Vascular intervention real-time monitoring method and system based on intelligent sensing technology

Abstract

The invention relates to the technical field of image processing, in particular to a vascular intervention real-time monitoring method and system based on an intelligent sensing technology. Comprising the steps of collecting two-dimensional images of multiple angles in a target area within a preset time period, selecting a reference image, and obtaining blood vessel blood pressure; mapping the blood vessel blood pressure and the reference image to the same time coordinate axis, obtaining a dynamic period of the blood vessel blood pressure, and obtaining a blood vessel three-dimensional image and a three-dimensional feature at each relative position in the dynamic period based on the dynamic period and the reference image; acquiring a reference path from the catheter assembly to the preset position of the blood vessel by monitoring the real-time dynamic period of the blood pressure of the blood vessel and the first model; monitoring the movement track of the head of the catheter assembly in real time to obtain a path error; and inputting the path error, the blood vessel dynamic image at the preset position and the catheter assembly characteristics into a second model to obtain an interventional operation risk probability value. The problems of real-time monitoring and risk prediction of the vascular intervention operation are solved, and the accuracy and safety of the operation are improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a real-time monitoring method and system for vascular intervention based on intelligent sensing technology. Background Art

[0002] Vascular interventional surgery is a common minimally invasive treatment method, which is widely used in the treatment of cardiovascular diseases, cerebrovascular diseases and peripheral vascular diseases. However, the existing technology still has many shortcomings in vascular interventional surgery. There is a lack of real-time and accurate monitoring methods during the operation. Doctors mainly rely on imaging equipment for navigation, but these devices have radiation exposure risks and cannot provide real-time feedback on subtle changes inside the blood vessels. In addition, they still need to be improved in terms of accuracy, real-time and safety, and cannot meet the needs of complex surgical scenarios. Similar existing technologies include a Chinese patent application with publication number CN117426874A, which discloses a cardiovascular interventional robot control monitoring system, including a catheter system for delivering interventional instruments for vascular interventional surgery to a target target area through a patient's blood vessels, an image navigation system for constructing a three-dimensional image of the distribution of blood vessels in a patient's body, and a control system for controlling the catheter system to perform vascular interventional surgery; the three-dimensional image accuracy is improved by combining image fusion three-dimensional navigation and electromagnetic tracking navigation, and the artificial intelligence-assisted algorithm is improved accordingly, so that the catheter's path in the blood vessel is planned more intelligently, the risk of damage to the vascular wall caused by interventional surgery is avoided to the greatest extent, and the catheter is optimized to withstand stress. In addition, similar prior art includes a German patent with publication number DE112022000032T5, which discloses an unmanned interventional operating room system, including: a catheter room and a control room, with an observation window between the catheter room and the control room; a robot, wherein the catheter room is provided with an interventional surgery robot, a main robot, a puncture robot and a catheter guidewire exchange robot that cooperate with each other; a DSA device and a contrast agent injection device are installed on the catheter bed; a monitoring device, wherein the control room is provided with a monitoring device, which is communicated with the robot, the DSA device and the contrast agent injection device to provide information of each device and the robot, which is synchronized and updated in real time for monitoring and display by the doctor; a controller, wherein the controller for human-machine interaction between the doctor and the robot is arranged in the control room to realize manual monitoring, and through the combination with the robot, the purpose of unmanned interventional surgery is achieved, and the impact of interventional surgery on the health of the doctor is reduced. Both of the above patent documents realize the control or guidance of the surgical operation through the monitoring data obtained during the interventional surgery, thereby reducing the surgical risk, but the above two technical solutions do not take into account the dynamic changes of the interventional blood vessels and the operation coordination with the catheter, thereby further reducing the probability of surgical risk. Summary of the invention

[0003] The present application provides a real-time monitoring method for vascular intervention based on sensor technology, which aims to solve the problem of lack of real-time and accurate monitoring means in existing vascular intervention surgery. Through this technology, the accuracy and safety of surgery can be significantly improved and the risk of surgery can be reduced. The method includes:

[0004] Step S1: within a preset time period, a plurality of two-dimensional images at various angles in a target area are collected by an image acquisition unit, a reference image is selected from the plurality of two-dimensional images, and blood vessel blood pressure is also acquired by a monitoring unit in a catheter assembly;

[0005] Step S2: mapping the blood vessel blood pressure and the plurality of reference images onto the same time coordinate axis according to the acquisition time, obtaining a dynamic cycle of the blood vessel blood pressure, obtaining a three-dimensional image and three-dimensional features of the blood vessel at each relative position within the dynamic cycle based on the dynamic cycle and the plurality of reference images, and training a first model based on the dynamic cycle and the three-dimensional image and three-dimensional features of the blood vessel corresponding to each relative position;

[0006] Step S3: monitoring the real-time dynamic cycle of the blood pressure of the blood vessel, acquiring a three-dimensional dynamic image of the blood vessel based on the real-time dynamic cycle and the first model, and acquiring a reference path for the catheter assembly to reach a preset position of the blood vessel based on the three-dimensional dynamic image;

[0007] Step S4: when the catheter assembly reaches the preset position, the movement trajectory of the catheter assembly head is monitored in real time, and a path error is acquired based on the movement trajectory and the reference path;

[0008] Step S5: inputting the path error, the characteristics of the catheter assembly, and the three-dimensional image at the preset position into a second model to obtain an interventional surgery risk probability value, wherein the second model is an interventional surgery risk prediction model.

[0009] As a preferred technical solution of the present invention, after step S5, the following is further included:

[0010] When the interventional surgery risk probability value is greater than or equal to a set threshold, improving the control accuracy of the catheter assembly, and repeating the steps S3 and S4 to reacquire the path error, reacquire the interventional surgery risk probability value based on the reacquired path error and step S5, and repeating this step until the interventional surgery risk probability value is less than the set threshold;

[0011] When the interventional surgery risk probability value cannot be satisfied to be less than the set threshold value by improving the control accuracy of the catheter assembly, repeat steps S1-S2 to retrain the first model, and repeat steps S3-S5 to reacquire the interventional surgery risk probability value based on the reacquired first model.

[0012] As a preferred technical solution of the present invention, obtaining a three-dimensional image and three-dimensional features of a blood vessel at each relative position in the dynamic cycle based on the dynamic cycle and a plurality of reference images includes:

[0013] Mapping the blood vessel blood pressure and the reference image onto the same time axis according to the acquisition time, analyzing the change cycle of the blood vessel blood pressure value over time by a dynamic time warping algorithm, and using the change cycle as the dynamic cycle of the blood vessel blood pressure;

[0014] The multiple reference images are divided into groups according to the dynamic cycle, wherein the reference images in each group correspond to the same time in the same dynamic cycle, and blood vessel features are extracted from each reference image in the group, and all the blood vessel features corresponding to each group are integrated to obtain the three-dimensional blood vessel features of the group corresponding to each time in the dynamic cycle, wherein the blood vessel features include the diameter of the blood vessel, the centerline position of the blood vessel and the morphology of the blood vessel, and the three-dimensional blood vessel features include the three-dimensional information of the blood vessel.

[0015] As a preferred technical solution of the present invention, the training of the first model includes:

[0016] Based on the multiple dynamic cycles included in the preset time period and the three-dimensional images of the blood vessels at each relative position in each dynamic cycle and the corresponding three-dimensional features, and performing three-dimensional image and three-dimensional feature interpolation on the discontinuous positions of the three-dimensional images of the blood vessels between adjacent relative positions, continuous frames of the three-dimensional images of the blood vessels and the corresponding three-dimensional features are obtained, and the multiple dynamic cycles and the continuous frames of the three-dimensional images of the blood vessels corresponding to each dynamic cycle and the corresponding three-dimensional features are used as training data to train the first model.

[0017] As a preferred technical solution of the present invention, obtaining the reference path includes:

[0018] According to the real-time monitoring of the numerical change of the blood pressure of the blood vessel, the real-time dynamic cycle of the blood vessel blood pressure is obtained, and the real-time dynamic cycle is input into the first model to obtain a three-dimensional dynamic image of the blood vessel changing with time, and based on the three-dimensional dynamic image, the three-dimensional dynamic image is mapped to a three-dimensional coordinate system and the dynamic distribution of the blood vessel in the three-dimensional coordinate system is obtained, and the overlapping distribution area of ​​each relative position in the real-time dynamic cycle is calculated according to the dynamic distribution, and the first size of the overlapping distribution area is calculated. When the first size is greater than the second size of the catheter assembly head and the difference between the first size and the second size is greater than or equal to a set value, the center line of the overlapping distribution area is used as the reference path for the catheter assembly to reach the preset position. Otherwise, the dynamic center line of the blood vessel is obtained according to the three-dimensional dynamic image, and the dynamic center line is used as the reference path of the catheter assembly, and the forward speed of the catheter assembly is adjusted based on the change speed of the center line of the dynamic image.

[0019] As a preferred technical solution of the present invention, obtaining the path error includes:

[0020] The real-time image of the location of the catheter assembly head is tracked and acquired by an image acquisition unit in real time, and the motion trajectory of the catheter assembly head is extracted from the real-time image, and the maximum distance between the motion trajectory and the reference path in the real-time dynamic period is acquired, and the maximum distance is used as the path error.

[0021] As a preferred technical solution of the present invention, in step S5, the characteristics of the catheter assembly include the type of catheter, the rigidity of the corresponding guidewire, and the size and material of the catheter assembly and the guidewire.

[0022] As a preferred technical solution of the present invention, the training of the second model includes:

[0023] Historical surgical data is obtained from a storage unit, wherein the historical surgical data includes historical data of multiple patients, and the historical data includes characteristics of a catheter assembly used in the operation, multiple two-dimensional images of a target blood vessel, a planned path of the operation, an actual path of the operation, and a result of the operation. Based on the multiple two-dimensional images of the target blood vessel, a three-dimensional dynamic image corresponding to the target blood vessel is obtained through steps S1 to S3, and a historical path error is obtained through the planned path and the actual path. The path error, the three-dimensional dynamic image of the target blood vessel, and the characteristics of the catheter assembly are used as training data to train the second model.

[0024] The present invention also provides a real-time monitoring system for vascular intervention based on intelligent sensing technology, which is used to implement the above method. The system includes:

[0025] An image acquisition unit, used to collect a plurality of two-dimensional images at various angles in a target area within a preset time period, and select a reference image from the plurality of two-dimensional images;

[0026] a monitoring unit, located in the catheter assembly, for obtaining blood pressure in the blood vessel;

[0027] a calculation unit, configured to map the blood vessel blood pressure and the plurality of reference images onto the same time coordinate axis according to the acquisition time, obtain a dynamic cycle of the blood vessel blood pressure, and obtain a three-dimensional image and three-dimensional features of the blood vessel at each relative position within the dynamic cycle based on the dynamic cycle and the plurality of reference images;

[0028] A model training unit, configured to train a first model based on the dynamic cycle and the three-dimensional image and three-dimensional features of the blood vessel corresponding to each relative position;

[0029] The calculation unit is further used to monitor the real-time dynamic cycle of the blood pressure of the blood vessel, and obtain a three-dimensional dynamic image of the blood vessel based on the real-time dynamic cycle and the first model, and obtain a reference path for the catheter assembly to reach a preset position of the blood vessel based on the three-dimensional dynamic image; when the catheter assembly reaches the preset position, the movement trajectory of the head of the catheter assembly is monitored in real time, and a path error is obtained based on the movement trajectory and the reference path; the prediction unit is used to input the path error, the characteristics of the catheter assembly, and the three-dimensional image at the preset position into a second model to obtain an interventional surgery risk probability value, wherein the second model is an interventional surgery risk prediction model.

[0030] The present invention also provides a computer-readable storage medium, on which instructions are stored, and the above method is implemented when the instructions are executed by a processor.

[0031] Effect

[0032] The present invention collects two-dimensional images of multiple angles in the target area within a preset time period, and combines the monitoring unit in the catheter to obtain blood pressure in the blood vessels, thereby realizing multi-dimensional data collection during vascular intervention surgery. This multimodal data fusion method makes up for the deficiency of traditional imaging equipment in being unable to provide real-time feedback on subtle changes inside the blood vessels, and provides more comprehensive information support for precise navigation of surgery;

[0033] Secondly, the method of mapping vascular blood pressure and reference images onto the same time coordinate axis can accurately obtain the dynamic cycle of vascular blood pressure, and based on this, construct a three-dimensional image and three-dimensional features of the blood vessels. It not only considers the static structure of the blood vessels, but also pays attention to their dynamic changes, making the surgical path planning more in line with the actual physiological state of the blood vessels, thereby effectively reducing the risk of catheter damage to the vascular wall.

[0034] In addition, through the collaborative work of the first model and the second model, the system can monitor the movement trajectory of the catheter in real time, calculate the path error, and then predict the risk probability value of the interventional surgery. This real-time risk assessment mechanism provides an important decision-making reference for surgeons, making the surgical process more controllable. When the risk probability value exceeds the set threshold, the system can automatically adjust the control accuracy of the catheter, and even retrain the model to optimize the surgical path, further improving the safety and success rate of the operation; it also realizes real-time monitoring and risk prediction of vascular interventional surgery through intelligent sensing technology and advanced image processing algorithms, significantly improving the accuracy and safety of the operation. The widespread application of this technology is expected to promote vascular interventional surgery to develop in a more intelligent, precise and safe direction, and provide strong technical support for minimally invasive treatment of cardiovascular diseases, cerebrovascular diseases and peripheral vascular diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0036] Figure 1 This is a flow chart of a real-time monitoring method for vascular intervention based on intelligent sensor technology in an embodiment of the present application;

[0037] Figure 2 A flowchart of a method for calculating a blood vessel dynamic cycle, a blood vessel three-dimensional image, and three-dimensional features in an embodiment of the present application;

[0038] Figure 3 This is a flowchart of a reference path calculation method in an embodiment of the present application;

[0039] Figure 4 This is a structural diagram of a real-time monitoring system for vascular intervention based on intelligent sensor technology in an embodiment of the present application. DETAILED DESCRIPTION

[0040] The embodiment of the present application provides a method and system for real-time monitoring of vascular intervention based on intelligent sensor technology. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than the content illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0041] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 As shown, an embodiment of the real-time monitoring method for vascular intervention based on intelligent sensing technology in the embodiment of the present application includes:

[0042] Step S1: within a preset time period, a plurality of two-dimensional images at various angles in a target area are collected by an image acquisition unit, a reference image is selected from the plurality of two-dimensional images, and blood vessel blood pressure is also acquired by a monitoring unit in a catheter;

[0043] Specifically, when performing a vascular interventional surgery, a preset time period is first set, such as 10 minutes. During this time period, multiple two-dimensional images of various angles in the target vascular area are collected through an image acquisition unit (such as a multi-angle X-ray imaging device). It is assumed that an image is collected every 10° from 0° to 360°, and a total of 36 two-dimensional images are collected. Then, an image with the highest clarity and the most obvious vascular features is selected from these 36 two-dimensional images as a reference image. At the same time, vascular blood pressure data is obtained through a monitoring unit in the catheter (such as a pressure sensor). It is assumed that blood pressure data is collected every 0.1 second within 10 seconds, and a total of 100 blood pressure data points are collected. Among them, the above-mentioned monitoring unit can also monitor parameters such as blood flow velocity and temperature. Through the above-mentioned technical solution, high-quality reference images and corresponding vascular blood pressure can be obtained, laying the foundation for further obtaining dynamic cycles and vascular three-dimensional images.

[0044] Step S2: mapping the blood vessel blood pressure and the reference image onto the same time coordinate axis according to the acquisition time, obtaining the dynamic cycle of the blood vessel blood pressure, and obtaining the three-dimensional image and three-dimensional features of the blood vessel at each relative position in the dynamic cycle based on the dynamic cycle and each reference image;

[0045] Specifically, the 100 blood vessel blood pressure data points and the reference image acquired in step S1 are mapped to the same time coordinate axis according to the acquisition time, and the dynamic time bending algorithm is used to analyze the change cycle of the blood vessel blood pressure value over time. It is assumed that a dynamic cycle is 4 seconds, that is, a contraction and relaxation cycle of the heart. Then, multiple reference images are divided into groups according to the relative position of the acquisition time in the corresponding dynamic cycle. For example, the reference images with an acquisition time of 0-0.5 seconds are divided into one group, and those with an acquisition time of 0.5-1 seconds are divided into another group, and so on. Blood vessel features, such as blood vessel diameter, centerline position, etc., are extracted from each group of reference images, and each group of reference images and their corresponding blood vessel features are integrated to obtain a three-dimensional image and three-dimensional features of the blood vessel at each relative position in the dynamic cycle. Through the above technical solution, a foundation is laid for obtaining the relationship between the dynamic cycle of the blood vessel and the above three-dimensional dynamic image of the blood vessel.

[0046] Step S3: training a first model based on the dynamic cycle and the three-dimensional image and three-dimensional features of the blood vessel corresponding to each time, monitoring the real-time dynamic cycle of the blood vessel blood pressure, and acquiring a reference path for the catheter assembly to reach a preset position of the blood vessel based on the real-time dynamic cycle, the first model and the size of the catheter assembly;

[0047] Specifically, based on the dynamic cycle obtained in step S2 and the three-dimensional image and three-dimensional features of the blood vessel at each relative position, the first model is trained. It is assumed that the first model is a neural network model based on deep learning. During the operation, the numerical changes of the blood pressure of the blood vessel are monitored in real time to obtain the real-time dynamic cycle. The real-time dynamic cycle is input into the trained first model, and the model outputs a three-dimensional dynamic image of the blood vessel changing with time. The three-dimensional dynamic image is mapped to a three-dimensional coordinate system based on the three-dimensional dynamic image, and the dynamic distribution of the blood vessel in the three-dimensional coordinate system is obtained. The overlapping distribution area and its first size of each relative position in the real-time dynamic cycle are calculated according to the dynamic distribution, and the reference path is obtained according to the relationship between the first size of the overlapping distribution area and the second size of the catheter assembly. Through the above technical solution, the reference path of the catheter assembly to the preset position can be accurately obtained to avoid damage to the blood vessel.

[0048] Step S4: when the catheter assembly reaches the preset position, the movement trajectory of the catheter head is monitored in real time, and a path error is acquired based on the movement trajectory and the reference path;

[0049] Specifically, when the catheter reaches the preset position, the real-time image of the position of the catheter head is tracked and acquired in real time through the image acquisition unit. Assuming that an image is acquired every 0.05 seconds, the motion trajectory of the catheter head is extracted from the real-time image. The motion trajectory is compared with the reference path obtained in step S3 to obtain the maximum distance between the motion trajectory and the reference path within the real-time dynamic period. Assuming that the maximum distance is 0.2 mm, the maximum distance is used as the path error. Through the above technical solution, the path error of the above catheter assembly in the blood vessel can be accurately obtained, thereby laying a foundation for further obtaining the operation risk probability of the interventional surgery through the second model.

[0050] Step S5: inputting the path error, the dynamic image of the blood vessel at the preset position and the characteristics of the catheter into a second model to obtain an interventional surgery risk probability value, wherein the second model is an interventional surgery risk prediction model.

[0051] Specifically, the path error obtained in step S4, the dynamic image of the blood vessel at the preset position (including characteristics such as the blood vessel diameter and the centerline position), and the characteristics of the catheter (such as the catheter type is a certain model of vascular catheter, the guidewire rigidity is medium, the size and material of the catheter and guidewire are 2 mm in diameter, 100 cm in length, and the material is a medical polymer material) are input into the second model. It is assumed that the second model is a risk prediction model based on machine learning. After model calculation, the risk probability value of the interventional surgery is output. Through the above technical solution, the risk probability of the above interventional surgery can be accurately obtained, which is convenient for adjusting the control accuracy of the above catheter assembly in advance, thereby reducing the operational risk of the above interventional surgery.

[0052] Furthermore, after step S5, the method further includes:

[0053] When the interventional surgery risk probability value is greater than or equal to the set threshold, the control accuracy of the catheter is improved, and the steps S3 and S4 are repeated to reacquire the path error, and the interventional surgery risk probability value is reacquired based on the reacquired path error and step S5, and this step is repeated until the interventional surgery risk probability value is less than the set threshold;

[0054] When the interventional surgery risk probability value cannot be satisfied to be less than the set threshold value by improving the control accuracy of the catheter, repeat steps S1-S2 to retrain the first model, and repeat steps S3-S5 to reacquire the interventional surgery risk probability value based on the reacquired first model.

[0055] Specifically, when the risk probability of the interventional surgery is greater than or equal to the set threshold, that is, when the risk probability value of the interventional surgery is large, it is possible that the control accuracy of the catheter assembly is not high enough or the accuracy of the first model is not high enough, resulting in a large path error, thereby making the risk probability of the interventional surgery large. Since the control accuracy of the catheter assembly is easy to adjust, the control accuracy of the catheter assembly is first improved, that is, by increasing the first control frequency of the catheter assembly to the second control frequency, thereby adjusting the forward speed and direction of the catheter assembly, and repeating the above steps S3 and S4 to re-acquire the above path error, and based on the re-acquired path error and the above step S5, the risk probability value of the interventional surgery is obtained. When the risk probability value of the interventional surgery is still greater than or equal to the set threshold, repeat this step. The speed and control accuracy of the catheter assembly are made to satisfy that the risk probability value of the interventional surgery is less than or equal to the set threshold, wherein the set threshold is 90%. When the above conditions cannot be met by adjusting the accuracy of the catheter assembly, it may be that the accuracy of the first model is insufficient, resulting in a large path error. Therefore, by repeating steps S1 to S5, the first model is retrained. Since the training data of the second model is processed by the first model, the second model also needs to be synchronously updated based on the retrained first model, and the risk probability value of the interventional surgery is reacquired. Through the technical solution, the operational risk of the interventional surgery, i.e., the risk probability value of the interventional surgery, can be accurately predicted by adjusting the accuracy of the catheter assembly and the accuracy of the first model.

[0056] Furthermore, based on the dynamic cycle and the plurality of reference images, a three-dimensional image and three-dimensional features of a blood vessel at each relative position in the dynamic cycle are obtained, such as Figure 2 As shown, including:

[0057] Mapping the blood vessel blood pressure and the reference image onto the same time axis according to the acquisition time, analyzing the change cycle of the blood vessel blood pressure value over time by a dynamic time warping algorithm, and using the change cycle as the dynamic cycle of the blood vessel blood pressure;

[0058] The multiple reference images are divided into groups according to the relative positions of the acquisition time in the corresponding dynamic cycle, wherein the reference images in each group have the same relative positions in the corresponding dynamic cycle, and vascular features are extracted from each reference image in the group, and each reference image and the corresponding vascular features in each group are respectively integrated to obtain the vascular three-dimensional image and three-dimensional features at each relative position in the dynamic cycle corresponding to the group.

[0059] Specifically, since the coronary arteries will also contract and relax during the contraction and relaxation of the heart, in order to improve the operational accuracy of interventional surgery, it is necessary to grasp the dynamic changes of the blood vessels. Since the changes in vascular blood pressure correspond to the dynamic changes of the above-mentioned blood vessels, the blood pressure and the reference image are mapped to the same time axis according to the acquisition time, and the values ​​of the vascular blood pressure are connected to obtain a blood pressure value curve, and the change cycle of the vascular blood pressure over time is analyzed by a dynamic time bending algorithm, and the change cycle is used as the dynamic cycle of the vascular blood pressure. Since the dynamic cycle of the vascular blood pressure is synchronized with the dynamic changes of the blood vessels, the multiple reference images are divided into groups according to the relative positions of the acquisition time in the corresponding dynamic cycle, and the relative positions of the reference images in each of the above groups are the same, wherein the calculation of the relative position includes obtaining the corresponding The difference between the acquisition time of the reference image and the start time of the above-mentioned dynamic cycle, and the ratio of the above-mentioned difference to the corresponding dynamic cycle is used as the relative position of the above-mentioned reference image. Since the reference images at the same relative position in the above-mentioned dynamic cycle correspond to the same dynamic features of the above-mentioned blood vessels, and since different reference images in the same group correspond to the above-mentioned blood vessels at different angles, the vascular features of each reference image in the above-mentioned group are extracted, and the reference images at different angles are integrated to obtain the above-mentioned vascular three-dimensional image, and the vascular features of the blood vessels at multiple angles in the above-mentioned reference images are integrated to obtain the above-mentioned vascular three-dimensional features, wherein the above-mentioned vascular three-dimensional features include the diameter of the blood vessel, the centerline position of the blood vessel and the morphology of the blood vessel. Through the above-mentioned technical scheme, the vascular three-dimensional image and vascular three-dimensional features at each relative position of the above-mentioned blood vessel in the corresponding dynamic cycle can be accurately obtained, laying a foundation for the accurate training of the above-mentioned first model.

[0060] Furthermore, the training of the first model includes:

[0061] Based on the multiple dynamic cycles included in the preset time period and the three-dimensional images of the blood vessels at each relative position in each dynamic cycle and the corresponding three-dimensional features, and performing three-dimensional image and three-dimensional feature interpolation on the discontinuous positions of the three-dimensional images of the blood vessels between adjacent relative positions, continuous frames of the three-dimensional images of the blood vessels and the corresponding three-dimensional features are obtained, and the multiple dynamic cycles and the continuous frames of the three-dimensional images of the blood vessels corresponding to each dynamic cycle and the corresponding three-dimensional features are used as training data to train the first model.

[0062] Specifically, since the three-dimensional image of the blood vessel at each relative position in the above-mentioned dynamic cycle is integrated by the reference image in the two-dimensional image acquired by the image acquisition unit, and since the above-mentioned two-dimensional image is acquired periodically, the above-mentioned three-dimensional image of the blood vessel in the above-mentioned dynamic cycle is discontinuous. In order to provide sufficient and rich training data for the above-mentioned first model, the three-dimensional images of the blood vessels at adjacent relative positions of each dynamic cycle in the above-mentioned preset time period and the corresponding three-dimensional features are interpolated, so as to obtain continuous frames of the three-dimensional images of the blood vessels and the corresponding three-dimensional features, and use them as training data to train the above-mentioned first model. The above-mentioned dynamic cycle is also input into the above-mentioned first model, and based on the projections of the acquired three-dimensional dynamic images of the blood vessels at multiple angles, the above-mentioned projections are compared with the corresponding training data to judge the accuracy of the above-mentioned first model. When the accuracy of the above-mentioned first model is less than the set accuracy, the above-mentioned first model is retrained by repeating the above-mentioned steps S1 and S2. Through the above-mentioned technical solution, a high-precision blood vessel three-dimensional dynamic image prediction model, i.e., the first model, can be obtained, which provides a basis for accurately obtaining the reference path of the above-mentioned catheter assembly, thereby improving the control accuracy of interventional surgery and reducing surgical risks.

[0063] Further, the reference path is obtained, such as Figure 3 As shown, including:

[0064] According to the real-time monitoring of the numerical change of the blood pressure of the blood vessel, the real-time dynamic cycle of the blood vessel blood pressure is obtained, and the real-time dynamic cycle is input into the first model to obtain a three-dimensional dynamic image of the blood vessel changing with time, and based on the three-dimensional dynamic image, it is mapped to a three-dimensional coordinate system and the dynamic distribution of the blood vessel in the three-dimensional coordinate system is obtained, and the overlapping distribution area of ​​each relative position in the real-time dynamic cycle is calculated according to the dynamic distribution, and the first size of the overlapping distribution area is calculated. When the first size is greater than the second size of the catheter assembly head and the difference between the first size and the second size is greater than or equal to a set value, the center line of the overlapping distribution area is used as the reference path for the catheter assembly to reach the preset position. Otherwise, the dynamic center line of the blood vessel is obtained according to the three-dimensional dynamic image, and the dynamic center line is used as the reference path of the catheter assembly, and the forward speed of the catheter assembly is adjusted based on the change speed of the dynamic center line.

[0065] Specifically, by real-time monitoring of the changes in the blood pressure of the above-mentioned blood vessels, and obtaining the above-mentioned real-time dynamic cycle by the method of obtaining the dynamic cycle in step S2, and inputting the above-mentioned real-time dynamic cycle into the above-mentioned first model, and obtaining the above-mentioned three-dimensional dynamic image output by the above-mentioned first model, wherein the above-mentioned three-dimensional dynamic image is marked with corresponding three-dimensional features, and the above-mentioned three-dimensional dynamic image is mapped to the above-mentioned three-dimensional coordinate system, so as to facilitate quantification of the dynamic distribution of the above-mentioned blood vessels in the above-mentioned three-dimensional dynamic image, and obtain the overlapping distribution area of ​​each relative position according to the dynamic distribution of the above-mentioned blood vessels in the above-mentioned real-time dynamic cycle, and calculate the first size of the above-mentioned overlapping distribution area, that is, the cross-sectional size of the blood vessel corresponding to the above-mentioned overlapping distribution area, and compare the above-mentioned first size with the second size of the above-mentioned catheter assembly head, that is, the outer size of the catheter assembly head. When the above-mentioned first size is greater than the above-mentioned second size and the difference between the two is greater than the above-mentioned set value, at this time, the above-mentioned overlapping distribution area is used as a channel for the above-mentioned catheter assembly head to reach the above-mentioned preset position, and no Will cause damage to the above-mentioned blood vessel, for example: blood vessel perforation, etc., and the central position of the above-mentioned overlapping distribution area, that is, the center line, is used as the reference path of the head of the above-mentioned catheter assembly. Otherwise, when the above-mentioned first size is less than or equal to the above-mentioned second size or the difference between the two is less than the above-mentioned preset value, that is, when the head of the above-mentioned catheter assembly cannot reach the above-mentioned preset position through the above-mentioned overlapping distribution area, in order to avoid causing damage to the blood vessel in the process of guiding the head of the above-mentioned catheter assembly to reach the above-mentioned preset position, the dynamic center line of the above-mentioned blood vessel is used as the above-mentioned reference path, wherein the above-mentioned preset position is the surgical position, that is, the lesion position, because the above-mentioned dynamic center line, that is, the above-mentioned blood vessel, changes with time, therefore, by adjusting the forward speed of the above-mentioned catheter assembly head, the movement trajectory of the above-mentioned catheter assembly head is made to fit the above-mentioned dynamic center line to the maximum extent, wherein the above-mentioned forward speed includes the speed magnitude and direction. Through the above-mentioned technical solution, the reference path of the above-mentioned catheter assembly can be obtained, laying a foundation for further reducing the operational risk of interventional surgery.

[0066] Further, obtaining the path error includes:

[0067] The real-time image of the location of the catheter assembly head is tracked and acquired by an image acquisition unit in real time, and the motion trajectory of the catheter assembly head is extracted from the real-time image, and the maximum distance between the motion trajectory and the reference path in the real-time dynamic period is acquired, and the maximum distance is used as the path error.

[0068] Specifically, since the head of the catheter assembly is made of a material that is not penetrable by X-rays, the position and motion trajectory of the head of the catheter assembly can be identified through the real-time image, and the motion trajectory is compared with the reference path to obtain the distance between each position of the motion trajectory within the real-time dynamic period and the corresponding reference path, and the maximum value of the distances, i.e., the maximum distance, is used as the path error, wherein the path error reflects the control accuracy of the catheter assembly and indirectly reflects the operational error during the surgical operation. Through the technical solution, the path error of the head of the catheter assembly can be obtained in real time, laying a foundation for further accurately predicting the risk probability of interventional surgery through the path error.

[0069] Furthermore, in step S5, the characteristics of the catheter assembly include the type of catheter, the rigidity of the corresponding guidewire, and the size and material of the catheter and the guidewire.

[0070] Specifically, since the above surgical intervention risk probability value is the risk probability of successful surgical operation, the above path error reflects the operation accuracy of the above catheter assembly, and the above blood vessel dynamic image

[0071] Furthermore, the training of the second model includes:

[0072] Historical surgical data is obtained from a storage unit, wherein the historical surgical data includes historical data of multiple patients, and the historical data includes characteristics of a catheter assembly used in the operation, multiple two-dimensional images of a target blood vessel, a planned surgical path, an actual surgical path, and a surgical result; a three-dimensional dynamic image corresponding to the target blood vessel is obtained through steps S1 to S3 based on the multiple two-dimensional images of the target blood vessel; a historical path error is also obtained through the planned path and the actual path; the path error, the three-dimensional dynamic image of the target blood vessel, the characteristics of the catheter assembly, and the surgical result are used as training data to train the second model.

[0073] Specifically, multiple groups of training data are extracted from the above-mentioned historical surgical data, wherein the above-mentioned path error reflects the control accuracy of the above-mentioned catheter assembly, and the three-dimensional dynamic image of the above-mentioned target blood vessel can reflect the difficulty of the surgical operation at the above-mentioned preset position, that is, the patient's position, wherein the three-dimensional dynamic image of the above-mentioned target blood vessel is marked with the three-dimensional characteristics of the above-mentioned target blood vessel, and the above-mentioned target blood vessel is the target position of the interventional surgery, and the characteristics of the above-mentioned catheter assembly can reflect the accuracy of the surgical operation. Through the above-mentioned technical solution, an accurate interventional surgery risk probability prediction model can be obtained, thereby facilitating the accurate prediction of interventional surgery operation risks.

[0074] The present invention also provides a real-time monitoring system for vascular intervention based on intelligent sensing technology, which is used to implement the above method, such as Figure 4As shown, the system comprises:

[0075] An image acquisition unit, used to collect a plurality of two-dimensional images at various angles in a target area within a preset time period, and select a reference image from the plurality of two-dimensional images;

[0076] a monitoring unit, located in the catheter assembly, for obtaining blood pressure in the blood vessel;

[0077] a calculation unit, configured to map the blood vessel blood pressure and the plurality of reference images onto the same time coordinate axis according to the acquisition time, obtain a dynamic cycle of the blood vessel blood pressure, and obtain a three-dimensional image and three-dimensional features of the blood vessel at each relative position within the dynamic cycle based on the dynamic cycle and the plurality of reference images;

[0078] A model training unit, configured to train a first model based on the dynamic cycle and the three-dimensional image and three-dimensional features of the blood vessel corresponding to each relative position;

[0079] The calculation unit is further used for monitoring the real-time dynamic cycle of the blood pressure of the blood vessel, acquiring a three-dimensional dynamic image of the blood vessel based on the real-time dynamic cycle and the first model, and acquiring a reference path for the catheter assembly to reach a preset position of the blood vessel based on the three-dimensional dynamic image; in the process of the catheter assembly reaching the preset position, monitoring the motion trajectory of the catheter assembly head in real time, and acquiring a path error based on the motion trajectory and the reference path;

[0080] The prediction unit is used to input the path error, the characteristics of the catheter assembly, and the three-dimensional image at the preset position into a second model to obtain an interventional surgery risk probability value, wherein the second model is an interventional surgery risk prediction model.

[0081] The present invention also provides a computer-readable storage medium, on which instructions are stored, and the above method is implemented when the instructions are executed by a processor.

[0082] In summary, the present invention realizes multi-dimensional data collection of vascular interventional surgery by collecting two-dimensional images of multiple angles in the target area within a preset time period, and combining the monitoring unit in the catheter to obtain vascular blood pressure. This multimodal data fusion method makes up for the deficiency that traditional imaging equipment cannot feedback subtle changes in the blood vessels in real time, and provides more comprehensive information support for precise navigation of surgery; secondly, the method of mapping vascular blood pressure and reference image to the same time coordinate axis can accurately obtain the dynamic cycle of vascular blood pressure, and based on this, construct a three-dimensional image and three-dimensional features of the blood vessel, which not only considers the static structure of the blood vessel, but also pays attention to its dynamic changes, so that the surgical path planning is more in line with the actual physiological state of the blood vessel, thereby effectively reducing the risk of damage to the vascular wall by the catheter; in addition, through the collaborative work of the first model and the second model, the system can monitor the movement trajectory of the catheter in real time, calculate the path error, and then predict the risk probability value of interventional surgery. This real-time risk assessment mechanism provides an important decision-making reference for surgeons, making the surgical process more controllable. When the risk probability value exceeds the set threshold, the system can automatically adjust the control accuracy of the catheter or even retrain the model to optimize the surgical path. Through the mutual coordination of the above technical solutions, the accuracy and safety of the operation can be significantly improved and the surgical risk can be reduced.

[0083] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0084] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0085] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A real-time monitoring method for vascular intervention based on sensing technology, characterized in that: The method comprises: Step S1: within a preset time period, a plurality of two-dimensional images at various angles in a target area are collected by an image acquisition unit, a reference image is selected from the plurality of two-dimensional images, and blood vessel blood pressure is also acquired by a monitoring unit in a catheter assembly; Step S2: mapping the blood vessel blood pressure and the plurality of reference images onto the same time coordinate axis according to the acquisition time, obtaining a dynamic cycle of the blood vessel blood pressure, obtaining a three-dimensional image and three-dimensional features of the blood vessel at each relative position within the dynamic cycle based on the dynamic cycle and the plurality of reference images, and training a first model based on the dynamic cycle and the three-dimensional image and three-dimensional features of the blood vessel corresponding to each relative position; Step S3: training a first model based on the dynamic cycle and the three-dimensional image and three-dimensional features of the blood vessel corresponding to each time, monitoring the real-time dynamic cycle of the blood vessel blood pressure, and acquiring a reference path for the catheter assembly to reach a preset position of the blood vessel based on the real-time dynamic cycle and the first model; Step S4: when the catheter assembly reaches the preset position, the movement trajectory of the catheter assembly head is monitored in real time, and a path error is acquired based on the movement trajectory and the reference path; Step S5: inputting the path error, the dynamic image of the blood vessel at the preset position and the characteristics of the catheter assembly into a second model to obtain an interventional surgery risk probability value, wherein the second model is an interventional surgery risk prediction model.

2. The method according to claim 1, characterized in that After step S5, the method further includes: When the interventional surgery risk probability value is greater than or equal to a set threshold, improving the control accuracy of the catheter assembly, and repeating the steps S3 and S4 to reacquire the path error, reacquire the interventional surgery risk probability value based on the reacquired path error and step S5, and repeating this step until the interventional surgery risk probability value is less than the set threshold; When the interventional surgery risk probability value cannot be satisfied to be less than the set threshold value by improving the control accuracy of the catheter assembly, repeat steps S1-S2 to retrain the first model, and repeat steps S3-S5 to reacquire the interventional surgery risk probability value based on the reacquired first model.

3. The method according to claim 1, characterized in that Acquiring a three-dimensional image and three-dimensional features of a blood vessel at each relative position in the dynamic cycle based on the dynamic cycle and the plurality of reference images, including: Mapping the blood vessel blood pressure and the reference image onto the same time axis according to the acquisition time, analyzing the change cycle of the blood vessel blood pressure value over time by a dynamic time warping algorithm, and using the change cycle as the dynamic cycle of the blood vessel blood pressure; The multiple reference images are divided into groups according to the relative positions of the acquisition time in the corresponding dynamic cycle, wherein the reference images in each group have the same relative positions in the corresponding dynamic cycle, and vascular features are extracted from each reference image in the group, and each reference image and the corresponding vascular features in each group are respectively integrated to obtain the vascular three-dimensional image and three-dimensional features at each relative position in the dynamic cycle corresponding to the group.

4. The method according to claim 1, characterized in that The training of the first model includes: Based on the multiple dynamic cycles included in the preset time period and the three-dimensional images of the blood vessels at each relative position in each dynamic cycle and the corresponding three-dimensional features, and performing three-dimensional image and three-dimensional feature interpolation on the discontinuous positions of the three-dimensional images of the blood vessels between adjacent relative positions, continuous frames of the three-dimensional images of the blood vessels and the corresponding three-dimensional features are obtained, and the multiple dynamic cycles and the continuous frames of the three-dimensional images of the blood vessels corresponding to each dynamic cycle and the corresponding three-dimensional features are used as training data to train the first model.

5. The method according to claim 1, characterized in that The acquisition of the reference path includes: According to the real-time monitoring of the numerical change of the blood pressure of the blood vessel, the real-time dynamic cycle of the blood vessel blood pressure is obtained, and the real-time dynamic cycle is input into the first model to obtain a three-dimensional dynamic image of the blood vessel changing with time, and based on the three-dimensional dynamic image, the three-dimensional dynamic image is mapped to a three-dimensional coordinate system and the dynamic distribution of the blood vessel in the three-dimensional coordinate system is obtained, and the overlapping distribution area of ​​each relative position in the real-time dynamic cycle is calculated according to the dynamic distribution, and the first size of the overlapping distribution area is calculated. When the first size is greater than the second size of the catheter assembly head and the difference between the first size and the second size is greater than or equal to a set value, the center line of the overlapping distribution area is used as the reference path for the catheter assembly to reach the preset position. Otherwise, the dynamic center line of the blood vessel is obtained according to the three-dimensional dynamic image, and the dynamic center line is used as the reference path of the catheter assembly, and the forward speed of the catheter assembly is adjusted based on the change speed of the center line of the dynamic image.

6. The method according to claim 1, characterized in that The path error includes: The real-time image of the location of the catheter assembly head is tracked and acquired by an image acquisition unit in real time, and the motion trajectory of the catheter assembly head is extracted from the real-time image, and the maximum distance between the motion trajectory and the reference path in the real-time dynamic period is acquired, and the maximum distance is used as the path error.

7. The method according to claim 1, characterized in that In step S5, the characteristics of the catheter assembly include the type of catheter, the rigidity of the corresponding guidewire, and the size and material of the catheter and the guidewire.

8. The method according to claim 1, characterized in that The training of the second model comprises: Historical surgical data is obtained from a storage unit, wherein the historical surgical data includes historical data of multiple patients, and the historical data includes characteristics of a catheter assembly used in the operation, multiple two-dimensional images of a target blood vessel, a planned path of the operation, an actual path of the operation, and a result of the operation. Based on the multiple two-dimensional images of the target blood vessel, a three-dimensional dynamic image corresponding to the target blood vessel is obtained through steps S1 to S3, and a historical path error is obtained through the planned path and the actual path. The path error, the three-dimensional dynamic image of the target blood vessel, and the characteristics of the catheter assembly are used as training data to train the second model.

9. A real-time monitoring system for vascular intervention based on intelligent sensing technology, used to implement the method according to any one of claims 1 to 8, characterized in that: The system comprises: Step S1: within a preset time period, a plurality of two-dimensional images at various angles in a target area are collected by an image acquisition unit, a reference image is selected from the plurality of two-dimensional images, and blood vessel blood pressure is also acquired by a monitoring unit in a catheter assembly; Step S2: mapping the blood vessel blood pressure and the plurality of reference images onto the same time coordinate axis according to the acquisition time, obtaining a dynamic cycle of the blood vessel blood pressure, obtaining a three-dimensional image and three-dimensional features of the blood vessel at each relative position within the dynamic cycle based on the dynamic cycle and the plurality of reference images, and training a first model based on the dynamic cycle and the three-dimensional image and three-dimensional features of the blood vessel corresponding to each relative position; Step S3: monitoring the real-time dynamic cycle of the blood pressure of the blood vessel, acquiring a three-dimensional dynamic image of the blood vessel based on the real-time dynamic cycle and the first model, and acquiring a reference path for the catheter assembly to reach a preset position of the blood vessel based on the three-dimensional dynamic image; Step S4: when the catheter assembly reaches the preset position, the movement trajectory of the catheter assembly head is monitored in real time, and a path error is acquired based on the movement trajectory and the reference path; Step S5: inputting the path error, the characteristics of the catheter assembly, and the dynamic image of the blood vessel at the preset position into a second model to obtain an interventional surgery risk probability value, wherein the second model is an interventional surgery risk prediction model.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images