment indicators after mapping; When the indicator does not meet the preset threshold, a rollback mechanism is triggered, which includes switching to a simpler geometric transformation model or adjusting the model parameters.

[0012] In one embodiment, the method further includes: The geometric transformation model is used to reverse map the pixel grid of the ultrasound image, and the deformed image with the same viewpoint as the pattern image is generated by interpolation resampling; the interpolation method includes bilinear interpolation or bicubic interpolation.

[0013] In one embodiment, the geometric transformation model can be automatically selected based on the number of key point pairs, fitting error, or deformation energy index. Optional models include affine transformation, perspective transformation, and thin plate spline transformation.

[0014] Secondly, this application also provides a spatial mapping system for ultrasound images and pattern diagrams based on key point constraints, the system comprising: The acquisition module is used to acquire the ultrasound image to be mapped and the anatomical structure segmentation results corresponding to the ultrasound image; The first determining module is used to determine the first set of key points in the ultrasound image coordinate system for describing the anatomical morphology; The second determining module is used to obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics. The model building module is used to build a geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system based on the first set of key points and the second set of key points. The mapping module is used to map the anatomical structure outline, lesion outline, and pixel grid to the pattern diagram coordinate system using the geometric transformation model, thereby generating a standardized structural representation.

[0015] Thirdly, this application also provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps: Obtain the ultrasound image to be mapped and the segmentation results of the corresponding anatomical structures in the ultrasound image; In the ultrasound image coordinate system, identify the first set of key points used to describe the anatomical morphology; Obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics; A geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system is established based on the first set of key points and the second set of key points. Using the geometric transformation model, the anatomical structure outline, lesion outline, and pixel grid are mapped to the pattern diagram coordinate system to generate a standardized structural representation.

[0016] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps: Obtain the ultrasound image to be mapped and the segmentation results of the corresponding anatomical structures in the ultrasound image; In the ultrasound image coordinate system, identify the first set of key points used to describe the anatomical morphology; Obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics; A geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system is established based on the first set of key points and the second set of key points. Using the geometric transformation model, the anatomical structure outline, lesion outline, and pixel grid are mapped to the pattern diagram coordinate system to generate a standardized structural representation.

[0017] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps: Obtain the ultrasound image to be mapped and the segmentation results of the corresponding anatomical structures in the ultrasound image; In the ultrasound image coordinate system, identify the first set of key points used to describe the anatomical morphology; Obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics; A geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system is established based on the first set of key points and the second set of key points. Using the geometric transformation model, the anatomical structure outline, lesion outline, and pixel grid are mapped to the pattern diagram coordinate system to generate a standardized structural representation.

[0018] The aforementioned method, system, computer equipment, storage medium, and computer program product for spatial mapping of ultrasound images and schematic diagrams based on key point constraints acquire the ultrasound image to be mapped and the corresponding anatomical structure segmentation results in the ultrasound image coordinate system; determine a first set of key points in the ultrasound image coordinate system to describe the anatomical morphology; acquire a preset schematic diagram and a second set of key points in the preset schematic diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in anatomical semantics; establish a geometric transformation model from the ultrasound image coordinate system to the schematic diagram coordinate system based on the first set of key points and the second set of key points; and use the geometric transformation model to map the anatomical structure contour, lesion contour, and pixel grid to the schematic diagram coordinate system to generate a standardized structural representation. This solves the problems in existing technologies where schematic diagrams are difficult to reconcile with the patient's actual uterine morphology and lesion locations are inaccurate, achieving precise and automatic mapping of anatomical structures and lesion contours, and possessing quality assessment and robust processing capabilities to generate standardized, quantifiable, and comparable ultrasound structural reports. Attached Figure Description

[0019] Figure 1 A flowchart of a method for spatial mapping of ultrasound images and pattern diagrams based on key point constraints, provided in one embodiment of the present invention; Figure 2 This is an architecture diagram of an ultrasound image and pattern graph spatial mapping system based on key point constraints, provided in one embodiment of the present invention. Figure 3 This is a schematic diagram of a real gynecological ultrasound image input.

[0020] Figure 4 This is a pre-set standardized uterine and endometrial pattern template. Figure 5The visualization of 8 key points is shown; Figure 6 The results of mapping using an affine transformation model are shown. Figure 7 The results of mapping using a perspective transformation model are shown; Figure 8 The results are shown after mapping using the thin-plate spline transformation model. Figure 9 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0022] The purpose of this invention is to provide a spatial mapping method based on key point constraints, which realizes morphological alignment and structural mapping between ultrasound images and pattern diagrams through the correspondence of key points and optional thin plate spline deformation models, taking into account both global consistency and local deformation expression.

[0023] The following describes in detail the ultrasonic image and pattern diagram spatial mapping method based on key point constraints of the present invention.

[0024] Please refer to Figure 1 This application provides a method for spatial mapping of ultrasound images and pattern diagrams based on key point constraints, the method comprising: S100, acquire the ultrasound image to be mapped and the anatomical structure segmentation result corresponding to the ultrasound image.

[0025] In this invention, the ultrasound image to be mapped is first acquired, and the ultrasound image is segmented into anatomical structures to obtain the anatomical structure contour of the ultrasound image.

[0026] This step aims to obtain raw data. The input includes a gynecological ultrasound image to be processed (typically a midline sagittal section of the uterus, such as...). Figure 3 The image is shown below, along with the anatomical structure segmentation results obtained after processing the image using a pre-trained semantic segmentation model (such as U-Net). The segmentation results are typically provided as a mask or a set of contour points, containing at least the uterine region, the endometrial region, and, if lesions (such as fibroids or polyps) are present, their regions. This can be represented as a set of structures. ,in This refers to the uterine region. For the endometrial region, This represents the k-th fibroid / polyp region.

[0027] However, images acquired under different devices and settings have varying resolutions, and directly using pixel coordinates for calculations would cause the model to heavily rely on absolute numerical scales. Therefore, normalization preprocessing is necessary. Specifically, the mask boundaries are extracted as point sets and polygons are constructed. After topological repair (such as self-intersection repair and hole filtering), the contour point set is obtained. For each point p = (x, y), normalization is performed, transforming its coordinates to a preset normalized coordinate system. For example, the coordinates of key points are made to fall into the normalized coordinate system of [-1, 1], so as to reduce the impact of resolution changes on the transformation solution.

[0028] The specific normalization formula can be:

[0029] in It can be obtained from the minimum bounding rectangle of the uterine outline, ensuring a consistent numerical scale under different resolutions and scaling conditions.

[0030] S200, in the ultrasound image coordinate system, identifies the first set of key points used to describe the anatomical morphology.

[0031] Key points on the ultrasound side can be derived from the contour and the central axis of the intima, or from a self-key point extraction model based on deep learning, or from manually selected points.

[0032] This step involves extracting feature points with stable anatomical semantics from the segmented anatomical structures. The first set of key points includes at least: the cephalic endpoint, foot endpoint, anterior wall midpoint, and posterior wall midpoint of the uterine contour; and / or, the cephalic endpoint, foot endpoint, anterior wall midpoint, and posterior wall midpoint of the endometrial contour. Crucially, the anterior and posterior wall midpoints are determined based on the endometrial midline.

[0033] The specific construction method is as follows (which can be combined with...) Figure 5 understand): Lateral point of uterine head The cephalic intersection point extending along the midline of the endometrium to the uterine boundary.

[0034] Uterine foot lateral point The foot-side intersection of the endometrial midline extending to the uterine boundary.

[0035] Anterior wall point of the uterus Using the central axis as the cutting curve, the uterine boundary is divided into two arcs, and the midpoint of the arc length is taken as the anterior wall / posterior wall.

[0036] Posterior wall of the uterus Same as above.

[0037] Endometrial head / foot / anterior / posterior points The same strategy was used to obtain the result at the endometrial boundary.

[0038] Thus, we have identified eight key points strongly correlated with the uterus's own anatomical structure (endometrial midline) (such as...). Figure 5 As shown in the image, their definitions are independent of the rotation angle of the uterus in the image and are stable and reliable.

[0039] S300, Obtain a preset pattern diagram and a second set of key points from the preset pattern diagram. The second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics.

[0040] From the preset pattern template ( Figure 4 We then retrieve a predefined second set of key points that correspond one-to-one with the aforementioned eight points in anatomical semantics. This establishes a precise correspondence table, which forms the foundation of the entire spatial mapping. Similarly, the coordinates of key points in the pattern diagram are typically normalized based on their canvas to ensure they are in a comparable numerical space with the ultrasound-side key points.

[0041] S400, establish a geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system based on the first set of key points and the second set of key points.

[0042] This step supports a variety of geometric transformation models. The system can automatically select the appropriate model based on the number of key point pairs, fitting error, or deformation energy index. Optional models include affine transformation, perspective transformation, and thin plate spline transformation.

[0043] Number of control points: at least 3 pairs for affine lines, at least 4 pairs for perspective lines, and ≥6 pairs for TPS to enhance stability.

[0044] Deformation intensity: TPS is preferred when there are large differences in the shape of the uterus or when there is local nonlinear deformation.

[0045] Numerical stability and speed: Affine is the most stable and fastest; TPS requires solving linear systems, so it is recommended to introduce regularization and cache the decomposition results.

[0046] Given a set of transformation models, including affine transformations, perspective transformations, and thin-plate spline TPS transformations, the system selects one model from the set each time and solves for the transformation parameters from the ultrasonic coordinate system to the model diagram coordinate system based on control point pairs. The system then measures the mapping quality of the selected transformation model. Quality assessment methods include indicators such as control point reprojection error, contour IoU, boundary Hausdorff distance, and TPS bending energy. If the model meets the preset quality conditions, it is selected; otherwise, another model is selected from the set. If none of the transformation models meet the preset conditions, a failure message is displayed.

[0047] In this invention, the affine transformation model and the perspective transformation model are linear transformation models, obtained by minimizing the error between all keypoint pairs. Affine transformation is suitable for handling overall image deformations such as translation, rotation, scaling, and shearing. Its model is simple, solves quickly, and provides a unique and stable solution, making it ideal for situations where the uterine shape does not differ significantly from a standard template. Perspective transformation (or homography transformation) adds perspective distortion capabilities to the affine transformation, better simulating and correcting slight projection distortions caused by the ultrasound probe not being perpendicular to the human body surface. These two models constitute a basic and efficient global alignment tool.

[0048] Thin-plate spline transformation minimizes overall bending energy while accurately matching keypoints. When solving the linear system of the thin-plate spline transformation model, a regularization parameter is introduced to suppress overfitting and balance keypoint matching accuracy with deformation smoothness. The reason for introducing regularization is that in practical applications, the point positions output by automatic segmentation or keypoint detection algorithms may contain slight errors or noise. Without constraints, the thin-plate spline model may produce unnatural and clinically unreliable localized drastic deformations (i.e., overfitting) in an attempt to "overfit" these noisy points perfectly. The introduction of the regularization parameter λ is equivalent to adding a penalty for the "unsmoothness" of the transformation to the optimization objective. By adjusting the value of λ, a flexible trade-off can be struck between "strictly matching keypoints" and "maintaining overall deformation smoothness," thereby greatly improving the model's noise resistance and the visual plausibility of the output results while ensuring accuracy.

[0049] In this invention, the ultrasound image coordinate system is derived from the pixel coordinates of a real ultrasound image or its normalized coordinates. Each point is denoted as... .

[0050] Model diagram coordinate system: derived from the preset model diagram canvas coordinate system or its normalized coordinate system. Each point is denoted as... . This represents the mapped point (which falls within the pattern diagram coordinate system).

[0051] Affine transformation model The affine transformation matrix is ​​solved using the least squares method, minimizing the error between all keypoint pairs:

[0052] Perspective Transformation Model , H is the normalization factor, and H is the perspective transformation matrix.

[0053] Thin plate spline TPS:

[0054] in Thin Plate Spline (TPS) mapping function, representing a non-rigid geometric transformation from the ultrasound image coordinate system to the pattern diagram coordinate system.

[0055] The coordinates of the points to be mapped on the ultrasound side (which can be the outline points or pixels of the uterus / endometrium / lesion) are usually normalized before use in S100.

[0056] The global affine term parameters of TPS. Used to describe global alignment such as translation, rotation, scaling, and shearing (obtained by fitting control points together with nonlinear terms).

[0057] N: Number of control points (i.e., number of key point pairs).

[0058] The position of the j-th control point in the ultrasonic coordinate system is usually taken directly from the ultrasonic side key point. .

[0059] The nonlinear weight vector corresponding to the j-th control point determines the magnitude and direction of the contribution of that control point to the local deformation.

[0060] :point To the control point The Euclidean distance.

[0061] The radial basis kernel function of TPS is used to generate smooth nonlinear deformations; It is a very small positive number, used to avoid the instability of the logarithmic term values ​​when r=0.

[0062] It should be noted that the TPS transformation of thin plate splines has a wider range of applications.

[0063] S500: Using the geometric transformation model, the anatomical structure outline, lesion outline, and pixel grid are mapped to the pattern diagram coordinate system to generate a standardized structural representation.

[0064] Map contour points, lesion points, or the entire image to the pattern coordinate system and output the result. For contours / key points: apply transformations directly. The output includes: polygons in the pattern map coordinates, key points, rendered image, and transformation parameters (matrix / control points + weights).

[0065] Specifically, after solving the transformation model, it can be applied to various types of data to generate standardized output: 1. Contour mapping: Substitute each point from the contour points of the uterus, endometrium, and lesions into... This directly yields its coordinates on the pattern diagram. This is the most crucial step in generating the structural representation.

[0066] 2. Image rendering (pixel-level mapping): To generate visually uniform and standardized ultrasound images, the entire image needs to be transformed. A reverse mapping strategy is employed to avoid holes: for each target pixel on the pattern drawing canvas, its corresponding source image coordinates are calculated. Since the source image coordinates are not integers, bicubic interpolation is used to sample grayscale values ​​from the original image and assign them to the target pixels. This process generates... Figure 6-8 The image shown is exactly the same as the template's perspective. Figure 8 The mapping effect after using thin plate spline TPS transformation is shown, and it can be seen that the contour fits the template well and the fibroid is accurately located.

[0067] 3. Quality assessment and robust rollback (system safety net): To ensure reliable output, the system also needs to calculate the quality evaluation index after mapping; when the index does not meet the preset threshold, a rollback mechanism is triggered, which includes switching to a simpler geometric transformation model or adjusting the model parameters.

[0068] Specifically, the system calculates the following quality indicators (reprojection error):

[0069] in : The reprojection error of the i-th control point (Euclidean distance, in "coordinate units": normalized units in normalized space; pixels in canvas pixels). The average error (overall fit level). A 95th percentile error (robustly reflects whether the "worst batch" of points is out of line, and is more useful for finding outliers / noise points).

[0070] The design logic of the above mechanism is to ensure the high reliability of the system in complex clinical environments. Quality assessment metrics (such as keypoint reprojection error, contour overlap, and deformation energy) are used to quantify the quality of the mapping results. When the assessment finds that the current model may lead to abnormal fitting (such as excessive error or unreasonable deformation) due to poor input data quality (such as image blurring or segmentation errors), the system automatically triggers a rollback. Rollback strategies include switching to a simpler geometric transformation model (e.g., rolling back from a complex thin-plate spline to a stable affine transformation) or adjusting model parameters (e.g., increasing the regularization parameter λ to force smoothing). This constitutes the system's "safety net," ensuring that it can still output a result that, although accuracy may be compromised, is absolutely stable, reliable, and usable even in marginal situations, avoiding complete failure and serving as a key guarantee for the method's clinical applicability.

[0071] In summary, this application obtains the ultrasound image to be mapped and the corresponding anatomical structure segmentation results; determines a first set of key points in the ultrasound image coordinate system to describe the anatomical morphology; obtains a preset pattern diagram and a second set of key points in the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in anatomical semantics; establishes a geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system based on the first set of key points and the second set of key points; and uses the geometric transformation model to map the anatomical structure contour, lesion contour, and pixel grid to the pattern diagram coordinate system, generating a standardized structural representation. This solves the problems in the prior art where the pattern diagram is difficult to maintain consistency with the patient's actual uterine morphology and the lesion location is inaccurate, achieving accurate and automatic mapping of anatomical structures and lesion contours, and possessing quality assessment and robust processing capabilities to generate standardized, quantifiable, and comparable ultrasound structural reports.

[0072] At least some steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0073] Based on the same inventive concept, this application also provides a system for implementing the aforementioned key-point constraint-based spatial mapping system for ultrasound images and pattern diagrams. The solution provided by this system is similar to the implementation described in the above method. Therefore, the specific limitations of one or more key-point constraint-based spatial mapping system embodiments provided below can be found in the limitations of the key-point constraint-based spatial mapping method for ultrasound images and pattern diagrams described above, and will not be repeated here.

[0074] In one embodiment, such as Figure 2 As shown, a spatial mapping system for ultrasound images and pattern diagrams based on key point constraints is provided, including: The acquisition module 100 is used to acquire the ultrasound image to be mapped and the anatomical structure segmentation results corresponding to the ultrasound image.

[0075] The first determining module 200 is used to determine a first set of key points in the ultrasound image coordinate system to describe the anatomical morphology.

[0076] The second determining module 300 is used to obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics.

[0077] The model building module 400 is used to build a geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system based on the first set of key points and the second set of key points.

[0078] The mapping module 500 is used to map the anatomical structure outline, lesion outline, and pixel grid to the pattern diagram coordinate system using the geometric transformation model, thereby generating a standardized structural representation.

[0079] In one embodiment, the first determining module 200 is further configured to: normalize the first set of key points before establishing the geometric transformation model, and transform their coordinates to a preset normalized coordinate system.

[0080] In one implementation, the system further includes a quality assessment module for: Calculate the quality assessment indicators after mapping; When the indicator does not meet the preset threshold, a rollback mechanism is triggered, which includes switching to a simpler geometric transformation model or adjusting the model parameters.

[0081] In one implementation, the mapping module 500 is further configured to: The geometric transformation model is used to reverse map the pixel grid of the ultrasound image, and the deformed image with the same viewpoint as the pattern image is generated by interpolation resampling; the interpolation method includes bilinear interpolation or bicubic interpolation.

[0082] The modules in the aforementioned key-point-constrained ultrasound image and pattern spatial mapping system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0083] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, and network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores preset data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a method for spatial mapping of ultrasound images and pattern diagrams based on key point constraints.

[0084] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0085] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method for spatial mapping of ultrasound images and pattern diagrams based on key point constraints.

[0086] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described method for spatial mapping of ultrasound images and pattern diagrams based on key point constraints.

[0087] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the above-described method for mapping ultrasound images and pattern diagrams based on key point constraints.

[0088] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0089] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0090] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for spatial mapping of ultrasound images and pattern diagrams based on key point constraints, characterized in that, The method includes: Obtain the ultrasound image to be mapped and the segmentation results of the corresponding anatomical structures in the ultrasound image; In the ultrasound image coordinate system, identify the first set of key points used to describe the anatomical morphology; Obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics; A geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system is established based on the first set of key points and the second set of key points. Using the geometric transformation model, the anatomical structure outline, lesion outline, and pixel grid are mapped to the pattern diagram coordinate system to generate a standardized structural representation.

2. The method according to claim 1, characterized in that, The first set of key points includes at least: the cephalic endpoint, the pedunculate endpoint, the midpoint of the anterior wall, and the midpoint of the posterior wall of the uterine contour; and / or, the cephalic endpoint, the pedunculate endpoint, the midpoint of the anterior wall, and the midpoint of the posterior wall of the endometrial contour; The midpoints of the anterior and posterior walls are determined based on the division of the intima's central axis.

3. The method according to claim 1, characterized in that, Before establishing the geometric transformation model, the first set of key points are normalized, and their coordinates are transformed to a preset normalized coordinate system.

4. The method according to claim 1, characterized in that, The geometric transformation model is a linear transformation model, which is obtained by minimizing the error between all key point pairs; the linear transformation model includes an affine transformation model or a perspective transformation model.

5. The method according to claim 1, characterized in that, The geometric transformation model is a thin plate spline transformation model, which minimizes the overall bending energy while accurately matching key points.

6. The method according to claim 5, characterized in that, When solving the linear system of the thin plate spline transformation model, a regularization parameter is introduced to suppress overfitting and balance the key point matching accuracy and deformation smoothness.

7. The method according to claim 1, characterized in that, The method further includes: Calculate the quality assessment indicators after mapping; When the indicator does not meet the preset threshold, a rollback mechanism is triggered, which includes switching to a simpler geometric transformation model or adjusting the model parameters.

8. The method according to claim 1 or 7, characterized in that, The method further includes: The geometric transformation model is used to reverse map the pixel grid of the ultrasound image, and the deformed image with the same viewpoint as the pattern image is generated by interpolation resampling; the interpolation method includes bilinear interpolation or bicubic interpolation.

9. The method according to claim 1, characterized in that, The geometric transformation model can be automatically selected based on the number of key point pairs, fitting error, or deformation energy index. Optional models include affine transformation, perspective transformation, and thin plate spline transformation.

10. A spatial mapping system for ultrasound images and pattern diagrams based on key point constraints, characterized in that, The system includes: The acquisition module is used to acquire the ultrasound image to be mapped and the anatomical structure segmentation results corresponding to the ultrasound image; The first determining module is used to determine a first set of key points in the ultrasound image coordinate system to describe the anatomical morphology; The second determining module is used to obtain a preset pattern diagram and a second set of key points of the preset pattern diagram, wherein the second set of key points corresponds one-to-one with the first set of key points in terms of anatomical semantics. The model building module is used to build a geometric transformation model from the ultrasound image coordinate system to the pattern diagram coordinate system based on the first set of key points and the second set of key points. The mapping module is used to map the anatomical structure outline, lesion outline, and pixel grid to the pattern diagram coordinate system using the geometric transformation model, thereby generating a standardized structural representation.

11. A computer device comprising a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

Images