DSA dynamic coronary artery image tracking method and system

Pre-trained machine learning model predicts the movement position of the coronary artery in medical images, solving the complications caused by contrast agent use in coronary angiography, realizing contrast agent-free diagnosis to ensure patient safety.

CN118021341BActive Publication Date: 2024-12-06THE SECOND XIANGYA HOSPITAL OF CENT SOUTH UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410354211.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-27
Publication Date
2024-12-06
Estimated Expiration
2044-03-27

AI Technical Summary

Technical Problem

During coronary angiography, the use of a large number of contrast agents leads to the occurrence of complications in patients and affects life safety. The existing technology has not effectively paid attention to and solved this problem.

Method used

By obtaining the patient's coronary artery medical images, respiratory movement data and heartbeat movement data, using a machine learning model pre-trained based on historical patient training data, predicting and displaying the movement position of the coronary artery in the medical image, avoiding contrast injection.

Benefits of technology

There is no need to inject contrast agent during interventional surgery to avoid complications, ensure the patient's life safety to the greatest extent, and accurately assist the doctor in diagnosing the coronary artery lesions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118021341B_ABST
    Figure CN118021341B_ABST
Patent Text Reader

Abstract

The present disclosure relates to a DSA dynamic coronary artery image tracking method and system, wherein the method comprises: obtaining a medical image and respiratory motion data and cardiac motion data of a patient's coronary artery; inputting the medical image and respiratory motion data and cardiac motion data into a target recognition model to predict the movement position of the patient's coronary artery in the medical image; the target recognition model is pre-trained based on training data, the training data comprises DSA angiography images of different sections of the coronary artery of historical patients and corresponding position annotation data, the DSA angiography images of different sections of the coronary artery are arranged based on time series, and the corresponding position annotation data of the DSA angiography images of different sections of the coronary artery are determined based on the intersection data of the respiratory motion cycle data and the cardiac motion cycle data of the historical patients; and displaying the movement position of the patient's coronary artery in the medical image.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The disclosed embodiments relate to the technical field of medical image processing, and in particular to a DSA dynamic coronary artery image tracking method and system. Background Art

[0002] Coronary arteries are the blood vessels that supply blood to the heart. Coronary angiography is a method for diagnosing heart diseases such as coronary heart disease, and the medical community calls it the "gold standard." Currently, coronary angiography can be used to examine and identify vascular lesions. The examination method is to inject contrast agents into the opening of the coronary angiography to visualize the coronary arteries. This can help doctors identify the location, degree, and range of coronary artery lesions, provide a basis for diagnosis, and facilitate the formulation of subsequent treatment plans.

[0003] At present, during the operation, when using angiography to identify vascular lesions, a large amount of contrast agent needs to be injected to make the coronary artery visible. However, the large amount of contrast agent used leads to a high probability of patient-related complications, affecting the patient's life safety, but the industry has not paid attention to and ignored this point. Summary of the invention

[0004] In order to solve the above technical problems or at least partially solve the above technical problems, the embodiments of the present disclosure provide a DSA dynamic coronary artery image tracking method and system.

[0005] In a first aspect, an embodiment of the present disclosure provides a DSA dynamic coronary artery image tracking method, comprising:

[0006] Obtain medical images of the patient's coronary arteries as well as respiratory motion data and heartbeat motion data;

[0007] The medical image and the respiratory motion data and the heartbeat motion data are input into a target recognition model to predict the motion position of the patient's coronary artery in the medical image; wherein the target recognition model is obtained by pre-training a machine learning model based on training data of multiple historical patients, the training data includes digital subtraction DSA angiography images of different sections of the coronary arteries of the historical patients and corresponding position annotation data, the DSA angiography images of different sections of the coronary arteries of the multiple historical patients are arranged based on a time series, and the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are determined based on the intersection data of the respiratory motion cycle data and the heartbeat motion cycle data of the historical patients;

[0008] The movement position of the patient's coronary artery in the medical image is displayed.

[0009] In one embodiment, the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are marked by bone marker coordinates, and the bone marker coordinates are determined by at least three bone marker points in the medical image containing the coronary arteries of the historical patient.

[0010] In one embodiment, the total number of segments of the different segments of the coronary artery is greater than or equal to a preset number of segments; and / or the clarity of the at least three bony markers is greater than a preset clarity.

[0011] In one embodiment, the heartbeat movement cycle data is determined by electrocardiogram waveform data of historical patients, and the respiratory movement cycle data is determined by thoracic movement waveform data of historical patients; the thoracic movement waveform data is determined by chest movement amplitude detection data of historical patients.

[0012] In one embodiment, the machine learning model includes at least a Transformer network model.

[0013] In one embodiment, the medical image of the patient's coronary arteries comprises at least an X-ray image.

[0014] In one embodiment, the position annotation data corresponding to the DSA angiography images of different sections of the coronary artery include axial center point position data of different sections of the coronary artery.

[0015] In a second aspect, the present disclosure provides a DSA dynamic coronary artery image tracking system, comprising:

[0016] An acquisition module, used for acquiring medical images of the patient's coronary arteries as well as respiratory motion data and heartbeat motion data;

[0017] A recognition module, used for inputting the medical image and the respiratory motion data and the heartbeat motion data into a target recognition model to predict the motion position of the patient's coronary artery in the medical image; wherein the target recognition model is obtained by pre-training a machine learning model based on training data of multiple historical patients, the training data includes digital subtraction DSA angiography images of different sections of the coronary arteries of the historical patients and corresponding position annotation data, the DSA angiography images of different sections of the coronary arteries of the multiple historical patients are arranged based on a time series, and the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are determined based on the intersection data of the respiratory motion cycle data and the heartbeat motion cycle data of the historical patients;

[0018] The tracking module is used to display the movement position of the patient's coronary artery in the medical image.

[0019] In a third aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the DSA dynamic coronary artery image tracking method described in any of the above embodiments.

[0020] In a fourth aspect, an embodiment of the present disclosure provides an electronic device, including:

[0021] Processor; and

[0022] Memory for storing computer programs;

[0023] Wherein, the processor is configured to execute the DSA dynamic coronary artery image tracking method described in any of the above embodiments by executing the computer program.

[0024] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:

[0025] The DSA dynamic coronary artery image tracking method and system provided in the embodiments of the present disclosure obtain medical images of the patient's coronary arteries and respiratory motion data and cardiac motion data; input the medical images and the respiratory motion data and cardiac motion data into a target recognition model to predict the movement position of the patient's coronary artery in the medical image; wherein the target recognition model is obtained by pre-training a machine learning model based on training data of multiple historical patients, the training data includes digital subtraction DSA angiography images of different sections of the coronary arteries of the historical patients and corresponding position annotation data, the DSA angiography images of different sections of the coronary arteries of the multiple historical patients are arranged based on a time series, and the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are determined based on the intersection data of the respiratory motion cycle data and the cardiac motion cycle data of the historical patients; and the movement position of the patient's coronary artery in the medical image is displayed. In this way, the target recognition model obtained by pre-training the model with vascular images obtained by the DSA angiography method of a large number of historical patients before the interventional surgery and position annotation data can recognize and predict the movement position of the coronary artery of the current patient in the medical image and display it to assist the doctor in diagnosis. In this way, there is no need to inject contrast agent during the interventional surgery of the current patient, thereby avoiding the occurrence of related complications caused by contrast agents and ensuring the patient's life safety to the greatest extent. In addition, the model training takes into account the impact of the superposition of respiratory and ECG data on the movement position of the coronary artery, subdivides the image of the coronary artery into different segments, and forms targeted training data based on the positions of different segments of the coronary artery marked in different movement cycles caused by the superposition of respiratory and ECG. The target recognition model trained with such training data can accurately identify and determine the position of the coronary artery of the current patient in the medical image, thereby facilitating the auxiliary doctor to accurately determine the pathological conditions of the coronary artery. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0028] Figure 1 This is a flow chart of the DSA dynamic coronary artery image tracking method according to an embodiment of the present disclosure;

[0029] Figure 2 is a schematic diagram of an image of a coronary artery according to an embodiment of the present disclosure;

[0030] Figure 3 This is a flow chart of a DSA dynamic coronary artery image tracking method according to another embodiment of the present disclosure;

[0031] Figure 4 Schematic diagram of a DSA dynamic coronary artery image tracking system according to an embodiment of the present disclosure;

[0032] Figure 5 Schematic diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0033] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0034] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0035] It should be understood that, in the following, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0036] Figure 1 The following is a flow chart of a DSA dynamic coronary artery image tracking method according to an embodiment of the present disclosure. The method may be executed by a computing device. The DSA dynamic coronary artery image tracking method may include the following steps:

[0037] Step S101: obtaining medical images of the patient's coronary arteries and respiratory motion data and cardiac motion data.

[0038] Exemplarily, the medical image may be an X-ray image of the patient, but is not limited thereto. In one embodiment, the heartbeat motion data may be determined by the electrocardiogram waveform data of the patient, the respiratory motion data may be determined by the chest motion waveform data of the patient, and the chest motion waveform data may be determined by the chest motion amplitude detection data of the patient. The acquisition of these data may be understood by referring to the prior art and will not be described in detail here.

[0039] Step S102: inputting the medical image and the respiratory motion data and the cardiac motion data into a target recognition model to predict the motion position of the patient's coronary artery in the medical image; wherein the target recognition model is obtained by pre-training a machine learning model based on training data of multiple historical patients, the training data includes digital subtraction DSA angiography images of different sections of the coronary arteries of historical patients and corresponding position annotation data, the DSA angiography images of different sections of the coronary arteries of the multiple historical patients are arranged based on a time series, and the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are determined based on the intersection data of the respiratory motion cycle data and the cardiac motion cycle data of the historical patients.

[0040] Exemplarily, the machine learning model can be a convolutional neural network model, but is not limited thereto. The training data can be constructed as follows: DSA angiography images of multiple historical patients are obtained from the hospital, such as X-ray images containing coronary arteries. The medical images of the coronary arteries of each historical patient, such as X-ray images, can include multiple X-ray images arranged in time sequence, that is, one moment corresponds to one X-ray image, and then each X-ray image of the coronary artery is divided as finely as possible into multiple vascular segment images, such as images of different segments such as the coronary trunk, bifurcation, and branches, and then the multiple vascular segment images corresponding to each X-ray image at each moment are annotated to obtain position annotation data. Due to the influence of breathing and heartbeat, the positions of vascular segments in images at different moments are usually different. Therefore, the position annotation data in this embodiment takes into account the common influencing factors of the patient's breathing and heartbeat, and the position annotation data is determined based on the intersection data of the respiratory motion cycle data and the heartbeat motion cycle data of the historical patient.

[0041] As an example, for example, the respiratory motion cycle data and cardiac motion cycle data of historical patients are obtained, and the respiratory motion curve and cardiac motion curve are fitted respectively. The horizontal coordinates of each curve correspond to different moments, and the vertical coordinates correspond to the vascular motion position at each moment. Then, multiple intersections of the respiratory motion curve and the cardiac motion curve are determined, and the moment and vascular motion position corresponding to each intersection are obtained as intersection data, and then the position annotation data is determined in combination with multiple vascular segment images corresponding to the X-ray image at each moment. It can be understood that the intersection can be a point where motion is enhanced or a point where motion is offset and weakened.

[0042] After the training data is constructed, training can be carried out. After the training is completed, a target recognition model can be obtained. When applied, the medical image of the current patient and the respiratory movement data and heartbeat movement data can be input into the target recognition model to predict the movement position of the coronary artery of the current patient in the medical image.

[0043] Step S103: Displaying the movement position of the patient's coronary artery in the medical image.

[0044] For example, the movement position of the patient's coronary artery is displayed in the current patient's X-ray image to assist doctors in diagnosing or operating the movement of the guidewire in the coronary artery to facilitate stent installation. That is, the vascular position area in the image is no longer obtained through contrast agent imaging such as subtraction enhancement, but the vascular position area is obtained based on image recognition, so that doctors can observe and confirm the lesion, such as Figure 2 The two arrows indicate the locations of coronary artery obstruction, i.e., the locations of lesions.

[0045] In the above scheme of this embodiment, the target recognition model obtained by pre-training the model with the vascular images obtained by the DSA angiography method of a large number of historical patients before the interventional surgery and the position annotation data can recognize and predict the movement position of the coronary artery of the current patient in the medical image and display it to assist the doctor in diagnosis. In this way, vascular DSA angiography can be performed without the injection of contrast agent during the interventional surgery of the current patient, thereby avoiding the occurrence of related complications caused by the use of contrast agents and ensuring the patient's life safety to the greatest extent. In addition, the influence of the superposition of respiratory and ECG data on the movement position of the coronary artery is considered during model training, and the image of the coronary artery is subdivided into different segments. Targeted training data is formed based on the positions of different segments of the coronary artery marked in different movement cycles caused by the superposition of respiratory and ECG. The target recognition model trained with such training data can accurately identify and determine the position of the coronary artery of the current patient in the medical image, thereby facilitating the assistance of doctors in accurately determining the pathological conditions of the coronary artery.

[0046] In order to more accurately identify and determine the position of the patient's coronary artery in the medical image, based on the above embodiment, in one embodiment, the position annotation data corresponding to the DSA angiography image of different sections of the coronary artery is marked by bony marker coordinates, and the bony marker coordinates are determined by at least three bony marker points in the medical image containing the coronary artery of the historical patient.

[0047] Specifically, the training data is labeled based on the coordinates of the bone markers in each X-ray image containing the coronary artery as the positioning coordinates, so that the positions of different segments of the marked coronary artery in different motion cycles (respiration and heartbeat superposition cycles) are more accurate. Among them, the bone marker coordinates can also be called bone landmark coordinates, and bone landmarks (bony landmarks) in the medical field usually refer to bones in certain parts of the human body, which often form obvious protrusions or depressions on the human body surface, called bone landmarks. For example, important bone landmarks include the xiphoid process and the sternal angle. The sternal angle is located at the junction of the sternal manubrium (another bone landmark) and the sternal body (another bone landmark). It is an important bone landmark about two horizontal fingers downward from the jugular notch (also a bone landmark). The xiphoid process is located at the lower end of the sternal body and plays a role in protecting the heart.

[0048] When an X-ray image of the coronary artery is taken in the DSA system, in addition to the coronary artery, there will be images of other human bones or parts such as some bony landmark images (not shown). In this embodiment, for the coronary artery at the heart position, the position of bony landmarks such as the xiphoid process in the X-ray image of the coronary artery can be selected as the bony landmark coordinates. In one example, the geometric center point of the xiphoid process or the geometric center point of the sternal manubrium can be selected as the bony landmark coordinates, but it is not limited to this and can be set as needed.

[0049] During labeling, the bone marker coordinates are used as positioning coordinates to mark the positions of different segments of the coronary artery in different motion cycles. During specific labeling, lines can be drawn on the blood vessel positions in the images of different segments of the coronary artery, such as the axial center points of the blood vessels, and then the corresponding multiple lines are connected together as position labels. In addition, due to the influence of the heartbeat and breathing of historical patients, the X-ray image of the coronary artery is subdivided into multiple segment images, and the coronary artery part in the segment image is marked at different motion cycles based on the bone marker coordinates, where different motion cycles are represented by expansion or contraction cycles of the diameter of the coronary artery. In different cycles of expansion or contraction, the relative position of the same coronary artery segment in the image is different. In this embodiment, the position marking of the coronary artery segment based on the bone marker coordinates is more in line with the actual situation and the marked position is more accurate. In other words, the constructed training data is more in line with the actual situation and the training data is more targeted and more accurate. In addition, the bone marker coordinates are determined by at least three bone marker points in the image rather than a single bone marker point, that is, multi-dimensional positioning coordinates are used for labeling, which can also make the position labeling data more precise and accurate. This makes the training data more precise and accurate as a whole, thereby making the training results of the trained target recognition model better, and further enabling the target recognition model to more accurately identify and determine the position of the coronary artery in the medical image, so as to accurately assist the doctor's diagnostic operations.

[0050] In order to further accurately identify and determine the position of the patient's coronary artery in the medical image, in one embodiment, the total number of segments of the different segments of the coronary artery is greater than or equal to a preset number of segments.

[0051] It should be noted that the setting of training data has an impact on the training results of the model. In this embodiment, in order to enable the trained target recognition model to more accurately identify and determine the position of the coronary artery in the medical image, the image of the coronary artery, such as the X-ray image, can be subdivided into different segments in the setting of training data, and the total number of divided segments is set to be greater than or equal to the preset number of segments, where the preset number of segments can be set as needed. That is, the vascular segments of the image of the coronary artery are divided as finely as possible, so that the position annotation data obtained by annotation is more precise and accurate, and the training data is further refined and accurate, so that the training results of the trained target recognition model are better, and the target recognition model can more accurately identify and determine the position of the coronary artery in the medical image, so as to accurately assist the doctor's diagnosis operation.

[0052] Furthermore, in order to further improve the accuracy of identifying and determining the position of the coronary artery in the medical image, in one embodiment, the clarity of the at least three bony markers is greater than a preset clarity.

[0053] Exemplarily, clarity refers to image clarity, and the preset clarity can be set as needed, and there is no limitation on this. In this embodiment, three bony marker points are taken as an example, that is, three different bony marker points. For example, the three clearest different bony markers in the field of view of the image of the coronary artery, such as the geometric center of the xiphoid process, sternal angle and sternal manubrium, are selected as three different bony marker points. The three bony marker points are used as positioning coordinates to annotate the different segment images of the sample coronary artery medical image, such as the X-ray image. Among them, the three clearest different bony markers can be obtained based on the existing relevant image processing algorithm. In this embodiment, the images of different sections of the coronary arteries corresponding to the medical images of the sample coronary arteries, such as X-ray images, can be annotated based on the three bony markers. The specific annotation process can refer to the aforementioned embodiment, so that images of different sections of the coronary arteries and corresponding accurate position annotation data can be formed, and then model training is performed based on the images of different sections of the coronary arteries and the corresponding position annotation data. In this way, the annotated training data can be made more accurate, so that the training results of the trained target recognition model can be further improved, and then the target recognition model can more accurately identify and determine the position of the coronary artery in the medical image.

[0054] In one embodiment, the heartbeat movement cycle data is determined by electrocardiogram waveform data of historical patients, and the respiratory movement cycle data is determined by thoracic movement waveform data of historical patients; the thoracic movement waveform data is determined by chest movement amplitude detection data of historical patients.

[0055] Specifically, as an example, the ECG waveform data can be displayed on the real-time screen of the DSA system by accessing the surface simulated lead ECG, and the cardiac cycle of the instantaneous coronary artery movement can be marked, and the chest movement waveform can be recorded by detecting the amplitude of chest movement during breathing. In this way, accurate ECG waveform data and chest movement waveform data can be obtained, and then accurate heartbeat movement cycle data and respiratory movement cycle data of historical patients can be obtained. Combined with the medical images of the coronary arteries of the corresponding historical patients, such as X-ray images, they can be divided and labeled, and accurate position labeling data can be obtained to form accurate training data required for training, so that the training results of the trained target recognition model are better, and then the target recognition model can accurately determine the accuracy of the position of the coronary artery in the medical image, and assist doctors in diagnostic operations such as guide wire movement.

[0056] In order to further improve the accuracy of identifying and determining the position of the coronary artery in medical images, in one embodiment, the machine learning model may at least include a Transformer network model. The Transformer network model is a neural network that processes time series data with an attention mechanism. The attention mechanism first controls the global information and then allocates attention resources. In this embodiment, the training of the target recognition model is based on time series data. Selecting the Transformer network model for training based on this feature can better exploit the characteristics of the time series data, so that the trained target recognition model can accurately identify and determine the position of the coronary artery in the medical image, that is, further improve the accuracy of identifying and determining the position of the coronary artery in the medical image.

[0057] In one embodiment, the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries include the axial center point position data of different sections of the coronary arteries. Since the cross-section of the arterial blood vessel is basically a circle, in order to accurately identify the position of the blood vessel segment, the axial center point position data of different sections of the coronary artery is used as the position annotation data. Of course, the upper and lower vertex positions of the blood vessel can also be used, and there is no limitation to this. The method in this embodiment can simply and conveniently obtain training data, and make the training data more accurate, so that the trained target recognition model can accurately identify and determine the position of the coronary artery in the medical image, that is, further improve the accuracy of identifying and determining the position of the coronary artery in the medical image.

[0058] In one example, in conjunction with reference Figure 3 As shown, the training process of the Transformer network model may include the following steps: S301, extracting the position feature vector in the medical image of the coronary artery of the historical patient, wherein the position feature vector represents the relative position between the guidewire position in the image of different segments of the coronary artery and the position of the different segments of the coronary artery in different motion cycles; S302, training the Transformer network model based on the position feature vector and the corresponding position annotation data, and the Transformer network model outputs the guidewire position prediction value; S303, updating the Transformer network model based on the difference between the guidewire position prediction value and the actual guidewire position until the training is terminated when the difference meets the preset condition.

[0059] Exemplarily, the Transformer network model may include an encoder and a decoder. The actual guidewire position may be determined based on the image annotation data. The preset condition may be satisfied when the difference value is less than the preset value, and the preset value may be set as needed. In this embodiment, during the model training process, the position feature vector of the relative position between the guidewire position (i.e., the arterial blood vessel reference position) in the image representing different segments of the coronary artery and the position of different segments of the coronary artery in different motion cycles is first extracted, and then the Transformer network model is trained in combination with the corresponding position annotation data, such as the position of different segments of the coronary artery marked based on the coordinates of the bony markers in different motion cycles, such as by fusion and comparison training of the position feature vector and the feature vector of the dynamically changing position of different segments of the coronary artery in the position annotation data, so that the control of the training process enables the trained target recognition model to more accurately identify and determine the position of the coronary artery in the medical image, so as to assist the doctor in accurate diagnosis.

[0060] It should be noted that, although the steps of the method in the present disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple steps for execution, etc. In addition, it is also easy to understand that these steps may be, for example, executed synchronously or asynchronously in multiple modules / processes / threads.

[0061] like Figure 4 As shown, the embodiment of the present disclosure provides a DSA dynamic coronary artery image tracking system, comprising:

[0062] An acquisition module 401 is used to acquire medical images of the patient's coronary arteries and respiratory motion data and heartbeat motion data;

[0063] The recognition module 402 is used to input the medical image and the respiratory motion data and the heartbeat motion data into a target recognition model to predict the motion position of the patient's coronary artery in the medical image; wherein the target recognition model is obtained by pre-training a machine learning model based on training data of multiple historical patients, the training data includes digital subtraction DSA angiography images of different sections of the coronary arteries of the historical patients and corresponding position annotation data, the DSA angiography images of different sections of the coronary arteries of the multiple historical patients are arranged based on a time series, and the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are determined based on the intersection data of the respiratory motion cycle data and the heartbeat motion cycle data of the historical patients;

[0064] The tracking module 403 is used to display the movement position of the patient's coronary artery in the medical image.

[0065] In one embodiment, the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are marked by bone marker coordinates, and the bone marker coordinates are determined by at least three bone marker points in the medical image containing the coronary arteries of the historical patient.

[0066] In one embodiment, the total number of segments of the different segments of the coronary artery is greater than or equal to a preset number of segments; and / or the clarity of the at least three bony markers is greater than a preset clarity.

[0067] In one embodiment, the heartbeat movement cycle data is determined by electrocardiogram waveform data of historical patients, and the respiratory movement cycle data is determined by thoracic movement waveform data of historical patients; the thoracic movement waveform data is determined by chest movement amplitude detection data of historical patients.

[0068] In one embodiment, the machine learning model includes at least a Transformer network model.

[0069] In one embodiment, the medical image of the patient's coronary arteries comprises at least an X-ray image.

[0070] In one embodiment, the position annotation data corresponding to the DSA angiography images of different sections of the coronary artery include axial center point position data of different sections of the coronary artery.

[0071] In one example, the training process of the Transformer network model may include: extracting position feature vectors from medical images of coronary arteries of historical patients, wherein the position feature vectors represent the relative position between the guidewire position in the images of different segments of the coronary artery and the positions of the different segments of the coronary artery in different motion cycles; training the Transformer network model based on the position feature vectors and corresponding position annotation data, and outputting a guidewire position prediction value from the Transformer network model; and updating the Transformer network model based on the difference between the guidewire position prediction value and the actual guidewire position until the training is terminated when the difference meets a preset condition.

[0072] Regarding the system in the above embodiment, the specific manner in which each module performs operations and the corresponding technical effects brought about have been described in detail in the embodiment of the method, and will not be elaborated here.

[0073] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the present disclosure, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of a module or unit described above can be further divided into multiple modules or units for concretization. The components displayed as modules or units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. Those of ordinary skill in the art can understand and implement it without paying creative work.

[0074] The embodiments of the present disclosure further provide a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the DSA dynamic coronary artery image tracking method described in any one of the above embodiments is implemented.

[0075] Exemplarily, the readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0076] The computer readable storage medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by an instruction execution system, an apparatus, or a device or used in combination with it. The program code contained on the readable storage medium may be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0077] The present disclosure also provides an electronic device, including a processor and a memory, wherein the memory is used to store a computer program, wherein the processor is configured to execute the DSA dynamic coronary artery image tracking method in any of the above embodiments by executing the computer program.

[0078] Refer to the following Figure 5The electronic device 600 according to this embodiment of the present invention is described. Figure 5 The electronic device 600 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0079] like Figure 5 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different system components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0080] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 performs the steps according to various exemplary embodiments of the present invention described in the above method embodiment section of this specification. For example, the processing unit 610 can perform the following steps: Figure 1 The steps of the method shown in .

[0081] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0082] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 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 the implementation of a network environment.

[0083] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0084] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also 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 660. The network adapter 660 may communicate with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0085] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation methods of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, or a network device, etc.) to execute the method steps of the above-mentioned embodiments according to the implementation methods of the present disclosure.

[0086] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0087] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A DSA dynamic coronary artery image tracking method, characterized in that: include: Acquire a medical image of the patient's coronary arteries and respiratory motion data and heartbeat motion data; wherein the medical image of the patient's coronary arteries at least includes an X-ray image; The medical image and the respiratory motion data and the heartbeat motion data are input into a target recognition model to predict the movement position of the patient's coronary artery in the medical image; wherein the target recognition model is obtained by pre-training a machine learning model based on training data of multiple historical patients, the training data includes digital subtraction DSA angiography images of different sections of the coronary arteries of historical patients and corresponding position annotation data, the DSA angiography images of different sections of the coronary arteries of the multiple historical patients are arranged based on time series, and the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are based on the respiratory motion cycle of the historical patients The position marking data is determined based on the intersection data of the respiratory movement cycle data and the cardiac movement cycle data of the historical patients, including: fitting the respiratory movement curve and the cardiac movement curve based on the historical patients' respiratory movement cycle data and the cardiac movement cycle data, respectively, the horizontal coordinates of each curve correspond to different moments, and the vertical coordinates correspond to the blood vessel movement position at each moment; determining multiple intersections of the respiratory movement curve and the cardiac movement curve, obtaining the moment and blood vessel movement position corresponding to each intersection as the intersection data, and then determining the position marking data in combination with the corresponding blood vessel segment image in the DSA angiography image at each moment; The movement position of the patient's coronary artery in the medical image is displayed.

2. The method according to claim 1, characterized in that The position annotation data corresponding to the DSA angiography images of different sections of the coronary artery are marked by bone marker coordinates, and the bone marker coordinates are determined by at least three bone marker points in the medical image containing the coronary artery of the historical patient.

3. The method according to claim 2, characterized in that The total number of segments of the different segments of the coronary artery is greater than or equal to a preset number of segments; and / or the clarity of the at least three bony markers is greater than a preset clarity.

4. The method according to any one of claims 1 to 3, characterized in that The heartbeat movement cycle data is determined by electrocardiogram waveform data of historical patients, and the respiratory movement cycle data is determined by chest movement waveform data of historical patients; the chest movement waveform data is determined by chest movement amplitude detection data of historical patients.

5. The method according to claim 4, characterized in that The machine learning model at least includes a Transformer network model.

6. The method according to claim 4, characterized in that The position annotation data corresponding to the DSA angiography images of different sections of the coronary artery include the axial center point position data of different sections of the coronary artery.

7. A DSA dynamic coronary artery image tracking system, characterized in that: include: An acquisition module, used to acquire medical images of the patient's coronary arteries and respiratory motion data and cardiac motion data; wherein the medical images of the patient's coronary arteries at least include X-ray images; The recognition module is used to input the medical image and the respiratory motion data and the heartbeat motion data into a target recognition model to predict the movement position of the patient's coronary artery in the medical image; wherein the target recognition model is obtained by training a machine learning model based on training data of multiple historical patients in advance, and the training data includes digital subtraction DSA angiography images of different sections of the coronary arteries of historical patients and corresponding position annotation data, and the DSA angiography images of different sections of the coronary arteries of the multiple historical patients are arranged based on time series, and the position annotation data corresponding to the DSA angiography images of different sections of the coronary arteries are based on the respiratory motion data of the historical patients. The position marking data is determined based on the intersection data of the respiratory movement cycle data and the cardiac movement cycle data of the historical patients, including: based on the respiratory movement cycle data and the cardiac movement cycle data of the historical patients, respectively fitting the respiratory movement curve and the cardiac movement curve, the horizontal coordinate in each curve corresponds to different time, and the vertical coordinate corresponds to the blood vessel movement position at each time; determining multiple intersections of the respiratory movement curve and the cardiac movement curve, obtaining the time and blood vessel movement position corresponding to each intersection as the intersection data, and then combining the corresponding blood vessel segment image in the DSA angiography image at each time to determine the position marking data; The tracking module is used to display the movement position of the patient's coronary artery in the medical image.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the DSA dynamic coronary artery image tracking method according to any one of claims 1 to 6 is implemented.

9. An electronic device, characterized in that: include: processor; as well as Memory for storing computer programs; Wherein, the processor is configured to execute the DSA dynamic coronary artery image tracking method according to any one of claims 1 to 6 by executing the computer program.