Coronary artery lesion detection method and device based on multi-dimensional dynamic image and medium

Abstract

The invention relates to a coronary artery lesion detection method and device based on a multi-dimensional dynamic image and a medium, and the method comprises the steps: obtaining the coronary artery dynamic motion information of a plurality of testees based on the coronary artery multi-dimensional dynamic images of the testees; performing principal component analysis on each piece of dynamic motion information to obtain a corresponding principal component contribution rate; constructing a coronary artery lesion detection model based on the contribution rate of each principal component and the lesion label of the coronary artery of the corresponding detected person; and inputting the principal component contribution rate corresponding to the coronary artery dynamic motion information of a person to be detected into the coronary artery lesion detection model to obtain a result whether the coronary artery lesion exists or not and a lesion degree result. The problems that coronary artery lesion diagnosis is invasive and low in precision and efficiency in the prior art are solved, and injury-free, efficient and accurate early diagnosis of coronary artery lesion is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of medical image processing, and particularly to a method, device and medium for detecting coronary artery lesions based on multi-dimensional dynamic images. Background Art

[0002] Coronary artery lesion refers to a series of pathological processes in which the morphology, function or structure of the coronary artery is abnormal due to various reasons, resulting in myocardial ischemia, injury or necrosis.

[0003] Currently, the diagnosis of coronary artery lesions is mainly carried out in the following ways:

[0004] (1) Diagnosing whether the coronary artery has lesions through invasive and non-invasive static image morphological information. Taking coronary angiography, which is regarded as the "gold standard", as an example, coronary angiography combined with intravascular ultrasound or optical coherence tomography is an invasive diagnosis, which may cause kidney injury due to contrast agents and may cause vascular injury during the operation. Other proposed non-invasive diagnoses are challenged by insufficient image quality caused by too high image heart rate and / or immature image noise reduction technology, and it is difficult to identify early coronary artery lesions from morphology.

[0005] (2) Diagnosing coronary artery lesions through vascular biomechanical parameters. This is mainly because biomechanics is closely related to the progression of coronary artery disease. This method focuses on exploring the mechanical and hemodynamic parameters of the coronary artery, such as wall shear stress, wall stress, strain and pressure, to enrich the detection index system for detecting coronary artery stenosis. However, it is difficult and complex to obtain mechanical parameters or / and hemodynamic parameters in the in-vivo environment. Generally, modeling and finite element simulation are used for in-vitro simulation, which is time-consuming, and there is a lack of a large amount of clinical data to establish the relationship between changes in mechanical parameters and coronary artery lesions.

[0006] In view of the problems of invasiveness, low accuracy and efficiency in the diagnosis of coronary artery lesions in the related art, no effective solution has been proposed yet. Summary of the Invention

[0007] Based on this, it is necessary to provide a method, device and medium for detecting coronary artery lesions based on multi-dimensional dynamic images for the above technical problems.

[0008] In a first aspect, an embodiment of the present application provides a method for detecting coronary artery lesions based on multi-dimensional dynamic images, the method comprising:

[0009] Obtaining the coronary artery dynamic motion information of each of the subjects based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects;

[0010] Perform principal component analysis on each of the dynamic motion information to obtain the corresponding principal component contribution rate;

[0011] Construct a coronary artery lesion detection model based on each of the principal component contribution rates and the lesion labels of the corresponding subjects' coronary arteries;

[0012] Input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion.

[0013] In one embodiment, the obtaining of the coronary artery dynamic motion information of each of the subjects based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects includes:

[0014] Collect the coronary artery images of the subject at multiple uniform phases, and select a reference phase image therefrom;

[0015] Annotate the reference phase image and select one or more target nodes;

[0016] Register the coronary artery images collected at the remaining phases with the reference phase image respectively, and obtain the spatial coordinates of each of the target nodes in the coronary artery images collected at each of the phases;

[0017] Obtain the coronary artery dynamic motion information of the subject based on the spatial coordinates of each of the target nodes in the coronary artery images collected at each of the phases.

[0018] In one embodiment, the annotating the reference phase image and selecting one or more target nodes includes:

[0019] Annotate the reference phase image, and use the intersection point of the center line of the coronary artery and the center line of the small branch as the target node.

[0020] In one embodiment, the lesion label includes the coronary artery stenosis value and the degree of stenosis. The constructing of the coronary artery lesion detection model based on each of the principal component contribution rates and the labels of the corresponding subjects' coronary arteries includes:

[0021] Construct a first coronary artery lesion detection model based on each of the principal component contribution rates and the corresponding coronary artery stenosis values of the subjects to detect whether there is a lesion;

[0022] Construct a second coronary artery lesion detection model based on the principal component contribution rates of the subjects with lesions and the corresponding degrees of coronary artery stenosis to detect the degree of the lesion.

[0023] In one embodiment, inputting the principal component contribution rate corresponding to the coronary artery dynamic motion information of the person to be tested into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion includes:

[0024] Input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the person to be tested into the first coronary artery lesion detection model to determine whether there is a lesion;

[0025] In the case of a lesion, input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the person to be tested into the second coronary artery lesion detection model to obtain the degree of the lesion.

[0026] In one embodiment, the first coronary artery lesion detection model is constructed by logistic regression:

[0027]

[0028] Wherein, is the first principal component contribution rate of the dynamic motion information, , is the fitting parameter; is the coronary artery stenosis value, used to determine whether there is a lesion;

[0029] The second coronary artery lesion detection model is constructed by linear regression:

[0030]

[0031] Wherein, is the degree of coronary artery stenosis, and a and b are fitting parameters.

[0032] In one embodiment, the coronary artery dynamic motion information includes at least one of the cumulative displacement, incremental displacement, velocity, and acceleration of the coronary artery during the cardiac cycle.

[0033] In one embodiment, the method further includes:

[0034] Input the coronary artery dynamic motion information of each of the tested persons and the lesion labels of the corresponding coronary arteries of the tested persons into the model for training to obtain a coronary artery lesion detection model, which includes but is not limited to a machine learning model, a deep learning model, and a virtual vector machine.

[0035] In a second aspect, an embodiment of the present application further provides a coronary artery lesion detection device based on multi-dimensional dynamic images, and the device includes:

[0036] A dynamic information acquisition module, configured to obtain the coronary artery dynamic motion information of each of the tested persons based on the multi-dimensional dynamic images of the coronary arteries of multiple tested persons;

[0037] A principal component analysis module, configured to perform principal component analysis on each of the dynamic motion information respectively to obtain corresponding principal component contribution rates;

[0038] A model construction module, configured to construct a coronary artery lesion detection model based on each of the principal component contribution rates and the lesion labels of the coronary arteries of the corresponding subjects to be measured;

[0039] A lesion detection module, configured to input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be measured into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion.

[0040] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, and wherein when the computer program is executed by a processor, the method described in the first aspect above is implemented.

[0041] The above coronary artery lesion detection method, device and medium based on multi-dimensional dynamic images obtain the coronary artery dynamic motion information of each of the subjects to be measured by using the multi-dimensional dynamic images of the coronary arteries of multiple subjects to be measured; perform principal component analysis on each of the dynamic motion information respectively to obtain corresponding principal component contribution rates; construct a coronary artery lesion detection model based on each of the principal component contribution rates and the lesion labels of the coronary arteries of the corresponding subjects to be measured; input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be measured into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion. This solves the problems of invasiveness, low accuracy and low efficiency in the diagnosis of coronary artery lesions in the related art, and realizes the early diagnosis of coronary artery lesions without injury, with high efficiency and accuracy.

[0042] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more concise and understandable. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The schematic embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0044] Figure 1 is a hardware structure block diagram of a terminal device for a coronary artery lesion detection method based on multi-dimensional dynamic images in an embodiment;

[0045] Figure 2 is a schematic flowchart of a coronary artery lesion detection method based on multi-dimensional dynamic images in an embodiment;

[0046] Figure 3It is a schematic flowchart of coronary artery lesion detection based on multi-dimensional dynamic images in a specific embodiment;

[0047] Figure 4 It is a structural block diagram of a coronary artery lesion detection device based on multi-dimensional dynamic images in an embodiment. Detailed implementation manners

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described and explained below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments provided in the present application without creative efforts fall within the scope of protection of the present application.

[0049] Obviously, the accompanying drawings in the following description are only some examples or embodiments of the present application. For those of ordinary skill in the art, without creative efforts, the present application can also be applied to other similar scenarios based on these drawings. In addition, it can also be understood that although the efforts made in this development process may be complex and lengthy, for those of ordinary skill in the art related to the content disclosed in the present application, some design, manufacturing or production changes based on the technical content disclosed in the present application are only conventional technical means and should not be understood as the content disclosed in the present application being insufficient.

[0050] The method embodiments provided in this embodiment can be executed on a terminal, a computer or a similar computing device. For example, running on a terminal, Figure 1 It is a hardware structural block diagram of the terminal of the coronary artery lesion detection method based on multi-dimensional dynamic images in this embodiment. As Figure 1 shown, the terminal may include one or more ( Figure 1 only one is shown in Figure 1 a processor 102 and a memory 104 for storing data. Among them, the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA. The above terminal may also include a transmission device 106 for communication functions and an input / output device 108. Those of ordinary skill in the art can understand that Figure 1 the structure shown in Figure 1 is only schematic and does not limit the structure of the above terminal. For example, the terminal may also include more or fewer components than

[0051] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for detecting coronary artery lesions based on multi-dimensional dynamic images in this embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, implements the above-mentioned method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely disposed relative to the processor 102, and these remote memories can be connected to the terminal through a network. Examples of the above-mentioned network include but are not limited to the Internet, intranet, local area network, mobile communication network, and combinations thereof.

[0052] The transmission device 106 is used to receive or send data via a network. The above-mentioned network includes a wireless network provided by a communication provider of the terminal. In one instance, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one instance, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0053] An embodiment of the present application provides a method for detecting coronary artery lesions based on multi-dimensional dynamic images. Taking the terminal to which this method is applied Figure 1 as an example for illustration, as Figure 2 shown, the method includes the following steps:

[0054] Step 201, based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects, obtain the coronary artery dynamic motion information of each of the subjects.

[0055] Specifically, a large set of multi-dimensional dynamic images of the coronary arteries of multiple subjects are obtained through clinical data. This image set should have the spatial dimension information of the coronary arteries, as well as the dynamic information of the change of the spatial information in the time dimension. These methods include but are not limited to ultrasonic imaging, magnetic resonance imaging, and multi-dimensional dynamic CTA imaging.

[0056] Preferably, this application uses multi-dimensional dynamic CTA imaging technology to obtain a multi-dimensional dynamic image set of the coronary artery. Multi-dimensional dynamic CTA imaging is an upgraded version based on CTA technology. Through multi-phase CTA scanning, it can provide three-dimensional (3D) vascular spatial coordinates and the changes of these information in the time dimension within the cardiac cycle. In actual operation, the patient (the person being measured) undergoes an electrocardiogram-guided multi-dimensional dynamic CTA image scan, and multiple phases with uniform intervals are reconstructed to obtain the dynamic motion information of the coronary artery during heart beating.

[0057] In this step, the dynamic motion information of the coronary artery during heart beating is quantified. The dynamic motion information in this application includes quantification results such as the cumulative displacement, incremental displacement, velocity, and acceleration of the coronary artery within the cardiac cycle.

[0058] Step 202: Perform principal component analysis on each of the dynamic motion information respectively to obtain the corresponding principal component contribution rate.

[0059] Principal component analysis is a method for dimensionality reduction of data, mainly used to reduce the data dimension while retaining as much original information as possible. It projects the original data onto a new coordinate system through linear transformation, so that the variance of the data is concentrated on fewer dimensions as much as possible.

[0060] This application uses principal component analysis to perform dimensionality reduction on the quantified dynamic motion information of the coronary artery, and extracts characteristic indicators such as principal component results and principal component contribution rates as diagnostic indicators for coronary artery lesions.

[0061] For an input row column incremental displacement data set as an example, where M is each phase and N is the incremental displacement of each target node in each dimension, the calculation process of its first principal component contribution rate is as follows:

[0062] Calculation of the centering matrix: ; is the incremental displacement vector corresponding to the i-th phase; is the mean vector of the incremental displacements of all phases of each target node in each dimension;

[0063] Calculation of the formula covariance matrix: ;

[0064] Perform singular value decomposition on the covariance matrix through to obtain the eigenvalue matrix and the eigenvector matrix ;

[0065] Extract the eigenvalue in the eigenvalue matrix where is the maximum eigenvalue, and through calculate the contribution rate of the first principal direction. The contribution rates of other principal components can all be obtained by the above method.

[0066] The contribution rates of the principal components of the dynamic motion information (such as cumulative displacement, velocity, acceleration, etc.) of the remaining coronary arteries are calculated by the same method.

[0067] Step 203: Based on the contribution rates of the respective principal components and the lesion labels of the coronary arteries of the corresponding subjects, construct a coronary artery lesion detection model.

[0068] Step 204: Input the contribution rate of the principal component corresponding to the dynamic motion information of the coronary artery of the subject to be tested into the coronary artery lesion detection model to obtain the result of whether there is a lesion and the degree of the lesion.

[0069] Specifically, obtain the multi-dimensional dynamic image of the coronary artery of the subject to be tested, calculate the dynamic motion information of the coronary artery of the subject to be tested, perform principal component analysis on the dynamic motion information of the coronary artery of the subject to be tested, obtain the contribution rate of the principal component of the dynamic motion information of the coronary artery of the subject to be tested, and input the contribution rate of the principal component into the established coronary artery lesion detection model, and the result of whether there is a lesion and the degree of the lesion can be obtained.

[0070] The above coronary artery lesion detection method, device and medium based on multi-dimensional dynamic images extract the dynamic motion information of the coronary artery through multi-dimensional dynamic images, quantify the dynamic motion information of the coronary artery, perform dimensionality reduction, analysis, and feature extraction on the quantified dynamic motion of the coronary artery through principal component analysis and statistical models, and establish the correlation with coronary artery lesions, realizing non-invasive, efficient and accurate early diagnosis of coronary artery lesions and the degree of lesions, and solving the problems of invasiveness, low accuracy and efficiency in the diagnosis of coronary artery lesions in the related art. Specifically, (1) Different from the traditional invasive diagnosis methods, this application uses non-invasive methods to extract the dynamic motion information of the heart to diagnose coronary artery lesions, and these information contain abnormal dynamic motions caused by coronary artery lesions. (2) Different from the traditional methods of using various methods such as finite element simulation to simulate the in-vivo vascular parameters, this method uses the method of principal component analysis to extract new diagnostic indicators from the dynamic motion data of the coronary artery, and can efficiently diagnose coronary artery lesions. (3) By establishing the correlation between the results of principal component analysis and the clinical diagnosis of coronary artery lesions and the degree of lesions, it assists in the early diagnosis of coronary artery lesions.

[0071] In one embodiment, obtaining the dynamic motion information of the coronary artery of each of the subjects based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects includes the following steps:

[0072] Step 301: Collect coronary artery images of the subject at multiple uniform phases, and select a reference phase image therefrom.

[0073] Exemplarily, select the phase with the best image quality and the clearest coronary artery as the reference phase image.

[0074] Step 302: Annotate the reference phase image and select multiple target nodes.

[0075] It can be understood that the selected target nodes are key landmark points used to indicate the dynamic movement process of the coronary artery.

[0076] In one embodiment, the intersection of the centerline of the coronary artery and the centerline of the small branch is used as the target node.

[0077] Step 303: Register the coronary artery images collected at the remaining phases with the reference phase image respectively, and obtain the spatial coordinates of each target node in the coronary artery images collected at each phase.

[0078] Among them, registration means aligning the coronary artery images collected at different phases so that they can be compared and analyzed in the same reference coordinate system.

[0079] Step 304: Obtain the coronary artery dynamic movement information of the subject based on the spatial coordinates of each target node in the coronary artery images collected at each phase.

[0080] In one embodiment, the lesion label includes the coronary artery stenosis value and the stenosis degree. The construction of the coronary artery lesion detection model based on each principal component contribution rate and the label of the corresponding subject's coronary artery includes the following:

[0081] Construct a first coronary artery lesion detection model based on each principal component contribution rate and the corresponding coronary artery stenosis value of the subject, which is used to detect whether a lesion occurs; construct a second coronary artery lesion detection model based on the principal component contribution rate of the subject with the lesion and the corresponding coronary artery stenosis degree, which is used to detect the lesion degree.

[0082] In one embodiment, the input of the principal component contribution rate corresponding to the coronary artery dynamic movement information of the subject to be measured into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the lesion degree includes the following:

[0083] Input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to the first coronary artery lesion detection model to determine whether there is a lesion; in the case of a lesion, input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to the second coronary artery lesion detection model to obtain the degree of the lesion.

[0084] In one embodiment, the first coronary artery lesion detection model is constructed by using logistic regression:

[0085]

[0086] Wherein, is the first principal component contribution rate of the dynamic motion information, , is the fitting parameter; is the coronary artery stenosis value, which is used to determine whether there is a lesion; bringing the first principal component contribution rate of the dynamic motion information into the logistic regression model, the probability of coronary artery stenosis can be predicted. Comparing the coronary artery stenosis value with the empirical threshold, if it is greater than or equal to the empirical threshold, it is determined that there is a lesion, otherwise, there is no lesion. Among them, the empirical threshold is the best threshold obtained by performing ROC curve analysis on the of the subject obtained after logistic regression and the binary label of whether the subject has a lesion obtained clinically.

[0087] It should be noted that the above formula only takes the first principal component contribution rate corresponding to the dynamic motion information as an example for illustration, and this application does not make specific limitations. In practical applications, the first coronary artery lesion detection model can also be obtained by performing logistic regression fitting based on the first K principal component contribution rates and the coronary artery stenosis value.

[0088] The second coronary artery lesion detection model is constructed by using linear regression:

[0089]

[0090] Wherein, is the degree of coronary artery stenosis, and a and b are fitting parameters.

[0091] It should be noted that the above formula only takes the first principal component contribution rate corresponding to the dynamic motion information as an example for illustration, and this application does not make specific limitations. In practical applications, the second coronary artery lesion detection model can also be obtained by performing linear regression fitting based on the first K principal component contribution rates and the degree of coronary artery stenosis.

[0092] In one embodiment, the coronary artery dynamic motion information in step 201 includes, but is not limited to, at least one of the cumulative displacement, incremental displacement, velocity, and acceleration of the coronary artery during the cardiac cycle. Principal component analysis is performed on each dynamic motion information respectively to obtain the corresponding principal component contribution rate.

[0093] The diagnosis of coronary artery lesions in this application can specifically analyze different coronary artery branches such as the left circumflex artery, left main artery, right coronary artery, and left anterior descending artery. Among them, the left anterior descending artery is one of the main branches of the coronary artery, responsible for supplying blood to important areas such as the left anterior wall of the heart and the front part of the left ventricle.

[0094] The following takes the example of obtaining the incremental displacement dynamic data of the left anterior descending artery of the subject to illustrate:

[0095] Coronary artery images of the subject during heart beating are collected at multiple uniform phases, and the phase with the best image quality and the clearest coronary artery is selected as the reference phase image; the reference phase image is labeled using medical image processing software, and the intersection point of the center line of the left anterior descending artery and the center line of the small branches is used as the target node, and there are multiple target nodes (other positions convenient for unified operation can also be selected as the target nodes); the registration algorithm is used to register the coronary artery images collected at the remaining phases with the reference phase image respectively, and the spatial coordinates of each target node in the coronary artery images collected at each phase are obtained; the coordinate data of all phases are sorted by time, and based on the spatial coordinates of each target node in the coronary artery images collected at each phase, the incremental displacement between adjacent phase images is calculated to obtain the incremental displacement dynamic data.

[0096] Through the spatial coordinates of each target node in the coronary artery images of each phase obtained above, other dynamic motion information of the coronary artery can also be calculated, such as dynamic motion information including the cumulative displacement, velocity, acceleration, etc. of the coronary artery during the cardiac cycle.

[0097] In a specific embodiment, as Figure 3 shown, the diagnosis of the coronary artery lesion and the degree of the lesion of the subject includes the following steps:

[0098] Step 401, acquisition of multi-dimensional dynamic images of the coronary artery of the subject.

[0099] Taking the multi-dimensional dynamic CTA image of the coronary artery as an example, the CTA imaging scan reconstructs 34 phases, and each phase has a uniform 3% R-R interval.

[0100] Step 402, quantification of the coronary artery dynamic motion information of the subject.

[0101] Taking the acquisition of incremental displacement data of the left anterior descending artery as an example: Select a reference phase for manual annotation to construct a reference left anterior descending artery model, use a registration algorithm to register the remaining phase image sets with the reference model, take the intersection of the centerline of the left anterior descending artery and the centerline of the small branch as the target node, obtain spatial coordinate data, and calculate the incremental displacement between adjacent phases.

[0102] Step 403, perform principal component analysis on the coronary artery dynamic motion information.

[0103] Step 404, determine whether there is a lesion according to the result of the principal component analysis.

[0104] Taking the contribution rate of the first principal direction as an example, establish the relationship between the coronary artery and the presence of lesions by using statistical methods such as a logistic regression model to determine whether there is a lesion.

[0105] Step 405, if the coronary artery is diseased, perform quantification of the lesion degree.

[0106] Establish the relationship between the coronary artery lesion degree by using statistical methods such as a linear regression model to determine the lesion degree.

[0107] In one embodiment, the method further includes: obtaining the coronary artery dynamic motion information of each of the plurality of subjects based on the multi-dimensional dynamic images of the coronary arteries of the subjects; inputting the coronary artery dynamic motion information of each subject and the lesion label of the corresponding subject's coronary artery into models such as a machine learning model, a deep learning model, and a virtual vector machine for training to obtain a coronary artery lesion detection model, extracting features from the coronary artery dynamic motion information through the model, and establishing the correlation with the coronary artery lesion to achieve non-invasive, efficient, and accurate early diagnosis of the coronary artery lesion and the lesion degree.

[0108] In one embodiment, the contribution rate of the principal component corresponding to the coronary artery dynamic motion information of each subject and the lesion label of the corresponding subject's coronary artery can also be input into the model for training to obtain a coronary artery lesion detection model, and the model includes but is not limited to a machine learning model, a deep learning model, and a virtual vector machine.

[0109] It should be noted that the steps shown in the above process or the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0110] In one embodiment, as Figure 4 shown, the embodiment of the present application further provides a coronary artery lesion detection device based on multi-dimensional dynamic images, and the device includes:

[0111] The dynamic information acquisition module 10 is configured to obtain the coronary artery dynamic motion information of each of the subjects based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects;

[0112] The principal component analysis module 20 is configured to perform principal component analysis on each of the dynamic motion information respectively to obtain the corresponding principal component contribution rate;

[0113] The model construction module 30 is configured to construct a coronary artery lesion detection model based on each of the principal component contribution rates and the lesion labels of the corresponding subjects' coronary arteries;

[0114] The lesion detection module 40 is configured to input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion.

[0115] In one embodiment, the dynamic information acquisition module 10 is further configured to: collect the coronary artery images of the subject at multiple uniform phases, and select a reference phase image therefrom; annotate the reference phase image, and select multiple target nodes; register the coronary artery images collected at the remaining phases with the reference phase image respectively, and obtain the spatial coordinates of each of the target nodes in the coronary artery images collected at each of the phases; and obtain the coronary artery dynamic motion information of the subject based on the spatial coordinates of each of the target nodes in the coronary artery images collected at each of the phases.

[0116] In one embodiment, the dynamic information acquisition module 10 is further configured to: annotate the reference phase image, and use the intersection point of the center line of the coronary artery and the center line of the small branch as the target node.

[0117] In one embodiment, the model construction module 30 is further configured to: construct a first coronary artery lesion detection model based on each of the principal component contribution rates and the coronary artery stenosis values of the corresponding subjects for detecting whether there is a lesion; and construct a second coronary artery lesion detection model based on the principal component contribution rates of the subjects with lesions and the corresponding coronary artery stenosis degrees for detecting the degree of the lesion.

[0118] In one embodiment, the lesion detection module 40 is further configured to input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the first coronary artery lesion detection model to determine whether there is a lesion; and in the case of a lesion, input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the second coronary artery lesion detection model to obtain the degree of the lesion.

[0119] In one embodiment, the first coronary artery lesion detection model is constructed by using logistic regression:

[0120]

[0121] Among them, is the first principal component contribution rate of the dynamic motion information, , is the fitting parameter; is the coronary artery stenosis value, which is used to judge whether there is a lesion;

[0122] The second coronary artery lesion detection model is constructed by linear regression:

[0123]

[0124] Among them, is the degree of coronary artery stenosis, and a and b are fitting parameters.

[0125] In one embodiment, the coronary artery dynamic motion information includes at least one of the cumulative displacement, incremental displacement, velocity, and acceleration of the coronary artery during the heartbeat cycle.

[0126] In one embodiment, the model construction module 30 is further configured to: input each of the principal component contribution rates and the lesion labels of the coronary arteries of the corresponding subjects to the model for training to obtain a coronary artery lesion detection model, and the model includes, but is not limited to, a machine learning model, a deep learning model, and a virtual vector machine.

[0127] It should be noted that the above-mentioned each module can be a functional module or a program module, and can be implemented either by software or by hardware. For the modules implemented by hardware, the above-mentioned each module can be located in the same processor; or the above-mentioned each module can also be located in different processors in any combined form.

[0128] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps in any one of the above-mentioned embodiments of the coronary artery lesion detection method based on multi-dimensional dynamic images are implemented.

[0129] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above various methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0130] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0131] The above embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for detecting coronary artery lesions based on multi-dimensional dynamic images, characterized in that The method includes: Based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects, obtaining the coronary artery dynamic motion information of each of the subjects; Performing principal component analysis on each of the dynamic motion information respectively to obtain the corresponding principal component contribution rate; Based on each of the principal component contribution rates and the lesion labels of the corresponding subjects' coronary arteries, constructing a coronary artery lesion detection model; Inputting the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion.

2. The method according to claim 1, wherein The obtaining the coronary artery dynamic motion information of each of the subjects based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects includes: Collecting the coronary artery images of the subject at multiple uniform phases and selecting a reference phase image therefrom; Annotating the reference phase image and selecting one or more target nodes; Registering the coronary artery images collected at the remaining phases with the reference phase image respectively and obtaining the spatial coordinates of each of the target nodes in the coronary artery images collected at each of the phases; Based on the spatial coordinates of each of the target nodes in the coronary artery images collected at each of the phases, obtaining the coronary artery dynamic motion information of the subject.

3. The method according to claim 2, wherein The annotating the reference phase image and selecting one or more target nodes includes: Annotating the reference phase image and taking the intersection point of the center line of the coronary artery and the center line of the small branch as the target node.

4. The method according to claim 1, characterized in that, The lesion label includes a binary label of whether there is stenosis and the degree of stenosis. The constructing a coronary artery lesion detection model based on each of the principal component contribution rates and the label of the corresponding subject's coronary artery includes: Based on each of the principal component contribution rates and the coronary artery stenosis value of the corresponding subject, constructing a first coronary artery lesion detection model for detecting whether there is a lesion; Based on the principal component contribution rate of the subjects with lesions and the corresponding degree of coronary artery stenosis, constructing a second coronary artery lesion detection model for detecting the degree of the lesion.

5. The method according to claim 4, characterized in that, The inputting the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion includes: Inputting the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the first coronary artery lesion detection model to judge whether there is a lesion; In the case of a lesion, inputting the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be tested into the second coronary artery lesion detection model to obtain the degree of the lesion.

6. The method according to claim 4, wherein Using logistic regression to construct the first coronary artery lesion detection model: Among them, is the contribution rate of the first principal component of the dynamic motion information, , is the fitting parameter; is the coronary artery stenosis value, which is used to judge whether a lesion occurs; Using linear regression to construct the second coronary artery lesion detection model: wherein, is the degree of coronary artery stenosis, and a and b are fitting parameters.

7. The method according to claim 1, characterized in that, The coronary artery dynamic motion information includes but is not limited to at least one of the cumulative displacement, incremental displacement, velocity, and acceleration of the coronary artery during the cardiac cycle.

8. The method according to claim 1, characterized in that, The method further includes: Input the coronary artery dynamic motion information of each of the subjects to be measured and the lesion labels corresponding to the coronary arteries of the subjects to be measured into a model for training to obtain a coronary artery lesion detection model, which includes but is not limited to a machine learning model, a deep learning model, and a virtual vector machine.

9. A coronary artery lesion detection device based on multi-dimensional dynamic images, characterized in that, The device includes: A dynamic information acquisition module, configured to obtain the coronary artery dynamic motion information of each of the subjects to be measured based on the multi-dimensional dynamic images of the coronary arteries of multiple subjects to be measured; A principal component analysis module, configured to perform principal component analysis on each of the dynamic motion information respectively to obtain the corresponding principal component contribution rate; A model construction module, configured to construct a coronary artery lesion detection model based on each of the principal component contribution rates and the lesion labels corresponding to the coronary arteries of the subjects to be measured; A lesion detection module, configured to input the principal component contribution rate corresponding to the coronary artery dynamic motion information of the subject to be measured into the coronary artery lesion detection model to obtain the results of whether there is a lesion and the degree of the lesion.

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

Images