Blood vessel stenosis detection method, device and system based on angiography image segmentation
By adopting a combined segmentation-detection model in coronary stenosis detection, combining image segmentation and stenosis detection, the problem of low detection accuracy in the prior art is solved, and higher detection accuracy is achieved.
Patent Information
- Application Number
- CN202410999451.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-12-10
AI Technical Summary
The existing coronary stenosis detection methods are less accurate in complex vascular morphology and inconsistent contrast, resulting in difficulty in diagnosis.
A combined segmentation-detection model based on contrast images is adopted, combining the image segmentation model and the stenosis detection model, and network parameters are updated through backpropagation loss, so as to achieve the combination of vascular segmentation and stenosis detection, and improve detection accuracy.
Improve the accuracy of coronary stenosis detection, and higher recognition capabilities compared with conventional methods through more accurate vascular segmentation and stenosis detection.
Smart Images
Figure CN119107451A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of image processing, and in particular relates to a method, device and system for detecting vascular stenosis based on angiography image segmentation. Background Art
[0002] Segment identification of angiographic images is an important technique in cardiovascular imaging that allows physicians to accurately identify and analyze different segments of the coronary arteries. During an angiogram, a special dye or contrast agent is usually used to clearly show the blood vessels of the heart. These images are then used to evaluate the presence, extent, and location of coronary artery disease.
[0003] Coronary artery disease is a common form of cardiovascular disease, which can lead to serious consequences such as myocardial infarction and angina pectoris. Therefore, accurate diagnosis of coronary artery stenosis is crucial for the treatment and health management of patients. However, due to factors such as complex vascular morphology, inconsistent contrast and imaging, the commonly used coronary artery stenosis detection methods have certain limitations and the accuracy of stenosis detection is low. Summary of the invention
[0004] In view of this, the present invention provides a method, device and system for detecting vascular stenosis based on angiography image segmentation, aiming to solve the problem of low accuracy of angiography image stenosis detection in the prior art.
[0005] A first aspect of an embodiment of the present invention provides a method for detecting vascular stenosis based on angiography image segmentation, comprising: Obtaining angiographic images; The angiography image is input into the joint segmentation-detection model to obtain the stenosis detection result of the blood vessel; Among them, the joint segmentation-detection model includes an image segmentation model and a stenosis detection model; the detection loss of the stenosis detection model is back-propagated to update the network parameters of the stenosis detection model; the segmentation loss of the image segmentation model is back-propagated; the back-propagated detection loss and segmentation loss are used to update the network parameters of the image segmentation model; the input of the image segmentation model is angiography image, and the output is vascular segmentation image; the input of the stenosis detection model is angiography image and vascular segmentation image, and the output is stenosis detection result.
[0006] A second aspect of an embodiment of the present invention provides a vascular stenosis detection device based on angiography image segmentation, comprising: An acquisition module, used for acquiring angiography images; A detection module, used for inputting the angiography image into the joint segmentation-detection model to obtain a stenosis detection result of the blood vessel; Among them, the joint segmentation-detection model includes an image segmentation model and a stenosis detection model; the detection loss of the stenosis detection model is back-propagated to update the network parameters of the stenosis detection model; the segmentation loss of the image segmentation model is back-propagated; the back-propagated detection loss and segmentation loss are used to update the network parameters of the image segmentation model; the input of the image segmentation model is angiography image, and the output is vascular segmentation image; the input of the stenosis detection model is angiography image and vascular segmentation image, and the output is stenosis detection result.
[0007] A third aspect of an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the vascular stenosis detection method based on angiography image segmentation as described in the first aspect above are implemented.
[0008] A fourth aspect of an embodiment of the present invention provides an angiography image processing system, comprising a vascular angiography device and the electronic device according to the third aspect above.
[0009] A fifth aspect of an embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the steps of the vascular stenosis detection method based on angiography image segmentation as described in the first aspect above are implemented.
[0010] The embodiment of the present invention provides a method, device and system for detecting vascular stenosis based on angiography image segmentation. First, angiography images are obtained; then the angiography images are input into a joint segmentation-detection model to obtain vascular stenosis detection results; wherein the joint segmentation-detection model includes an image segmentation model and a stenosis detection model; the detection loss of the stenosis detection model is back-propagated to update the network parameters of the stenosis detection model; the segmentation loss of the image segmentation model is back-propagated; the back-propagated detection loss and segmentation loss are used to update the network parameters of the image segmentation model; the input of the image segmentation model is angiography images, and the output is vascular segmentation images; the input of the stenosis detection model is angiography images and vascular segmentation images, and the output is stenosis detection results. The present invention uses an image segmentation model and a stenosis detection model to detect vascular stenosis, and at the same time assists in segmentation by transferring the detection loss of the stenosis detection model to the image segmentation model, so that the image segmentation model can obtain a better segmentation effect, and at the same time, segmenting more accurate blood vessels can also help the stenosis detection model to detect stenosis more accurately, thereby realizing the combination of the two models. Therefore, the joint segmentation-detection model of the present invention has higher accuracy than conventional coronary artery stenosis detection methods. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0012] Figure 1 This is a diagram of an application scenario of vascular stenosis detection based on angiography image segmentation provided by an embodiment of the present invention; Figure 2 is a flow chart of an implementation method of a vascular stenosis detection method based on angiography image segmentation provided by an embodiment of the present invention; Figure 3 is a schematic structural diagram of a vascular stenosis detection device based on angiography image segmentation provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0013] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.
[0014] Figure 1 is an application scenario diagram of vascular stenosis detection based on angiography image segmentation provided by an embodiment of the present invention. Figure 1 As shown, in some embodiments, the vascular stenosis detection based on angiography image segmentation provided by the embodiment of the present invention can be applied to but not limited to this application scenario. The system may include: angiography equipment 11 and processing electronic equipment 12.
[0015] First, the catheter needs to be inserted through the patient's femoral artery or other peripheral artery and sent to the ascending aorta. Then, the left or right coronary artery orifice is found and inserted. Then, contrast agent is injected into the coronary artery. At this time, the angiography device 11 captures the developed angiography image and sends it to the electronic device 12. The electronic device 12 segments and processes the angiography image to obtain a clear cardiovascular segmentation map.
[0016] Among them, the electronic device 12 can be an electronic device or a server. The electronic device can be an inspection electronic device equipped on a medical X-ray inspection device, a doctor's office electronic device, etc. The server can be a management server of a hospital information management system of a hospital, or it can be a cloud server, which is not limited here.
[0017] Usually, when patients have symptoms of vascular discomfort, they need to undergo static and dynamic vascular examinations. Static examinations include CT, MR, etc., and dynamic examinations include DSA (Digital subtraction angiography) angiography technology. After the examination, the corresponding lesion score is performed to provide data reference for the doctor's diagnosis and treatment.
[0018] Figure 2 FIG. 1 is a flow chart of the implementation of the vascular stenosis detection method based on angiography image segmentation provided by an embodiment of the present invention. Figure 2 As shown, the vascular stenosis detection method based on angiography image segmentation includes: S210, acquiring angiography images; S220, inputting the angiography image into a joint segmentation-detection model to obtain a stenosis detection result of the blood vessel; Among them, the joint segmentation-detection model includes an image segmentation model and a stenosis detection model; the detection loss of the stenosis detection model is back-propagated to update the network parameters of the stenosis detection model; the segmentation loss of the image segmentation model is back-propagated; the back-propagated detection loss and segmentation loss are used to update the network parameters of the image segmentation model; the input of the image segmentation model is angiography image, and the output is vascular segmentation image; the input of the stenosis detection model is angiography image and vascular segmentation image, and the output is stenosis detection result.
[0019] In an embodiment of the present invention, the image segmentation model and the stenosis detection model may be supervised learning models, semi-supervised learning models, etc., which are not limited here. Both the image segmentation model and the stenosis detection model are provided with a loss function that can be forward propagated. The loss function may be mean square error loss, cross entropy loss, logarithmic loss, etc., which are not limited here. In order to enable the image segmentation model to utilize detection loss and segmentation loss at the same time, the present invention sets a multi-task learning framework in the image segmentation model. The model learns multiple tasks at the same time, and each task has its own loss function. These loss functions can be combined into a total loss function, and the model updates the weights of the model by minimizing this total loss.
[0020] In some embodiments, the total loss function may be calculated as follows: (1) in, Loss is the total loss function, Loss 1 is the segmentation loss, Loss 2 is the detection loss, ω 1 and ω 2 is the preset parameter.
[0021] In some embodiments, the total loss function may also be calculated by the following formula: (2) in, k is a consistency index, which is calculated based on the above segmentation loss and detection loss. Both segmentation loss and detection loss are the difference between the model output and the true value, that is, a pixel difference map of segmentation loss and a pixel difference map of detection loss can be obtained. These two pixel difference maps are overlapped. If a pixel in the image is both a detection loss and a segmentation loss, it means that the segmentation difference of this part has caused the detection error, which is recorded as a consistency pixel. The number of all consistent pixels divided by the number of pixels in the pixel difference map of segmentation loss is the consistency index. The more detection errors caused by segmentation differences and the more consistent pixels there are, the greater the consistency index.
[0022] In addition to the above-mentioned method of fusing the two loss functions, the model of the image segmentation model can also be changed with the detection loss in the following way. The output of the image segmentation model is an image with annotated confidence. Each pixel in the image is set with a confidence, which is usually binarized with a threshold of 0.5, that is, pixels greater than 0.5 are treated as blood vessels, and blood vessels less than 0.5 are treated as background.
[0023] In some embodiments, the image segmentation model is used to: segment the angiography image to obtain a vascular segmentation image; determine the segmentation probability within the target box based on the detection loss; and re-segment the corresponding part of the target box in the angiography image based on the segmentation probability to update the vascular segmentation image.
[0024] In an embodiment of the present invention, the detection loss corresponds to three convolutional layers in the image segmentation model. First, the image segmentation model segments the image in a conventional manner. At the same time, the stenosis detection model marks the target box where the stenosis is located and gives the detection loss. After these two parts are fed back to the image segmentation model, convolutional layer 1 separates the target box from the angiography image. Convolutional layer 2 adjusts the segmentation probability according to the detection loss, that is, adjusts the above-mentioned confidence level, and re-segments the image in the target box according to the new confidence level. Finally, convolutional layer 3 fuses the image in the target box with the original segmented image to obtain an updated vascular segmentation image.
[0025] If the detection loss corresponding to the target box is large, the confidence level can be reduced so that the image segmentation model can segment more image details from the image within the target box and make the narrow area clearer.
[0026] Stenosis is a change in blood vessel width, which is not very noticeable on angiography. However, after vascular segmentation, the width of each blood vessel can be easily obtained. Generally speaking, except for the bifurcation points of the blood vessels, the width of other blood vessels usually changes evenly. If the width of a blood vessel changes suddenly at a certain point, it indicates that this is a stenotic / lesion area.
[0027] A target frame can be used to mark the stenosis / lesion area, but a traditional target frame is usually a horizontal rectangle, and the direction of the blood vessel is not horizontal, which results in a large amount of background being framed in the target frame, which is not conducive to the subsequent identification of stenosis / lesions. Therefore, the present invention draws the target frame through the following embodiment.
[0028] In some embodiments, the stenosis detection model is used to: determine at least one target frame and the position, size and angle of each target frame based on the blood vessel segmentation image; and obtain the stenosis detection result by marking each target frame on the angiography image.
[0029] In some embodiments, determining at least one target frame and the position, size and angle of each target frame based on the blood vessel segmentation image includes: identifying blood vessel width data in the blood vessel segmentation image; determining at least one target frame and the position, size and angle of each target frame based on the blood vessel width data.
[0030] In the embodiment of the present invention, after a stenotic blood vessel is detected, the center of the stenotic blood vessel is the position of the target frame, the angle between the direction of the blood vessel and the horizontal direction is the angle of the target frame, and the length of the stenotic blood vessel and the width of the normal blood vessel where the stenosis is located are the size of the target frame. By designing the above-mentioned rotating frame, the target frame can be adaptively adjusted according to the shape and size of the blood vessel, and the background in the target frame can be removed as much as possible to ensure the identification of stenosis / lesions.
[0031] In some embodiments, the method further includes: acquiring real contrast images; forming a training set based on the real contrast images, and training the joint segmentation-detection model.
[0032] In an embodiment of the present invention, real angiography images labeled by experts can be obtained to form a training set for training the joint segmentation-detection model. However, the actual data volume of real angiography images is usually difficult to meet the training requirements. Insufficient data volume will cause the trained model to be unable to identify stenosis in blood vessels well. Therefore, the present invention trains the model through the following embodiment.
[0033] In some embodiments, the method further includes: drawing a pseudo-vessel segmentation result, wherein the pseudo-vessel segmentation result is preset with multiple pseudo-stenosis; determining a pseudo-angiographic image based on the pseudo-vessel segmentation result and a pre-trained generation model; forming a pseudo-training set based on the pseudo-angiographic image, the pseudo-vessel segmentation result and the pseudo-stenosis, and pre-training the joint segmentation-detection model; obtaining a real angiography image; forming a training set based on the real angiography image, and training the pre-trained joint segmentation-detection model.
[0034] In the embodiment of the present invention, since stenosis is mainly a change in blood vessel width, the present invention can draw a large number of pseudo-vessel segmentation results manually / computer-generated. In the pseudo-vessel segmentation results, the blood vessel width at the pseudo-stenosis is smaller than the normal blood vessel width. This is used as a training set to train the model, so that the model can focus on the change in blood vessel width.
[0035] In the embodiment of the present invention, since the pseudo-vessel segmentation result is not real data, it has certain deviations from the actual blood vessels. Therefore, after the above training, the model needs to be trained again based on the real angiography images to correct the deviation of the model.
[0036] In some embodiments, drawing the pseudo blood vessel segmentation result includes: drawing an initial pseudo blood vessel segmentation map in response to a user's drawing operation; forming a pseudo stenosis of corresponding size at the user's click position in the initial pseudo blood vessel segmentation map in response to a user's click operation; marking target boxes for each pseudo stenosis camouflage on the initial pseudo blood vessel segmentation map, and recording the position, size and angle of each target box to obtain the pseudo blood vessel segmentation result.
[0037] In addition to the above manual drawing method, pseudo blood vessel segmentation results can also be generated by automatic drawing through algorithms. In some embodiments, an image drawing model can be trained through an adversarial network. After the training is completed, the user only needs to set the number of pseudo blood vessel segmentation images required, and the model can be output.
[0038] The specific training is as follows: Step 1: Initialize the parameters of the generator and discriminator.
[0039] Step 2: During the training process, the generator and discriminator are trained alternately.
[0040] Train the discriminator: Take a batch of images from the real dataset and get a batch of generated images from the generator. Then, use these images to train the discriminator so that it can better distinguish between real images and generated images.
[0041] Training the generator: fix the parameters of the discriminator, use the random position, size, and angle of the pseudo-stenosis to form a random noise vector, and then use the random noise vector as input to generate images through the generator. These images will be sent to the discriminator. The goal of the generator is to make the discriminator judge these images as real images. Among them, the role of the discriminator is to judge whether the loading of pseudo-stenosis on the generated image is reasonable, that is, whether the pseudo-stenosis overlaps with the blood vessels in the generated image. If it overlaps, it is considered that the generated image is real, otherwise it is a false image.
[0042] Repeat the second step until the generator and the discriminator reach a balance, at which point the generator is able to generate very realistic images.
[0043] After the training is completed, the user only needs to set the number of pseudo-vessel segmentation images required, and then generate a batch of random pseudo-stenosis and random noise vectors, and then load them onto the generated image of the generator to obtain the drawn pseudo-vessel segmentation results. The above-mentioned real training set can be obtained from real angiography images or can be manually drawn by the user, which is not limited here.
[0044] In summary, the beneficial effects of the present invention are specifically as follows: The present invention adopts an image segmentation model and a stenosis detection model to realize the detection of vascular stenosis. At the same time, the detection loss of the stenosis detection model is transmitted to the image segmentation model to assist the segmentation, so that the image segmentation model can obtain a better segmentation effect. At the same time, more accurate blood vessel segmentation can also help the stenosis detection model to detect stenosis more accurately, thereby realizing the combination of the two models. Therefore, the combined segmentation-detection model of the present invention is more accurate than the conventional coronary artery stenosis detection method.
[0045] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.
[0046] Figure 3 FIG. 1 is a schematic diagram of the structure of a vascular stenosis detection device based on angiography image segmentation provided by an embodiment of the present invention. Figure 3 As shown, in some embodiments, a vascular stenosis detection device 3 based on angiography image segmentation includes: An acquisition module 310 is used to acquire angiography images; A detection module 320, for inputting the angiography image into a joint segmentation-detection model to obtain a stenosis detection result of a blood vessel; Among them, the joint segmentation-detection model includes an image segmentation model and a stenosis detection model; the detection loss of the stenosis detection model is back-propagated to update the network parameters of the stenosis detection model; the segmentation loss of the image segmentation model is back-propagated; the back-propagated detection loss and segmentation loss are used to update the network parameters of the image segmentation model; the input of the image segmentation model is angiography image, and the output is vascular segmentation image; the input of the stenosis detection model is angiography image and vascular segmentation image, and the output is stenosis detection result.
[0047] Optionally, the stenosis detection model is used to: determine at least one target frame and the position, size and angle of each target frame based on the blood vessel segmentation image; and obtain the stenosis detection result by marking each target frame on the angiography image.
[0048] Optionally, the stenosis detection model is used to: identify blood vessel width data in the blood vessel segmentation image; and determine at least one target frame and the position, size and angle of each target frame based on the blood vessel width data.
[0049] Optionally, the image segmentation model is used to: segment the angiography image to obtain a vascular segmentation image; determine the segmentation probability within the target box based on the detection loss; and re-segment the corresponding part of the target box in the angiography image based on the segmentation probability to update the vascular segmentation image.
[0050] Optionally, the device further includes: a training module for acquiring real angiography images; forming a training set based on the real angiography images, and training the joint segmentation-detection model.
[0051] Optionally, a training module is used to: draw a pseudo-vessel segmentation result, wherein the pseudo-vessel segmentation result is preset with multiple pseudo-stenosis; determine a pseudo-angiography image based on the pseudo-vessel segmentation result and a pre-trained generation model; compose a pseudo-training set based on the pseudo-angiography image, the pseudo-vessel segmentation result and the pseudo-stenosis, and pre-train the joint segmentation-detection model; obtain a real angiography image; compose a training set based on the real angiography image, and train the pre-trained joint segmentation-detection model.
[0052] Optionally, a training module is used to: draw an initial pseudo vessel segmentation map in response to a user's drawing operation; form a pseudo stenosis of corresponding size at the user's click position in the initial pseudo vessel segmentation map in response to a user's click operation; mark target boxes for each pseudo stenosis camouflage on the initial pseudo vessel segmentation map, and record the position, size and angle of each target box to obtain a pseudo vessel segmentation result.
[0053] The vascular stenosis detection device based on angiography image segmentation provided in this embodiment can be used to execute the above method embodiments. Its implementation principle and technical effects are similar and will not be described in detail in this embodiment.
[0054] Figure 4 Schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Figure 4 As shown, an electronic device 4 provided by one embodiment of the present invention includes: a processor 40, a memory 41, and a computer program 42 stored in the memory 41 and executable on the processor 40. When the processor 40 executes the computer program 42, the steps in the above-mentioned embodiments of the vascular stenosis detection method based on angiography image segmentation are implemented, for example Figure 2 Alternatively, when the processor 40 executes the computer program 42, the functions of each module / unit in the above-mentioned system embodiments are realized, for example Figure 3 The functions of each module are shown.
[0055] Exemplarily, the computer program 42 may be divided into one or more modules / units, one or more modules / units are stored in the memory 41, and are executed by the processor 40 to implement the present invention. One or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program 42 in the electronic device 4.
[0056] The electronic device 4 may be an electronic device or a server. The electronic device may be a mobile phone, an MCU, an ECU, an industrial computer, etc., which are not limited here. The server may be a physical server, a cloud server, etc., which are not limited here. The electronic device 4 may include, but is not limited to, a processor 40 and a memory 41. Those skilled in the art will understand that Figure 4 It is only an example of the electronic device 4 and does not constitute a limitation of the electronic device 4. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.
[0057] The processor 40 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.
[0058] The memory 41 may be an internal storage unit of the electronic device 4, such as a hard disk or memory of the electronic device 4. The memory 41 may also be an external storage device of the electronic device 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 4. Further, the memory 41 may also include both an internal storage unit and an external storage device of the electronic device 4. The memory 41 is used to store computer programs and other programs and data required by the electronic device. The memory 41 may also be used to temporarily store data that has been output or is to be output.
[0059] An embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the steps in the embodiment of the vascular stenosis detection method based on angiography image segmentation are implemented.
[0060] The computer readable storage medium stores a computer program 42, which includes program instructions. When the program instructions are executed by the processor 40, all or part of the processes in the above-mentioned embodiment method are implemented. The computer program 42 can also be used to instruct related hardware to complete. The computer program 42 can be stored in a computer readable storage medium. When the computer program 42 is executed by the processor 40, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program 42 includes computer program code, which can be in source code form, object code form, executable file or some intermediate form. The computer readable medium can include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium.
[0061] The computer-readable storage medium may be an internal storage unit of the electronic device of any of the aforementioned embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.
[0062] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.
[0063] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.
[0064] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0065] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.
[0066] In the embodiments provided by the present invention, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0067] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0068] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0069] If the integrated module / 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 present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. Computer-readable media can include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.
[0070] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such 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 invention, and should be included in the protection scope of the present invention.
Claims
1. A method for detecting vascular stenosis based on angiography image segmentation, characterized in that: include: Obtaining angiographic images; Inputting the angiography image into a joint segmentation-detection model to obtain a stenosis detection result of the blood vessel; The joint segmentation-detection model includes an image segmentation model and a stenosis detection model; the detection loss of the stenosis detection model is back-propagated to update the network parameters of the stenosis detection model; the segmentation loss of the image segmentation model is back-propagated; the back-propagated detection loss and segmentation loss are used to update the network parameters of the image segmentation model; The input of the image segmentation model is the angiography image, and the output is the blood vessel segmentation image; the input of the stenosis detection model is the angiography image and the blood vessel segmentation image, and the output is the stenosis detection result.
2. The method for detecting vascular stenosis based on contrast image segmentation according to claim 1, characterized in that: The stenosis detection model is used to: Determine at least one target frame and the position, size and angle of each target frame according to the blood vessel segmentation image; The stenosis detection result is obtained by marking each target frame on the angiography image.
3. The method for detecting vascular stenosis based on contrast image segmentation according to claim 2, characterized in that: Determining at least one target frame and the position, size and angle of each target frame according to the blood vessel segmentation image, including: identifying blood vessel width data in the blood vessel segmentation image; At least one target frame and the position, size and angle of each target frame are determined according to the blood vessel width data.
4. The method for detecting vascular stenosis based on contrast image segmentation according to claim 3, characterized in that: The image segmentation model is used to: Segmenting the angiography image to obtain a blood vessel segmentation image; Determining a segmentation probability within a target box according to the detection loss; According to the segmentation probability, the corresponding portion of the target frame in the angiography image is re-segmented to update the blood vessel segmentation image.
5. The method for detecting vascular stenosis based on contrast image segmentation according to claim 1, characterized in that: The method further comprises: Acquire real angiographic images; The joint segmentation-detection model is trained by forming a training set based on the real angiography images.
6. The method for detecting vascular stenosis based on contrast image segmentation according to claim 1, characterized in that: The method further comprises: Drawing a pseudo-vessel segmentation result, wherein the pseudo-vessel segmentation result is preset with multiple pseudo-stenosis; Determining a pseudo-angiographic image according to the pseudo-vessel segmentation result and a pre-trained generative AI model; According to the pseudo angiographic image, the pseudo blood vessel segmentation result and the pseudo stenosis, a pseudo training set is formed to pre-train the joint segmentation-detection model; Acquire real angiographic images; The real angiography images are used to form a training set, and the pre-trained joint segmentation-detection model is trained.
7. The method for detecting vascular stenosis based on contrast image segmentation according to claim 5, characterized in that: The step of drawing the pseudo blood vessel segmentation result includes: The pseudo blood vessel segmentation result is drawn according to a pre-trained image drawing model; wherein the image drawing model is obtained according to adversarial network training.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the vascular stenosis detection method based on contrast image segmentation as described in any one of claims 1 to 7 are implemented.
9. A contrast image processing system, characterized in that: The device comprises an angiography device and the electronic device as claimed in claim 8 above.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the vascular stenosis detection method based on angiography image segmentation as described in any one of claims 1 to 7 are implemented.