Coronary artery segmentation method for contrast images, electronic device, processing system, and storage medium

Abstract

The application provides a contrast image coronary artery segmentation method, an electronic device, a processing system and a storage medium. The method comprises the following steps: grouping multiple cardiovascular angiography images of a preset cardiac cycle to obtain multiple cardiovascular angiography image segments; calculating a first key-value structure, a third key-value structure of each cardiovascular angiography image segment and a second key-value structure of a corresponding memory angiography image segment; extracting a target coronary artery feature value of each cardiovascular angiography image segment according to each first key-value structure, a corresponding second key-value structure and a given coronary artery feature value of the corresponding memory angiography image segment; performing coronary artery feature refinement extraction according to each third key-value structure and a corresponding target coronary artery feature value to obtain a refined coronary artery feature value of each cardiovascular angiography image segment; and decoding each refined coronary artery feature value to obtain a coronary artery segmentation result of each cardiovascular angiography image segment. The application can improve the accuracy of cardiovascular angiography image coronary artery segmentation.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, and particularly relates to a method for coronary artery segmentation in angiography images, an electronic device, a processing system, and a storage medium. Background Technology

[0002] Contrast imaging is a widely used type of medical imaging. Specifically, cardiovascular contrast imaging involves inserting a catheter through the femoral artery or other peripheral arteries in the thigh, advancing to the ascending aorta, and then inserting it into the left or right coronary artery ostium. Contrast agent is then injected, allowing the coronary arteries to be visualized under X-rays, thus obtaining the contrast image. Because other tissues in the body, such as bones, are also imaged by X-rays, contrast images often contain significant noise and are prone to blurring; therefore, image processing is necessary.

[0003] Image segmentation technology is fundamental to angiography image processing. Current techniques typically employ methods such as thresholding, region growing, statistical region fusion, and matched filtering to segment angiography images. However, due to the complexity of cardiovascular angiography images, traditional methods often result in coronary artery segmentation with significant noise or missing pixels, leading to low accuracy in coronary artery segmentation. Summary of the Invention

[0004] In view of this, the present invention provides a method for coronary artery segmentation of angiography images, an electronic device, a processing system, and a storage medium, aiming to solve the problem of low accuracy in angiography image segmentation in the prior art.

[0005] A first aspect of this invention provides a method for coronary artery segmentation in angiography images, comprising:

[0006] Acquire multiple frames of cardiovascular angiography images for a preset cardiac cycle, and group the multiple frames of cardiovascular angiography images to obtain multiple cardiovascular angiography image fragments;

[0007] The first key value structure of each cardiovascular angiography image fragment and the second key value structure of the corresponding memory angiography image fragment are calculated through the first channel of the key encoder, and the third key value structure of each cardiovascular angiography image fragment is calculated through the second channel of the key encoder.

[0008] Calculate the first correlation matrix between each first key-value structure and the corresponding second key-value structure, and extract the target coronary artery feature value for each cardiovascular angiography image segment based on the first correlation matrix and the given coronary artery feature value of the corresponding memory angiography image segment.

[0009] Based on each of the third key value structures and the corresponding target coronary artery feature values, the coronary artery features of each cardiovascular angiography image segment are refined and extracted to obtain the refined coronary artery feature values ​​of each cardiovascular angiography image segment.

[0010] Each refined coronary artery feature value is decoded to obtain the coronary artery segmentation result for each cardiovascular angiography image segment.

[0011] A second aspect of the present invention provides a coronary artery segmentation apparatus for angiography images, comprising:

[0012] The slicing module 31 is used to acquire multiple frames of cardiovascular angiography images of a preset cardiac cycle, and to group the multiple frames of cardiovascular angiography images to obtain multiple cardiovascular angiography image segments.

[0013] The encoding module 32 is used to calculate the first key value structure of each cardiovascular angiography image segment and the second key value structure of the corresponding memory angiography image segment through the first channel of the key encoder, and to calculate the third key value structure of each cardiovascular angiography image segment through the second channel of the key encoder.

[0014] Matching module 33 is used to calculate a first correlation matrix between each first key-value structure and the corresponding second key-value structure, and extract the target coronary artery feature value of each cardiovascular angiography image segment based on the first correlation matrix and the given coronary artery feature value of the corresponding memory angiography image segment.

[0015] The refinement module 34 is used to refine and extract the coronary artery features of each cardiovascular angiography image segment according to each of the third key value structures and the corresponding target coronary artery feature values, so as to obtain the refined coronary artery feature values ​​of each cardiovascular angiography image segment.

[0016] The decoding module 35 is used to decode each of the refined coronary artery feature values ​​to obtain the coronary artery segmentation result of each cardiovascular angiography image segment.

[0017] A third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the angiography image coronary artery segmentation method of the first aspect above.

[0018] A fourth aspect of the present invention provides an imaging image processing system, including a medical X-ray examination device and the electronic device described in the third aspect above.

[0019] A fifth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the angiography image coronary artery segmentation method of the first aspect above.

[0020] This invention provides a method, electronic device, processing system, and storage medium for coronary artery segmentation of angiography images. The method involves first acquiring multiple frames of cardiovascular angiography images within a preset cardiac cycle, and grouping these frames to obtain multiple cardiovascular angiography image segments. Then, a first key-value structure and a second key-value structure corresponding to the memorized angiography image segment are calculated using the first channel of a key encoder. A third key-value structure is then calculated for each cardiovascular angiography image segment using the second channel of the key encoder. Next, a first correlation matrix is ​​calculated between each first key-value structure and its corresponding second key-value structure. Based on the first correlation matrix and the given coronary artery feature values ​​of the corresponding memorized angiography image segment, target coronary artery feature values ​​are extracted for each cardiovascular angiography image segment. Furthermore, the coronary artery features of each cardiovascular angiography image segment are refined based on each third key-value structure and its corresponding target coronary artery feature values, resulting in refined coronary artery feature values ​​for each cardiovascular angiography image segment. Finally, each refined coronary artery feature value is decoded to obtain the coronary artery segmentation result for each cardiovascular angiography image segment. By considering the semantic structural relationships between different segments of the coronary artery in a cardiovascular angiography image through multiple segments, coronary artery segmentation can be performed according to the segments, thereby reducing noise or pixel loss during the segmentation process and improving the accuracy of coronary artery segmentation in cardiovascular angiography images. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is an application scenario diagram of the coronary artery segmentation method for angiography images provided in the embodiments of the present invention;

[0023] Figure 2 This is a flowchart illustrating the implementation of the coronary artery segmentation method for angiography images provided in this embodiment of the invention.

[0024] Figure 3 This is a schematic diagram of the coronary artery segmentation device for angiography provided in an embodiment of the present invention;

[0025] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0026] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0027] Figure 1 This is an application scenario diagram of the coronary artery segmentation method for angiography images provided in an embodiment of the present invention. For example... Figure 1 As shown, in some embodiments, the coronary artery segmentation method for angiography provided by the present invention can be applied to, but is not limited to, this application scenario. The system may include: a medical X-ray examination device 11 and an electronic device 12.

[0028] First, a catheter needs to be inserted through the patient's femoral artery or other peripheral arteries and advanced to the ascending aorta. Then, the left or right coronary artery orifice is located and inserted. Contrast agent is then injected into the coronary artery. At this time, the medical X-ray examination device 11 captures the contrast-enhanced cardiovascular angiography image and sends it to the electronic device 12. The electronic device 12 segments the cardiovascular angiography image to obtain a clear cardiovascular segmentation map.

[0029] Among them, electronic device 12 can be a terminal or a server. The terminal can be an examination terminal equipped on medical X-ray examination equipment 11, a doctor's office terminal, etc., and the server can be the management server of the hospital's hospital information management system or a cloud server, which is not limited here.

[0030] Figure 2 This is a flowchart illustrating the implementation of the coronary artery segmentation method for angiography images provided in this embodiment of the invention. Figure 2 As shown, the coronary artery segmentation method in angiography images is applied to... Figure 1 The method of the electronic device 12 shown may include:

[0031] In step 201, multiple cardiovascular angiography images of a preset cardiac cycle are acquired, and the multiple cardiovascular angiography images are grouped to obtain multiple cardiovascular angiography image segments.

[0032] In this embodiment of the invention, the multi-frame cardiovascular angiography images with a preset cardiac cycle can be: Figure 1 The images can be taken in real time by the medical X-ray examination equipment 11, or they can be obtained from the hospital's information management system; there is no limitation on this.

[0033] For example, the preset cardiac cycle can be 1 to 2 cardiac cycles.

[0034] Optionally, grouping multiple cardiovascular angiography images to obtain multiple cardiovascular angiography image segments may include: grouping multiple cardiovascular angiography images according to a preset frame interval to obtain multiple cardiovascular angiography image segments.

[0035] For example, assuming 60 frames of cardiovascular angiography images are obtained within 2 cardiac cycles, with a preset frame interval of 5 frames, the 1st, 6th, 11th, etc. of the cardiovascular angiography images can be divided into a cardiovascular angiography image segment, the 2nd, 7th, 12th, etc. can be divided into a cardiovascular angiography image segment, and so on.

[0036] In this embodiment, by grouping multiple frames of cardiovascular angiography images according to a preset frame interval, frames of cardiovascular angiography images that may have semantic structural relationships can be grouped together for coronary artery segmentation, thereby reducing noise or pixel loss during the coronary artery segmentation process and improving the accuracy of coronary artery segmentation in cardiovascular angiography images.

[0037] In step 202, the first key value structure of each cardiovascular angiography image segment and the second key value structure of the corresponding memory angiography image segment are calculated through the first channel of the key encoder, and the third key value structure of each cardiovascular angiography image segment is calculated through the second channel of the key encoder.

[0038] In this embodiment, the memory angiography image fragment can be understood as multiple frames of angiography images with the coronary artery location already marked. In order to facilitate the subsequent matching of the acquired cardiovascular angiography image fragments with the memory angiography image fragments, the same key encoder can be used to calculate the first key value structure of each cardiovascular angiography image fragment and the second key value structure of the corresponding memory angiography image fragment.

[0039] To facilitate the subsequent refinement of coronary artery features of each cardiovascular angiography image segment by utilizing the spatiotemporal relationship of each frame of cardiovascular angiography image segment, the third key value structure of each cardiovascular angiography image segment can be calculated through the second channel of the key encoder.

[0040] In step 203, a first correlation matrix between each first key-value structure and the corresponding second key-value structure is calculated. Based on the first correlation matrix and the given coronary artery feature value of the corresponding memory angiography image fragment, the target coronary artery feature value of each cardiovascular angiography image fragment is extracted.

[0041] In this embodiment, by calculating the first correlation matrix between each first key value structure and the corresponding second key value structure, each cardiovascular angiography image fragment can be matched with the corresponding memory angiography image fragment, thereby using the given coronary artery feature value at the matching position in the corresponding memory angiography image fragment as the target coronary artery feature value of the cardiovascular angiography image fragment.

[0042] Optionally, calculating the first correlation matrix between each first key-value structure and the corresponding second key-value structure may include:

[0043] according to Calculate the first correlation matrix between each of the first key-value structures and the corresponding second key-value structures.

[0044] Where A(kQ,kM) i,j Let k be the correlation between position i in the first key-value structure and position j in the corresponding second key-value structure. i Q is the key value at position i in each of the first key-value structures, k j M is the bond value at position j in the corresponding second bond value structure, and τ is the temperature parameter.

[0045] By using the temperature parameter τ, the rapid increase of the exponential function can be avoided.

[0046] In this embodiment, by taking the inner product of each first key value structure and the corresponding second key value structure to obtain the corresponding first correlation matrix, the angle information between each cardiovascular angiography image fragment and the corresponding memory angiography image fragment in the mapping space can be considered, so that the obtained first correlation matrix can better represent the matching degree between each cardiovascular angiography image fragment and the corresponding memory angiography image fragment.

[0047] Optionally, based on the first correlation matrix and the given coronary artery feature values ​​of the corresponding memory angiography image fragment, extracting the target coronary artery feature values ​​for each cardiovascular angiography image fragment may include:

[0048] According to v Q =Read(k Q ,k M ,v M )=A(k Q ,k M )v M Extract the target coronary artery feature values ​​for each cardiovascular angiography image fragment.

[0049] Among them, v Q For each cardiovascular angiography image fragment, the target coronary artery feature value, Read(k) Q ,k M ,v M ) is based on each of the first key-value structures k Q The corresponding second key-value structure k M and the corresponding given coronary artery feature value v M Feature extraction, A(k) Q ,k M ) represents the first correlation matrix, v M The given coronary artery characteristic value.

[0050] In step 204, the coronary artery features of each cardiovascular angiography image segment are refined and extracted based on each third key value structure and the corresponding target coronary artery feature value, thereby obtaining the refined coronary artery feature value of each cardiovascular angiography image segment.

[0051] In this embodiment, even though the most relevant target coronary artery feature values ​​for each cardiovascular angiography image segment are extracted in step 203 by memorizing the angiography image segments, coronary artery segmentation based solely on the target coronary artery feature values ​​of each cardiovascular angiography image segment is prone to errors when there are new targets, occlusions, or large deformations of objects. Therefore, considering the spatiotemporal relationship between each frame of cardiovascular angiography images in a cardiovascular angiography image segment, the coronary artery features of each cardiovascular angiography image segment are further refined and extracted to obtain more accurate refined coronary artery feature values, thereby enhancing the accuracy of coronary artery segmentation.

[0052] Optionally, the coronary artery features of each cardiovascular angiography image segment are refined based on each third key value structure and the corresponding target coronary artery feature value to obtain refined coronary artery feature values ​​for each cardiovascular angiography image segment. This may include:

[0053] The second correlation matrix is ​​calculated based on each third key value structure, and the attention-weighted coronary artery feature value of each cardiovascular angiography image segment is obtained based on the second correlation matrix and the corresponding target coronary artery feature value.

[0054] Based on each attention-weighted coronary artery feature value and the attention-weighted coronary artery feature value processed by the feedforward network, refined coronary artery feature values ​​for each cardiovascular angiography image segment are obtained.

[0055] In this embodiment, by calculating the second correlation matrix based on each third key value structure, and obtaining the attention-weighted coronary feature value of each cardiovascular angiography image segment based on the second correlation matrix and the corresponding target coronary feature value, the similarity transformation of the target coronary feature value can be performed based on the spatiotemporal relationship between each frame of cardiovascular angiography images in the cardiovascular angiography image segment, thereby enhancing the coronary features with high similarity in the target coronary feature value, which is beneficial to obtaining more accurate refined coronary feature values.

[0056] Optionally, calculating a second correlation matrix based on each third key-value structure, and obtaining attention-weighted coronary artery feature values ​​for each cardiovascular angiography image segment based on the second correlation matrix and the corresponding target coronary artery feature values, may include:

[0057] according to Obtain attention-weighted coronary artery feature values ​​for each cardiovascular angiography image fragment.

[0058] Wherein, vattn is the attention-weighted coronary artery feature value for each cardiovascular angiography image segment. This is the second correlation matrix. For each of the third key-value structures The result of the pooling process is ψ(vQ), which is the result of convolution processing on the target coronary artery feature value vQ, where vQ is the target coronary artery feature value.

[0059] In this embodiment, by performing a pooling operation on each third key value structure before calculating the second correlation matrix, the computational load can be reduced. Moreover, by performing a similarity transformation on the target coronary artery feature values ​​using the second correlation matrix calculated after the pooling operation, more important coronary artery features can be preserved while reducing the computational load.

[0060] Optionally, based on each attention-weighted coronary artery feature value and the attention-weighted coronary artery feature value processed by the feedforward network, refined coronary artery feature values ​​for each cardiovascular angiography image segment can be obtained, which may include:

[0061] according to Obtain refined coronary artery feature values ​​for each cardiovascular angiography image fragment.

[0062] in, FFN(vattn) represents the refined coronary artery feature values ​​for each cardiovascular angiography image segment, while FFN(vattn) represents the attention-weighted coronary artery feature values ​​after processing by the feedforward network.

[0063] In step 205, each refined coronary artery feature value is decoded to obtain the coronary artery segmentation result for each cardiovascular angiography image segment.

[0064] In this embodiment, each refined coronary artery feature value is decoded to obtain the coronary artery image corresponding to each cardiovascular angiography image segment as the coronary artery segmentation result of that cardiovascular angiography image segment.

[0065] In this embodiment of the invention, multiple cardiovascular angiography images of a preset cardiac cycle are first acquired and grouped to obtain multiple cardiovascular angiography image segments. Then, the first key-value structure of each cardiovascular angiography image segment and the second key-value structure of the corresponding memory angiography image segment are calculated through the first channel of the key encoder, and the third key-value structure of each cardiovascular angiography image segment is calculated through the second channel of the key encoder. Next, the first correlation matrix between each first key-value structure and the corresponding second key-value structure is calculated. Based on the first correlation matrix and the given coronary artery feature value of the corresponding memory angiography image segment, the target coronary artery feature value of each cardiovascular angiography image segment is extracted. Then, based on each third key-value structure and the corresponding target coronary artery feature value, the coronary artery features of each cardiovascular angiography image segment are refined to obtain the refined coronary artery feature value of each cardiovascular angiography image segment. Finally, each refined coronary artery feature value is decoded to obtain the coronary artery segmentation result of each cardiovascular angiography image segment. By considering the semantic structural relationships between different segments of the coronary artery in a cardiovascular angiography image through multiple segments, coronary artery segmentation can be performed according to the segments, thereby reducing noise or pixel loss during the segmentation process and improving the accuracy of coronary artery segmentation in cardiovascular angiography images.

[0066] It should be understood that the sequence number of each step in the above embodiments does not imply 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 embodiments of the present invention.

[0067] Figure 3 This is a schematic diagram of the coronary artery segmentation device for angiography provided in an embodiment of the present invention. Figure 3 As shown, in some embodiments, the coronary artery segmentation device for angiography includes:

[0068] The slicing module 31 is used to acquire multiple frames of cardiovascular angiography images of a preset cardiac cycle, and to group the multiple frames of cardiovascular angiography images to obtain multiple cardiovascular angiography image segments.

[0069] The encoding module 32 is used to calculate the first key value structure of each cardiovascular angiography image segment and the second key value structure of the corresponding memory angiography image segment through the first channel of the key encoder, and to calculate the third key value structure of each cardiovascular angiography image segment through the second channel of the key encoder.

[0070] Matching module 33 is used to calculate a first correlation matrix between each first key-value structure and the corresponding second key-value structure, and extract the target coronary artery feature value of each cardiovascular angiography image segment based on the first correlation matrix and the given coronary artery feature value of the corresponding memory angiography image segment.

[0071] The refinement module 34 is used to refine and extract the coronary artery features of each cardiovascular angiography image segment according to each of the third key value structures and the corresponding target coronary artery feature values, so as to obtain the refined coronary artery feature values ​​of each cardiovascular angiography image segment.

[0072] The decoding module 35 is used to decode each of the refined coronary artery feature values ​​to obtain the coronary artery segmentation result of each cardiovascular angiography image segment.

[0073] This embodiment first acquires multiple frames of cardiovascular angiography images within a preset cardiac cycle, and groups these images to obtain multiple cardiovascular angiography image segments. Then, it calculates the first key-value structure and the second key-value structure of each cardiovascular angiography image segment using the first channel of a key encoder, and calculates the third key-value structure of each segment using the second channel of the key encoder. Next, it calculates the first correlation matrix between each first key-value structure and the corresponding second key-value structure. Based on the first correlation matrix and the given coronary artery feature values ​​of the corresponding memory angiography image segment, it extracts the target coronary artery feature values ​​for each segment. Then, it refines the coronary artery features of each segment based on each third key-value structure and the corresponding target coronary artery feature values, obtaining refined coronary artery feature values ​​for each segment. Finally, it decodes each refined coronary artery feature value to obtain the coronary artery segmentation result for each segment. By considering the semantic structural relationships between different segments of the coronary artery in a cardiovascular angiography image through multiple segments, coronary artery segmentation can be performed according to the segments, thereby reducing noise or pixel loss during the segmentation process and improving the accuracy of coronary artery segmentation in cardiovascular angiography images.

[0074] Optionally, the slicing module 31 is used to group multiple cardiovascular angiography images according to a preset frame interval to obtain multiple cardiovascular angiography image segments.

[0075] Optional, matching module 33, used for matching according to Calculate the first correlation matrix between each of the first key-value structures and the corresponding second key-value structures;

[0076] Where A(kQ,kM)i,j is the correlation between position i in the first bond value structure and position j in the corresponding second bond value structure, kiQ is the bond value at position i in the first bond value structure, kjM is the bond value at position j in the corresponding second bond value structure, and τ is the temperature parameter.

[0077] Optionally, the matching module 33 is used to extract the target coronary artery feature value of each cardiovascular angiography image segment according to vQ = Read(kQ,kM,vM) = A(kQ,kM)vM;

[0078] Wherein, vQ is the target coronary artery feature value for each cardiovascular angiography image segment, Read(kQ,kM,vM) is the feature extraction based on each first key-value structure kQ, the corresponding second key-value structure kM and the corresponding given coronary artery feature value vM, A(kQ,kM) is the first correlation matrix, and vM is the given coronary artery feature value.

[0079] Optionally, the refinement module 34 is used to calculate a second correlation matrix based on each of the third key value structures, and to obtain the attention-weighted coronary artery feature value of each cardiovascular angiography image segment based on the second correlation matrix and the corresponding target coronary artery feature value.

[0080] Based on each attention-weighted coronary artery feature value and the attention-weighted coronary artery feature value processed by the feedforward network, refined coronary artery feature values ​​for each cardiovascular angiography image segment are obtained.

[0081] Optional, refinement module 34, used to... Obtain attention-weighted coronary artery feature values ​​for each cardiovascular angiography image segment;

[0082] Wherein, vattn is the attention-weighted coronary artery feature value for each cardiovascular angiography image segment. This is the second correlation matrix. For each of the third key-value structures The result of the pooling process is ψ(vQ), which is the result of convolution processing on the target coronary artery feature value vQ, where vQ is the target coronary artery feature value.

[0083] Optional, refinement module 34, used to... Obtain refined coronary artery feature values ​​for each cardiovascular angiography image segment;

[0084] in, FFN(vattn) represents the refined coronary artery feature values ​​for each cardiovascular angiography image segment, while FFN(vattn) represents the attention-weighted coronary artery feature values ​​after processing by the feedforward network.

[0085] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. For example... Figure 4 As shown, an embodiment of the present invention provides an electronic device 4, which 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, it implements the steps in the various angiography image coronary artery segmentation method embodiments described above, for example... Figure 2 Steps 201 to 205 are shown. Alternatively, when processor 40 executes computer program 42, it implements the functions of each module / unit in the above system embodiments, for example... Figure 3 The functions of modules 31 to 35 are shown.

[0086] For example, computer program 42 may be divided into one or more modules / units, one or more of which are stored in memory 41 and executed by processor 40 to complete the present invention. One or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 42 in electronic device 4.

[0087] Electronic device 4 can be a terminal or a server. The terminal can be a mobile phone, MCU, ECU, etc., and the server can be a physical server or a cloud server; no limitation is made here. Electronic device 4 may include, but is not limited to, processor 40 and memory 41. Those skilled in the art will understand that... Figure 4 This is merely an example of electronic device 4 and does not constitute a limitation on electronic device 4. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic devices may also include input / output devices, network access devices, buses, etc.

[0088] The processor 40 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), 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.

[0089] The memory 41 can be an internal storage unit of the electronic device 4, such as a hard disk or RAM. The memory 41 can also be an external storage device of the electronic device 4, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory 41 can include both internal and external storage units 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 can also be used to temporarily store data that has been output or will be output.

[0090] This invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the embodiments of the coronary artery segmentation method for angiography images.

[0091] A computer-readable storage medium stores a computer program 42. The computer program 42 includes program instructions. When executed by the processor 40, the program instructions implement all or part of the processes in the methods described in the above embodiments. The computer program 42 can also instruct related hardware to complete the process. The computer program 42 can be stored in a computer-readable storage medium. When executed by the processor 40, the computer program 42 can implement the steps of the various method embodiments described above. The computer program 42 includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0092] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic 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 can also be used to temporarily store data that has been output or will be output.

[0093] It should be understood that the sequence number of each step in the above embodiments does not imply 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 embodiments of the present invention.

[0094] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to 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 embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0095] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0096] Those skilled in the art will recognize that the units and algorithm steps of the various examples 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 implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0097] In the embodiments provided by this 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 merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0098] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0100] If an integrated module / unit is implemented as 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, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0101] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for coronary artery segmentation in angiographic images, characterized in that, include: Acquire multiple frames of cardiovascular angiography images for a preset cardiac cycle, and group the multiple frames of cardiovascular angiography images to obtain multiple cardiovascular angiography image fragments; The first key value structure of each cardiovascular angiography image fragment and the second key value structure of the corresponding memory angiography image fragment are calculated through the first channel of the key encoder, and the third key value structure of each cardiovascular angiography image fragment is calculated through the second channel of the key encoder. Calculate the first correlation matrix between each first key-value structure and the corresponding second key-value structure, and extract the target coronary artery feature value for each cardiovascular angiography image segment based on the first correlation matrix and the given coronary artery feature value of the corresponding memory angiography image segment. Based on each of the third key value structures and the corresponding target coronary artery feature values, the coronary artery features of each cardiovascular angiography image segment are refined and extracted to obtain the refined coronary artery feature values ​​of each cardiovascular angiography image segment. Each refined coronary artery feature value is decoded to obtain the coronary artery segmentation result for each cardiovascular angiography image segment.

2. The coronary artery segmentation method for angiography images according to claim 1, characterized in that, Multiple cardiovascular angiography images were grouped to obtain multiple cardiovascular angiography image segments, including: Multiple cardiovascular angiography images are grouped according to a preset frame interval to obtain multiple cardiovascular angiography image segments.

3. The coronary artery segmentation method for angiography images according to claim 1, characterized in that, The calculation of the first correlation matrix between each of the first key-value structures and the corresponding second key-value structures includes: according to Calculate the first correlation matrix between each of the first key-value structures and the corresponding second key-value structures; Where A(k) Q ,k M ) i,j Let k be the correlation between position i in the first key-value structure and position j in the corresponding second key-value structure. i Q For each key value at position i in the first key-value structure, k j M This corresponds to the bond value at position j in the second bond value structure, where τ is the temperature parameter. l Traverse all positions of the second key-value structure using the index.

4. The coronary artery segmentation method for angiography images according to claim 3, characterized in that, The step of extracting the target coronary artery feature value for each cardiovascular angiography image segment based on the first correlation matrix and the given coronary artery feature value of the corresponding memory angiography image segment includes: According to v Q =Read(k Q ,k M ,v M )=A(k Q ,k M )v M Extract the target coronary artery feature values ​​for each cardiovascular angiography image fragment; Among them, v Q For each cardiovascular angiography image fragment, the target coronary artery feature value, Read(k) Q ,k M ,v M ) is based on each of the first key-value structures k Q The corresponding second key-value structure k M and the corresponding given coronary artery feature value v M Feature extraction, A(k) Q ,k M ) represents the first correlation matrix, v M The given coronary artery characteristic value.

5. The coronary artery segmentation method for angiography images according to claim 1, characterized in that, The step of refining and extracting coronary artery features from each cardiovascular angiography image segment based on each of the third key value structures and the corresponding target coronary artery feature values ​​to obtain refined coronary artery feature values ​​for each cardiovascular angiography image segment includes: A second correlation matrix is ​​calculated based on each of the third key-value structures, and attention-weighted coronary artery feature values ​​are obtained for each cardiovascular angiography image segment based on the second correlation matrix and the corresponding target coronary artery feature values. Based on each attention-weighted coronary artery feature value and the attention-weighted coronary artery feature value processed by the feedforward network, refined coronary artery feature values ​​for each cardiovascular angiography image segment are obtained.

6. The coronary artery segmentation method for angiography images according to claim 5, characterized in that, The step of calculating a second correlation matrix based on each of the third key-value structures, and obtaining attention-weighted coronary artery feature values ​​for each cardiovascular angiography image segment based on the second correlation matrix and the corresponding target coronary artery feature values, includes: according to Obtain attention-weighted coronary artery feature values ​​for each cardiovascular angiography image segment; Among them, v attn Attention-weighted coronary artery feature values ​​for each cardiovascular angiography image segment, This is the second correlation matrix. For each of the third key-value structures The result of pooling is ψ(v) Q ) represents the characteristic value v of the target coronary artery. Q The result of convolution processing, v Q The target coronary artery feature value.

7. The coronary artery segmentation method for angiography images according to claim 6, characterized in that, The process of obtaining refined coronary artery feature values ​​for each cardiovascular angiography image segment based on each attention-weighted coronary artery feature value and the attention-weighted coronary artery feature values ​​processed by the feedforward network includes: according to Obtain refined coronary artery feature values ​​for each cardiovascular angiography image segment; in, Refined coronary artery feature values ​​for each cardiovascular angiography image fragment, FFN(v attn ) represents the attention-weighted coronary artery feature value after processing by the feedforward network.

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, it implements the steps of the angiography image coronary artery segmentation method as described in any one of claims 1 to 7.

9. An imaging processing system comprising a medical X-ray examination device and the electronic device as described in claim 8 above.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the angiography image coronary artery segmentation method as described in any one of claims 1 to 7.

Images