Medical image data processing method, system and application

Abstract

The present disclosure provides a method, system, and application for processing medical image data. The processing method includes: obtaining a plurality of target sector images of a preset subject at a current standard position under a preset surgical procedure; the target sector images include intracardiac echocardiography (ICE); and inputting each target sector image frame into a pre-trained sector parameter prediction model to output corresponding target sector parameter information. In the present disclosure, by constructing a reliable sector parameter prediction model, the accuracy and efficiency of analyzing and processing each input target sector image frame are improved, providing the surgeon with reliable and comprehensive auxiliary information to assist the surgeon in analyzing cardiac structure and guiding surgical procedures. This method, combined with AI algorithms, helps the surgeon conduct preoperative rehearsal planning and simulation operations without the need for additional hardware equipment, thereby improving operational proficiency, surgical success rate, and reducing surgical costs.

Description

Technical Field

[0001] The present disclosure relates to the field of medical equipment technology, and in particular to a method, system and application for processing medical image data. Background Art

[0002] ICE (Intracardiac Echocardiography) imaging equipment is an effective real-time imaging tool in cardiac catheterization and electrophysiology laboratories. Specifically, it uses an ultrasound transducer at the tip of the catheter to provide clear images of cardiac structures during intracardiac interventional procedures, facilitating procedures such as transseptal puncture and atrial fibrillation ablation. Due to the unique environment in which intracardiac ultrasound catheters are placed, ICE imaging typically only reveals localized cardiac anatomy. Therefore, interpreting ICE images is more challenging than external imaging. This also means that intraoperative catheter manipulation requires a high level of proficiency, often requiring extensive training and practice.

[0003] Existing intracardiac ultrasound catheter devices require a specialized three-dimensional mapping system for navigation and three-dimensional reconstruction, which usually requires additional hardware such as magnetic positioning, IMU (inertial measurement unit), electrode catheters, etc., and there is a problem of high surgical costs; in addition, the hardware calibration between these devices is complex and the operation process is cumbersome, further increasing the difficulty and time of the operation. Summary of the Invention

[0004] The technical problem to be solved by the present disclosure is to overcome the defects existing in the prior art and provide a method, system and application for processing medical image data.

[0005] The present disclosure solves the above technical problems through the following technical solutions:

[0006] The present disclosure provides a method for processing medical image data, the processing method comprising:

[0007] Acquire a plurality of target sector images of a preset object at a current standard point under a preset surgical procedure; wherein the target sector images include intracardiac echocardiogram (ICE);

[0008] Each frame of the target sector image is input into a pre-trained sector parameter prediction model to output corresponding target sector parameter information.

[0009] Optionally, the step of constructing the sector parameter prediction model includes:

[0010] Acquire several groups of historical sample data corresponding to the preset surgical procedure;

[0011] Each set of historical sample data includes a plurality of sample sector images at different spatial positions at different preset points, and sample sector parameter information corresponding to each of the sample sector images is obtained;

[0012] Using different sample sector images and corresponding sample sector parameter information as sample training data, training to obtain the sector parameter prediction model;

[0013] and / or,

[0014] The processing method further comprises:

[0015] Constructing a standard three-dimensional structural model of the preset object;

[0016] After the step of outputting the corresponding target sector parameter information, the method further includes:

[0017] Controlling the display of the spatial position of the current standard point in the standard three-dimensional structure model; and / or the target sector parameter information at the position.

[0018] Optionally, the sample sector parameter information includes a sample sector type, the sector parameter prediction model includes a sector classification model, and the target sector parameter information includes a target sector type;

[0019] and / or,

[0020] The sample sector parameter information includes sample structure anatomical information of different segmented regions in the sector, the sector parameter prediction model includes a segmentation model, and the target sector parameter information includes target contour information and corresponding target structure anatomical information of different segmented regions in the sector.

[0021] Optionally, after the step of outputting the corresponding target sector parameter information, the method further includes:

[0022] Determining a target local area in the preset object that matches the preset technique, and obtaining local contour information corresponding to the segmented area that matches the target local area;

[0023] Obtaining point space position information of each contour point on the local contour corresponding to the local contour information;

[0024] Obtaining sector spatial position information of the target sector corresponding to each frame of the target sector image;

[0025] Based on the point space position information and the sector space position information, a local three-dimensional structure model corresponding to the target local area is reconstructed.

[0026] Optionally, the step of obtaining sector spatial position information of the target sector corresponding to each frame of the target sector image includes:

[0027] Comparing each frame of the target sector image with a plurality of the sample sector images at the current standard point to obtain corresponding similarities;

[0028] Using the position of the sample sector image with the highest similarity as the sector space position information of the target sector image;

[0029] Alternatively, the step of obtaining the sector spatial position information of the target sector corresponding to each frame of the target sector image includes:

[0030] The target sector image of any frame is input into a pre-trained position estimation model to output the sector spatial position information of the target sector.

[0031] Optionally, the processing method further includes:

[0032] Acquire a plurality of standard points in the standard three-dimensional structure model that match the preset surgical procedure, a plurality of standard sector images at different spatial positions at each of the standard points, and standard sector spatial position information corresponding to each of the standard sector images;

[0033] The steps of training the position estimation model include:

[0034] Performing model training on each of the sample sector images and the corresponding standard sector spatial position information to obtain the position estimation model;

[0035] and / or,

[0036] After the step of reconstructing the local three-dimensional structure model corresponding to the target local area, the method further includes:

[0037] Controlling the display of the local three-dimensional structure model;

[0038] and / or,

[0039] The method for processing medical image data further includes:

[0040] generating preoperative planning reference information based on at least one of the target sector image, the target sector parameter information, and the local three-dimensional structure model;

[0041] and / or,

[0042] The preset object includes a heart;

[0043] and / or,

[0044] The target sector image is an image acquired in a preoperative simulation environment.

[0045] The present disclosure further provides a medical image data processing system, the processing system comprising:

[0046] A target image acquisition module is used to acquire a plurality of target sector images of a preset object at a current standard point under a preset surgical procedure; wherein the target sector images include intracardiac echocardiography (ICE);

[0047] The target parameter information output module is used to input each frame of the target sector image into a pre-trained sector parameter prediction model to output corresponding target sector parameter information.

[0048] Optionally, the processing system further includes:

[0049] A sample data acquisition module, used to acquire several groups of historical sample data corresponding to the preset procedure;

[0050] Each set of historical sample data includes a plurality of sample sector images at different spatial positions at different preset points, and sample sector parameter information corresponding to each of the sample sector images is obtained;

[0051] A model training module is used to use different sample sector images and corresponding sample sector parameter information as sample training data to train and obtain the sector parameter prediction model;

[0052] and / or,

[0053] The processing system further comprises:

[0054] A standard model building module, used to build a standard three-dimensional structural model of the preset object;

[0055] The first display control module is used to control the display of the spatial position of the current standard point in the standard three-dimensional structure model; and / or the target sector parameter information at the position.

[0056] Optionally, the sample sector parameter information includes a sample sector type, the sector parameter prediction model includes a sector classification model, and the target sector parameter information includes a target sector type;

[0057] and / or,

[0058] The sample sector parameter information includes sample structure anatomical information of different segmented regions in the sector, the sector parameter prediction model includes a segmentation model, and the target sector parameter information includes target contour information and corresponding target structure anatomical information of different segmented regions in the sector.

[0059] Optionally, the processing system further includes:

[0060] A local area determination module, configured to determine a target local area in the preset object that matches the preset surgical procedure;

[0061] A local contour acquisition module is used to obtain local contour information corresponding to the segmented area matched with the target local area;

[0062] a point position acquisition module, configured to acquire point spatial position information of each contour point on the local contour corresponding to the local contour information;

[0063] A sector position acquisition module, configured to acquire sector spatial position information of a target sector corresponding to each frame of the target sector image;

[0064] The local model reconstruction module is used to reconstruct a local three-dimensional structure model corresponding to the target local area based on the point space position information and the sector space position information.

[0065] Optionally, the sector position acquisition module is used to compare each frame of the target sector image with a plurality of the sample sector images at the current standard point to obtain corresponding similarities; and use the position of the sample sector image with the highest similarity as the sector spatial position information of the target sector image;

[0066] Alternatively, the sector position acquisition module is used to input the target sector image of any frame into a pre-trained position estimation model to output the sector spatial position information of the target sector.

[0067] Optionally, the processing system further includes:

[0068] an information acquisition module, configured to acquire a plurality of standard points in the standard three-dimensional structure model that match the preset surgical procedure, a plurality of standard sector images at different spatial positions at each of the standard points, and standard sector spatial position information corresponding to each of the standard sector images;

[0069] An estimation model training module, configured to perform model training on each of the sample sector images and the corresponding standard sector spatial position information to obtain the position estimation model;

[0070] and / or,

[0071] The processing system further comprises:

[0072] A second display control module, configured to control the display of the local three-dimensional structure model;

[0073] and / or,

[0074] The processing system further comprises:

[0075] a reference information generating module, configured to generate preoperative planning reference information based on at least one of the target sector image, the target sector parameter information, and the local three-dimensional structure model;

[0076] and / or,

[0077] The preset object includes a heart;

[0078] and / or,

[0079] The target sector image is an image acquired in a preoperative simulation environment.

[0080] The present disclosure also provides an intracardiac ultrasound catheter device, which includes the medical image data processing system as described above.

[0081] The present disclosure also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and configured to run on the processor, wherein the processor implements the above-mentioned method for processing medical image data when executing the computer program.

[0082] The present disclosure also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the above-mentioned method for processing medical image data is implemented.

[0083] The present disclosure also provides a computer program product, including a computer program, which implements the above-mentioned method for processing medical image data when executed by a processor.

[0084] On the basis of conforming to the common sense in this field, the above-mentioned preferred conditions can be arbitrarily combined to obtain the preferred embodiments of the present disclosure.

[0085] The positive progress of this disclosure is:

[0086] In the present disclosure, by constructing a reliable sector parameter prediction model, the accuracy and efficiency of the analysis and processing of each frame of intracardiac echocardiogram (ICE) collected in real time by the ultrasonic transducer are effectively improved, thereby providing reliable auxiliary information to the surgeon and other relevant personnel, assisting the surgeon in analyzing the heart structure, guiding the surgical process, etc., and achieving the purpose of helping the surgeon to conduct preoperative rehearsal planning and preoperative simulation operations in combination with AI (artificial intelligence) algorithms without the need for additional hardware equipment, thereby improving the surgeon's operational proficiency, the success rate of the operation, and greatly reducing the cost of the operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 This is a flowchart of a method for processing medical image data according to Embodiment 1 of the present disclosure;

[0088] Figure 2 Schematic diagram of different fan images according to Example 1 of the present disclosure;

[0089] Figure 3This is a first flow chart of a method for processing medical image data according to Embodiment 2 of the present disclosure;

[0090] Figure 4 This is a second flow chart of the method for processing medical image data according to Embodiment 2 of the present disclosure;

[0091] Figure 5 This is a schematic diagram of the output results of the sector classification model of Example 2 of the present disclosure;

[0092] Figure 6 This is a schematic diagram of the output results of the segmentation model of Example 2 of the present disclosure;

[0093] Figure 7 This is a third flow chart of the method for processing medical image data according to Embodiment 2 of the present disclosure;

[0094] Figure 8 Schematic diagram of a local three-dimensional structure of the left atrial region LA according to Example 2 of the present disclosure;

[0095] Figure 9 This is a schematic diagram of marking important locations in the heart according to Example 2 of the present disclosure;

[0096] Figure 10 This is a module diagram of a medical image data processing system according to Embodiment 3 of the present disclosure;

[0097] Figure 11 This is a module diagram of a medical image data processing system according to Embodiment 4 of the present disclosure;

[0098] Figure 12 This is a structural diagram of an electronic device according to embodiment 6 of the present disclosure. DETAILED DESCRIPTION

[0099] The present disclosure is further illustrated below by way of examples, but the present disclosure is not limited to the scope of the examples.

[0100] In the embodiments of the present disclosure, prefixes such as "first" and "second" are used only to distinguish different description objects and have no limiting effect on the position, order, priority, quantity or content of the described objects. In the embodiments of the present disclosure, the use of prefixes such as ordinal numbers to distinguish description objects does not constitute a restriction on the described objects. For the statement of the described objects, please refer to the description in the context of the embodiments, and the use of such prefixes should not constitute an unnecessary restriction. In addition, in the description of this embodiment, unless otherwise specified, the meaning of "plurality" is two or more.

[0101] Example 1

[0102] like Figure 1 As shown, the method for processing medical image data of this embodiment includes:

[0103] S101, obtaining a plurality of target sector images of a preset object at a current standard position under a preset technique;

[0104] Specifically, generally, the preset object includes the heart or a partial structure of the heart, etc.; the target sector image includes an intracardiac echocardiogram (ICE).

[0105] When the preset object is the heart, the medical image data processing solution of this embodiment is applied in the preoperative planning stage of intracardiac interventional surgery; different preset surgical procedures include but are not limited to atrial septal puncture and atrial fibrillation ablation.

[0106] like Figure 2 As shown, the data obtained when the catheter moves is a video, and the sector image obtained in each frame of the video is a section obtained by the ultrasound probe at the catheter head end at a position.

[0107] Among them, during the actual interventional operation, the ultrasonic transducer is placed at the corresponding standard point under the current procedure according to the operating process corresponding to the current procedure; different preset procedures can match different standard points; of course, according to the actual scene requirements, the corresponding standard points can be set to be the same for different preset procedures to simplify the overall data processing complexity and processing efficiency of medical image data.

[0108] In addition, these target sector images may be images acquired in a preoperative simulation environment, so as to assist the surgeon in processing relevant information in the preoperative simulation environment, thereby achieving a good preoperative rehearsal effect.

[0109] Taking transseptal puncture as an example, the different target sector images obtained by scanning are as follows: Figure 2 As shown in the figure, these sectors include the Home View, Ao / RVOT / PA View, LAA View (sector images obtained when the catheter is rotated to different positions), etc.

[0110] S102 : Input each frame of target sector image into a pre-trained sector parameter prediction model to output corresponding target sector parameter information.

[0111] Specifically, the target sector parameter information includes but is not limited to the sector type, structural anatomical information of tissues in different areas of the sector, etc. This information can be displayed to the operator in real time according to actual needs.

[0112] In the present disclosure, by constructing a reliable fan parameter prediction model, the accuracy and efficiency of the analysis and processing of each frame of intracardiac echocardiogram (ICE) collected in real time by the ultrasonic transducer are effectively improved, thereby providing reliable auxiliary information to the surgeon and other relevant personnel, assisting the surgeon in analyzing the heart structure, guiding the surgical process, etc., and achieving the purpose of helping the surgeon to conduct preoperative rehearsal planning and preoperative simulation operations in combination with AI algorithms without the need for additional hardware equipment, thereby improving the surgeon's operational proficiency, the success rate of the operation, and greatly reducing the cost of the operation.

[0113] Example 2

[0114] The method for processing image data in this embodiment is a further improvement of the first embodiment. Specifically:

[0115] In one feasible solution, Figure 3 As shown in FIG, the steps of constructing the sector parameter prediction model include:

[0116] S201, obtaining several groups of historical sample data corresponding to a preset surgical procedure;

[0117] Each set of historical sample data includes a number of sample sector images at different spatial positions at different preset points, and sample sector parameter information corresponding to each sample sector image is obtained;

[0118] S202, using different sample sector images and corresponding sample sector parameter information as sample training data to train and obtain a sector parameter prediction model;

[0119] There is no restriction on the type of model used, such as a large model, as long as it can meet the corresponding sector image processing requirements.

[0120] In the present disclosure, model training is performed based on several groups of sample sectors and their sector parameter information to obtain a sector parameter prediction model that can accurately output the corresponding sector association information in any sector, thereby ensuring the accuracy and reliability of the output results of the sector parameter prediction model.

[0121] In one feasible solution, Figure 4 As shown, the processing method also includes:

[0122] Construct a standard three-dimensional structural model of a preset object;

[0123] When the preset object is a heart, the standard three-dimensional structure model is a three-dimensional structured model of the heart.

[0124] After step S102, the following steps are also included:

[0125] S1031. Controlling the display of the spatial position of the current standard point in the standard three-dimensional structure model; and / or target sector parameter information at the current position.

[0126] In this solution, after the sector parameter prediction model outputs the sector parameter information corresponding to any intracardiac echocardiogram (ICE), the information at that point is controlled to be intuitively displayed in the standard three-dimensional structure model so that the surgeon can view it in time, allowing the surgeon to view and analyze the data more intuitively and clearly, allowing the surgeon to become more familiar with the heart structure, surgical procedures, etc., to ensure the rationality and safety of subsequent surgical operations, etc., and also enhance the sense of interaction and operating experience.

[0127] In an implementable solution, the sample sector parameter information includes the sample sector type, the sector parameter prediction model includes a sector classification model (or classification network), and the target sector parameter information includes the target sector type;

[0128] For example, the home view position and Ao view position of the standard point are adjacent standard sectors. The sectors between them have certain structural similarities and can be distinguished by pre-labeling. Among them, based on the type label information of the sector, the basic information of the sector can be clearly understood.

[0129] like Figure 5 As shown, the classification result of any input image output by the sector classification model is any one of the reference position A, left atrial appendage B, oval fossa C, etc., and the current position of the probe can also be known.

[0130] and / or,

[0131] The sample sector parameter information includes the sample structure anatomical information of different segmented areas in the sector, the sector parameter prediction model includes the segmentation model (or segmentation network), and the target sector parameter information includes the target contour information of different segmented areas in the sector and the corresponding target structure anatomical information.

[0132] Among them, the segmentation model is used to extract the boundaries of the segmented areas where different tissues in the fan are located, as well as the specific information of the tissues in each segmented area, so as to help the surgeon identify the regional information of the three-dimensional structure corresponding to the two-dimensional fan. Figure 6 As shown, the marked RA, LA, LSPV, etc. are all different segmented areas.

[0133] In the present disclosure, for any two-dimensional fan image input into the fan parameter prediction model, the corresponding fan type can be output, and the structural anatomical information of the tissues in different areas of the fan can also be output. Of course, parameter prediction models of other dimensions can also be trained according to actual needs, thereby ensuring the efficiency, accuracy and comprehensiveness of the analysis and processing of any fan image, so as to provide the surgeon with a clear understanding of the fan information, effectively assist the surgeon in performing more reasonable and accurate surgical operations in the future, and thereby ensure the rationality and safety of the surgeon's subsequent surgical operations.

[0134] In one feasible solution, Figure 7 As shown, after step S102, the following steps are further included:

[0135] S1032, determining a target local area in a preset object that matches a preset technique, and obtaining local contour information corresponding to the segmented area that matches the target local area;

[0136] The target local area is the surgical target area; for example, when the preset object corresponds to the heart, the operating catheter is moved to the surgical target area; wherein the target local area may be the left atrial surgical area. At this time, it is necessary to extract the local area corresponding to the left atrial surgical area from the multiple different segmented areas output by the model, and focus on reconstructing the three-dimensional local structure corresponding to the left atrial surgical area. Areas other than the left atrial surgical area are not considered.

[0137] S104, obtaining the point space position information of each contour point on the local contour corresponding to the local contour information;

[0138] S105 , obtaining sector spatial position information of the target sector corresponding to each frame of the target sector image; wherein the sector spatial position information is generally the angle of the sector in space.

[0139] S106 : Reconstructing a local three-dimensional structure model corresponding to the target local area based on the point space position information and the sector space position information.

[0140] Among them, steps S101-S102 are for any point, that is, the sector parameter information corresponding to the sector image collected by the catheter at any point can be obtained; in steps S103-S106, the target local area corresponds to a specific point or target point, and the point corresponds to the target local area; at the point in the target local area, the sector parameter information of each frame of the sector image obtained by finely rotating the catheter is also implemented by executing steps S101-S102.

[0141] Specifically, this solution can achieve accurate positioning of contour points on the local contour of the surgical target area and accurate positioning of the fan surface without the need for additional positioning sensors, thereby providing location information for navigation. While greatly reducing the cost of surgery, it also greatly simplifies the implementation process.

[0142] In addition, for the current preset surgical procedure, the accurate positioning of the contour points on the local contour of the surgical target area and the accurate positioning of the fan surface can achieve the effect of fine positioning reconstruction focusing on the local area of the surgical target area.

[0143] The surgical target area corresponds to LA, and the local contour corresponding to the left atrial area LA is obtained. The spatial position data corresponding to each contour point on the contour is filtered out from the three-dimensional point cloud data, and the spatial angle information of different sectors at the points corresponding to the surgical target area is obtained. The outer surface of the left atrial area LA is reconstructed to construct the local three-dimensional structural model corresponding to the left atrial area LA, which is Figure 8 The part shown in green.

[0144] In this scheme, by obtaining the point spatial position information of each contour point on the local contour in each target sector, as well as the sector spatial position information of each sector, a local three-dimensional structural model of the target area to be operated on is constructed, such as the three-dimensional local structure corresponding to the left atrial surgical area, to assist the surgeon in analyzing and processing the finely reconstructed local three-dimensional structural model, determine preoperative planning, intraoperative operations, etc., and further facilitate the surgeon to conduct preoperative rehearsal planning, improve the surgeon's operational proficiency, the success rate of the operation, and reduce the cost of the operation.

[0145] In one feasible solution, step S105 includes:

[0146] Compare each frame of the target sector image with several sample sector images at the current standard point to obtain the corresponding similarity;

[0147] The position of the sample sector image with the highest similarity is used as the sector space position information of the target sector image.

[0148] In this scheme, the target sector image of any frame at the current standard point is compared with different sample sector images at the current standard point in the standard three-dimensional structure model to obtain different similarities, and the sample sector image with the highest similarity is screened out to quickly determine the spatial position of each target sector, and can reduce the difficulty of estimating the positioning of incomplete sectors in a larger space, thereby assisting in reconstructing the local three-dimensional model of the surgical target area, ensuring the efficiency and reliability of local three-dimensional model reconstruction.

[0149] In one feasible solution, step S105 includes:

[0150] The target sector image of any frame is input into a pre-trained position estimation model to output the sector spatial position information of the target sector.

[0151] In this solution, the spatial position of the target sector image of any frame is predicted through a pre-built position estimation model, which also ensures that the spatial position of each target sector is accurately and quickly determined, and can reduce the difficulty of estimating the positioning of incomplete sectors in a larger space, thereby assisting in reconstructing the local three-dimensional model of the surgical target area, ensuring the efficiency and reliability of local three-dimensional model reconstruction.

[0152] In one feasible solution, the processing method further includes:

[0153] Acquire several standard points in the standard three-dimensional structure model that match the preset surgical procedure, as well as several standard sector images at different spatial positions at each standard point, and standard sector spatial position information corresponding to each standard sector image;

[0154] Among them, based on the standard three-dimensional structure model, several two-dimensional slices corresponding to several standard points (i.e., standard sector images corresponding to standard sectors) and their corresponding associated data (sector type, structural anatomical information corresponding to the contour, spatial position information, etc.) are sampled.

[0155] Taking the preset object as an example, since the anatomical structure of the human heart is generally consistent, some key positions and cross-sections can be determined in the standard heart, and these positional relationships can be represented by a tree structure. In this way, in the structured standard heart, the relationship between each point position and the parent node is determined; the relative relationship between these positions can be roughly obtained through the transfer matrix (for example, translation and rotation of the fan), and the structure can be set through a standardized model, and can also be seamlessly connected to sensor data by replacing the point position and transfer matrix with sensor data.

[0156] Specifically, after constructing a standard three-dimensional structural model of the heart, several standard points in the standard three-dimensional structural model are divided based on experience. Each standard point has a different sector as the catheter rotates.

[0157] Some ICE standard sectors can be preset. Of course, for a surgical procedure, the first choice is to select the required position, such as Figure 9 As shown, such as the reference position A, left atrial appendage B, oval fossa C, etc., training is performed by sampling the standard sectors of these places and their associated data as training data.

[0158] For example, the coordinates of the Home View position observed in the right atrium are designated as v1, and the coordinates of other positions are designated as v2, v3, etc. These fixed points serve as standard points, and v1-v2-v3 can be viewed as a tree structure with v1 as the root node. Each position has a different fan as the catheter rotates. For example, in v1, there is Home View -> Ao / RVOT / PA View -> LAAView -> ..., and the corresponding fan rotation angle can be obtained. These specific fan rotations serve as standard fans, with relatively fixed positions, and can be used as angle references. In other words, the structural information inside the heart can be represented by different standard points and corresponding angles.

[0159] The steps to train a location estimation model include:

[0160] Each sample sector image and the corresponding standard sector spatial position information are trained to obtain a position estimation model;

[0161] The standard sector spatial position information is the spatial angle information of the standard sector, that is, the position estimation model is used to output the spatial angle information of any input sector image.

[0162] In this solution, model training is performed based on standard sectors with different standard points predetermined in the standard three-dimensional structure model and their corresponding spatial angle information to quickly and accurately identify the position of any sector in space.

[0163] In one feasible solution, after the step of reconstructing the local three-dimensional structure model corresponding to the target local area, the method further includes:

[0164] Control the display of local three-dimensional structural model;

[0165] In this solution, important data during the processing of medical image data is displayed to the surgeon in a timely manner through visualization, allowing the surgeon to view and analyze the data more intuitively and clearly, and enabling the surgeon to become more familiar with the heart structure, surgical procedures, etc., to ensure the rationality and safety of subsequent surgical operations.

[0166] In one feasible solution, the method for processing medical image data further includes:

[0167] Preoperative planning reference information is generated based on at least one of the target sector image, target sector parameter information, and a local three-dimensional structure model.

[0168] Among them, without the need for additional positioning sensors, sector positioning and positioning of contour points on the local contour of the surgical target area are achieved, and then combined with the target sector image, target sector parameter information, local three-dimensional structure model data, and the generated preoperative planning reference information, that is, with the assistance of comprehensive information, timely and accurate provision is made to provide guidance for surgical process planning, so that preoperative planning of automatic navigation of intracardiac ultrasound catheters can be reliably achieved.

[0169] In this solution, through the automatic generation of preoperative planning reference information, the surgeon is provided with relevant reference information, such as key precautions reference information, risk warning reference information, catheter navigation path, etc., which is convenient for the surgeon to refer to and analyze, conduct preoperative planning, and practice planning in a more comprehensive and accurate manner, thereby ensuring the safety and success rate of surgical operations during subsequent operations.

[0170] In this embodiment, the process of providing auxiliary support for key preoperative information based on the preoperative simulation environment to achieve preoperative surgical planning is as follows:

[0171] S1. According to the surgical procedure, the catheter first reaches the first standard point and the formal procedure begins;

[0172] S2. Obtaining real-time sector images of the transducer at different rotational positions of the current ultrasound catheter head, and inputting these images into a pre-trained sector parameter prediction model to output the corresponding sector type, the contours of different segmented regions on the sector, and the structural anatomical information within each contour;

[0173] S3, operate the catheter according to the predefined surgical procedure, enter the next standard point, and repeat step S2 at the current standard point;

[0174] After reaching the standard point for the scheduled surgical operation, that is, the point corresponding to the surgical target area, the transducer of the rotating ultrasound catheter head is finely scanned at the standard point to obtain the sector type corresponding to the sector images continuously acquired at the standard point, the contours of different segmented areas on the sector, and the structural anatomical information within each contour;

[0175] S4, screening out a contour corresponding to the surgical target area from the multiple segmented areas, and obtaining point space position information of each contour point on the contour; and obtaining sector space position information of each sector at the standard point position;

[0176] Based on the sector space position information of each sector at the standard point and the point space position information of each contour point on the contour corresponding to the surgical target area in each sector at the standard point, these data are used as volume data and surface processed to obtain a local three-dimensional structural model corresponding to the surgical target area to assist subsequent surgical operations, etc.

[0177] The following describes the implementation principle of the medical image data processing solution of this embodiment with reference to specific examples:

[0178] (1) Constructing a standard three-dimensional structural model

[0179] After constructing a standard three-dimensional structural model of the heart, several standard points in the standard three-dimensional structural model are divided based on experience;

[0180] For example, the coordinates of the Home View position observed in the right atrium are defined as v1, and the other positions are v2, v3, ..., these fixed points are used as standard points, and v1-v2-v3 can be regarded as a tree structure with v1 as the root node.

[0181] Each position has a different fan view as the catheter rotates. For example, in v1, there is Home View -> Ao / RVOT / PA View -> LAA View -> ..., and the corresponding fan view rotation angle can be obtained. These specific fan views serve as standard fan views with relatively fixed positions and can be used as angle references. In other words, the structural information inside the heart can be represented by different standard points and corresponding angles.

[0182] (2) Constructing a fan parameter prediction model

[0183] Based on the standard 3D structural model, several sections are obtained, namely, several sample sector images and their corresponding sample sector parameter information (such as sector type, structural anatomical information, etc.), which are used as sample data for model training to obtain sector classification models and segmentation models;

[0184] Performing surgical operations at preset target points requires precise positioning modeling. At this time, the handle is used to rotate the catheter to scan and obtain multiple sector images. The sector classification model and segmentation model are used to output the corresponding sector classification, target contour information of different segmented areas in the sector, and corresponding target structure anatomical information for any sector image input.

[0185] (3) Reconstruction of local three-dimensional structural model

[0186] Each sample sector image and the corresponding standard sector spatial position information are trained to obtain a position estimation model;

[0187] The multiple sector images obtained by scanning are input into the trained position estimation model to output the spatial angle information of different sectors.

[0188] (4) Information display and navigation guidance

[0189] In the preoperative simulation planning process, the three-dimensional heart structure and corresponding standard points are displayed in real time;

[0190] At the same time, it can also display each frame of the two-dimensional sector image and the results of different model analysis outputs (such as sector type, the contours of different tissue areas on the sector, i.e. tissue information, the spatial position of each point on the contour, the spatial angle of the sector, etc.). That is, through visual intelligent guidance, the surgeon can become more familiar with the heart structure.

[0191] In addition, the surgical process guides the surgeon to perform further surgical operations through a predefined standard point sequence, combined with the results of different model analysis outputs (such as fan type, the contours of different tissue areas on the fan, i.e. tissue information, the spatial position of each point on the contour, the spatial angle of the fan, etc.), so that the surgeon becomes more familiar with the surgical process. That is, through visual intelligent guidance, it helps the surgeon to conduct preoperative rehearsal planning, improve the surgeon's operational proficiency and the success rate of the operation, etc.

[0192] Example 3

[0193] like Figure 10 As shown, the medical image data processing system of this embodiment includes:

[0194] The target image acquisition module 1 is used to acquire a plurality of target sector images of a preset object at a current standard point under a preset surgical procedure; wherein the target sector images include intracardiac echocardiogram (ICE);

[0195] The target parameter information output module 2 is used to input each frame of the target sector image into a pre-trained sector parameter prediction model to output corresponding target sector parameter information.

[0196] In the present disclosure, by constructing a reliable fan parameter prediction model, the accuracy and efficiency of the analysis and processing of each frame of intracardiac echocardiogram (ICE) collected in real time by the ultrasonic transducer are effectively improved, thereby providing reliable auxiliary information to the surgeon and other relevant personnel, assisting the surgeon in analyzing the heart structure, guiding the surgical process, etc., and achieving the purpose of helping the surgeon to conduct preoperative rehearsal planning and preoperative simulation operations in combination with AI algorithms without the need for additional hardware equipment, thereby improving the surgeon's operational proficiency, the success rate of the operation, and greatly reducing the cost of the operation.

[0197] Since the system embodiments generally correspond to the method embodiments, reference will be made to the description of the method embodiments for relevant details. The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components of the units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the disclosed solution.

[0198] Example 4

[0199] like Figure 11 As shown, the image data processing system of this embodiment is a further improvement of the embodiment 3, specifically:

[0200] In one feasible solution, the processing system further includes:

[0201] The sample data acquisition module 3 is used to acquire several groups of historical sample data corresponding to the preset surgical procedures;

[0202] Each set of historical sample data includes a number of sample sector images at different spatial positions at different preset points, and sample sector parameter information corresponding to each sample sector image is obtained;

[0203] The model training module 4 is used to use different sample sector images and corresponding sample sector parameter information as sample training data to train and obtain a sector parameter prediction model.

[0204] In one feasible solution, the processing system further includes:

[0205] A standard model building module 5 is used to build a standard three-dimensional structure model of a preset object;

[0206] The first display control module 6 is used to control the display of the spatial position of the current standard point in the standard three-dimensional structure model; and / or the target sector parameter information at the position.

[0207] In an implementable solution, the sample sector parameter information includes the sample sector type, the sector parameter prediction model includes a sector classification model, and the target sector parameter information includes the target sector type;

[0208] and / or,

[0209] The sample sector parameter information includes sample structural anatomical information of different segmented regions in the sector, the sector parameter prediction model includes a segmentation model, and the target sector parameter information includes target contour information of different segmented regions in the sector and corresponding target structural anatomical information.

[0210] In one feasible solution, the processing system further includes:

[0211] A local area determination module 7 is used to determine a target local area in a preset object that matches a preset surgical procedure;

[0212] A local contour acquisition module 8 is used to obtain local contour information corresponding to the segmented area matched with the target local area;

[0213] Point position acquisition module 9, used to obtain point space position information of each contour point on the local contour corresponding to the local contour information;

[0214] The sector position acquisition module 10 is used to obtain sector spatial position information of the target sector corresponding to each frame of the target sector image;

[0215] The local model reconstruction module 11 is used to reconstruct a local three-dimensional structure model corresponding to the target local area based on the point space position information and the sector space position information.

[0216] In an practicable solution, the sector position acquisition module 10 is used to compare each frame of the target sector image with a number of sample sector images at the current standard point to obtain the corresponding similarity; the position of the sample sector image with the highest similarity is used as the sector space position information of the target sector image;

[0217] Alternatively, the sector position acquisition module 10 is used to input the target sector image of any frame into a pre-trained position estimation model to output sector spatial position information of the target sector.

[0218] In one feasible solution, the processing system further includes:

[0219] The information acquisition module 12 is used to obtain several standard points in the standard three-dimensional structure model that match the preset surgical procedure, as well as several standard sector images at different spatial positions at each standard point, and the standard sector spatial position information corresponding to each standard sector image.

[0220] The estimation model training module 13 is used to perform model training on each sample sector image and the corresponding standard sector spatial position information to obtain a position estimation model;

[0221] and / or,

[0222] The processing system also includes:

[0223] A second display control module 14 is used to control the display of the local three-dimensional structure model;

[0224] and / or,

[0225] The processing system also includes:

[0226] A reference information generating module 15 is configured to generate preoperative planning reference information based on at least one of the target sector image, target sector parameter information, and a local three-dimensional structure model;

[0227] and / or,

[0228] Preset objects include heart;

[0229] and / or,

[0230] The target sector image is an image acquired in a preoperative simulation environment.

[0231] Since the system embodiments generally correspond to the method embodiments, reference will be made to the description of the method embodiments for relevant details. The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components of the units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the disclosed solution.

[0232] Example 5

[0233] The present disclosure also provides an intracardiac ultrasound catheter device, which includes the medical image data processing system as described above.

[0234] The intracardiac ultrasound catheter device in this solution is integrated with the above-mentioned medical image data processing system, so that it can efficiently and high-quality output the sector parameter information of each frame of the target sector image, thereby providing reliable auxiliary information to the surgeon and other relevant personnel, assisting the surgeon in analyzing the heart structure, guiding the surgical process, etc., and achieving the purpose of helping the surgeon to conduct preoperative rehearsal planning and preoperative simulation operations in combination with AI algorithms without the need for additional hardware equipment, thereby improving the surgeon's operational proficiency, the success rate of the operation, and greatly reducing the cost of the operation; at the same time, it also reduces the product investment cost of the intracardiac ultrasound catheter device and improves the overall product performance.

[0235] Example 6

[0236] Figure 12 This is a structural diagram of an electronic device showing an example embodiment of the present disclosure. The electronic device includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the medical image data processing method described in any of the above embodiments. Figure 12 The electronic device 90 shown is only an example and should not limit the functionality and scope of use of the embodiments of the present disclosure.

[0237] like Figure 12 As shown, the electronic device 90 may be a general-purpose computing device, such as a server device. Components of the electronic device 90 may include, but are not limited to, the at least one processor 91, the at least one memory 92, and a bus 93 connecting different system components (including the memory 92 and the processor 91).

[0238] The bus 93 includes a data bus, an address bus, and a control bus.

[0239] The memory 92 may include a volatile memory, such as a random access memory (RAM) 921 and / or a cache memory 922 , and may further include a read-only memory (ROM) 923 .

[0240] The memory 92 may also include a program tool 925 (or utility) having a set (at least one) of program modules 924, such program modules 924 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include an implementation of a network environment.

[0241] The processor 91 executes various functional applications and data processing by running the computer program stored in the memory 92, such as the medical image data processing method provided in any of the above embodiments.

[0242] The electronic device 90 can also communicate with one or more external devices 94 (e.g., a keyboard, pointing device, etc.). This communication can occur via an input / output (I / O) interface 95. Furthermore, the electronic device 90 can communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 96. As shown, the network adapter 96 communicates with other modules of the electronic device 90 via a bus 93. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device 90, including but not limited to microcode, device drivers, redundant processors, external disk drive arrays, RAID (RAID) systems, tape drives, and data backup storage systems.

[0243] It should be noted that although several units / modules or sub-units / modules of the electronic device are mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0244] Example 7

[0245] An embodiment of the present disclosure further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for processing medical image data provided by any of the above embodiments.

[0246] The readable storage medium may include, but is not limited to, a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0247] Example 8

[0248] An embodiment of the present disclosure further provides a computer program product, comprising a computer program, which implements any of the above-mentioned methods for processing medical image data when executed by a processor.

[0249] The program code for executing the computer program product of the present disclosure may be written in any combination of one or more programming languages, and the program code may be executed entirely on the user device, partially on the user device, as a standalone software package, partially on the user device and partially on a remote device, or entirely on the remote device.

[0250] While specific embodiments of the present disclosure have been described above, those skilled in the art will appreciate that these are merely illustrative and that the scope of protection of the present disclosure is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present disclosure, and such changes and modifications are intended to fall within the scope of protection of the present disclosure.

Claims

1. A method for processing medical image data, characterized in that: The processing method comprises: Acquire a plurality of target sector images of a preset object at a current standard point under a preset surgical procedure; wherein the target sector images include intracardiac echocardiogram (ICE); Inputting each frame of the target sector image into a pre-trained sector parameter prediction model to output corresponding target sector parameter information; the step of constructing the sector parameter prediction model includes: Acquire several groups of historical sample data corresponding to the preset surgical procedure; Each set of historical sample data includes a plurality of sample sector images at different spatial positions at different preset points, and sample sector parameter information corresponding to each of the sample sector images is obtained; Using different sample sector images and corresponding sample sector parameter information as sample training data, training to obtain the sector parameter prediction model; The sample sector parameter information includes a sample sector type, the sector parameter prediction model includes a sector classification model, and the target sector parameter information includes a target sector type; The sample sector parameter information includes sample structural anatomical information of different segmented regions in the sector, the sector parameter prediction model includes a segmentation model, and the target sector parameter information includes target contour information of different segmented regions in the sector and corresponding target structural anatomical information; The processing method further comprises: Constructing a standard three-dimensional structural model of the preset object; After the step of outputting the corresponding target sector parameter information, the method further includes: Controlling the display of the spatial position of the current standard point in the standard three-dimensional structure model; and / or the target sector parameter information at the position; After the step of outputting the corresponding target sector parameter information, the method further includes: Determining a target local area in the preset object that matches the preset technique, and obtaining local contour information corresponding to the segmented area that matches the target local area; Obtaining point space position information of each contour point on the local contour corresponding to the local contour information; Obtaining sector spatial position information of the target sector corresponding to each frame of the target sector image; Based on the point space position information and the sector space position information, a local three-dimensional structure model corresponding to the target local area is reconstructed.

2. The method for processing medical image data according to claim 1, wherein: The step of obtaining the sector spatial position information of the target sector corresponding to each frame of the target sector image comprises: Comparing each frame of the target sector image with a plurality of the sample sector images at the current standard point to obtain corresponding similarities; Using the position of the sample sector image with the highest similarity as the sector space position information of the target sector image; Alternatively, the step of obtaining the sector spatial position information of the target sector corresponding to each frame of the target sector image includes: The target sector image of any frame is input into a pre-trained position estimation model to output the sector spatial position information of the target sector.

3. The method for processing medical image data according to claim 2, wherein: The processing method further comprises: Acquire a plurality of standard points in the standard three-dimensional structure model that match the preset surgical procedure, a plurality of standard sector images at different spatial positions at each of the standard points, and standard sector spatial position information corresponding to each of the standard sector images; The steps of training the position estimation model include: Performing model training on each of the standard sector images and the corresponding standard sector spatial position information to obtain the position estimation model; and / or, After the step of reconstructing the local three-dimensional structure model corresponding to the target local area, the method further includes: Controlling the display of the local three-dimensional structure model; and / or, The processing method further comprises: generating preoperative planning reference information based on at least one of the target sector image, the target sector parameter information, and the local three-dimensional structure model; and / or, The preset object includes a heart; and / or, The target sector image is an image acquired in a preoperative simulation environment.

4. A medical image data processing system, characterized in that: The processing system comprises: A target image acquisition module is used to acquire a plurality of target sector images of a preset object at a current standard point under a preset surgical procedure; wherein the target sector images include intracardiac echocardiography (ICE); A target parameter information output module is used to input each frame of the target sector image into a pre-trained sector parameter prediction model to output corresponding target sector parameter information; The processing system further comprises: A sample data acquisition module, used to acquire several groups of historical sample data corresponding to the preset procedure; Each set of historical sample data includes a plurality of sample sector images at different spatial positions at different preset points, and sample sector parameter information corresponding to each of the sample sector images is obtained; A model training module is used to use different sample sector images and corresponding sample sector parameter information as sample training data to train and obtain the sector parameter prediction model; The sample sector parameter information includes a sample sector type, the sector parameter prediction model includes a sector classification model, and the target sector parameter information includes a target sector type; The sample sector parameter information includes sample structural anatomical information of different segmented regions in the sector, the sector parameter prediction model includes a segmentation model, and the target sector parameter information includes target contour information of different segmented regions in the sector and corresponding target structural anatomical information; The processing system further comprises: A standard model building module, used to build a standard three-dimensional structural model of the preset object; A first display control module is configured to control the display of the spatial position of the current standard point in the standard three-dimensional structure model; and / or the target sector parameter information at the position; A local area determination module, configured to determine a target local area in the preset object that matches the preset surgical procedure; A local contour acquisition module is used to obtain local contour information corresponding to the segmented area matched with the target local area; a point position acquisition module, configured to acquire point spatial position information of each contour point on the local contour corresponding to the local contour information; A sector position acquisition module, configured to acquire sector spatial position information of a target sector corresponding to each frame of the target sector image; The local model reconstruction module is used to reconstruct a local three-dimensional structure model corresponding to the target local area based on the point space position information and the sector space position information.

5. An intracardiac ultrasound catheter device, characterized in that: The intracardiac ultrasound catheter device comprises the medical image data processing system according to claim 4.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and configured to run on the processor, wherein: When the processor executes the computer program, the method for processing medical image data according to any one of claims 1 to 3 is implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for processing medical image data according to any one of claims 1 to 3 is implemented.

Images