Myocardial image detection device, system and storage medium

Abstract

The present invention belongs to the field of image detection technology, and discloses a myocardial image detection device, system and storage medium. The present invention obtains a myocardial segmentation image by segmenting the target myocardial image, and generates a myocardial outer ring image based on the myocardial segmentation image; samples the myocardial outer ring image to obtain a sampled image, and extracts multiple coronary centerline key points from the sampled image; intercepts a candidate area image from the myocardial outer ring image based on multiple coronary centerline key points; and performs feature analysis on the candidate area image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate area image. Since the area suspected of having a myocardial bridge is determined based on the coronary centerline key points, the candidate area image is intercepted from the myocardial outer ring image, and then the candidate area image is further feature analyzed through a preset classification model, it can be determined whether each candidate area has a myocardial bridge, and no manual participation is required as a whole, and no coronary segmentation is required, and the detection accuracy is high.

Description

Technical Field

[0001] The present invention relates to the field of image detection technology, and in particular to a myocardial image detection device, system and storage medium. Background Art

[0002] Currently, the diagnosis of myocardial bridge is mainly carried out through two methods:

[0003] 1. Relying on doctors to make manual diagnosis based on myocardial images, this solution requires doctors to have rich diagnostic experience and the diagnosis time will be longer. In addition, different doctors will diagnose different results for the same blood vessel, and the results of the same doctor's diagnosis at different times may also be different. Therefore, manual diagnosis is basically not reproducible and the effect is not ideal in actual use.

[0004] 2. The presence of a myocardial bridge is determined by obtaining the myocardial segmentation results and calculating whether the vascular segmentation is adhered to the myocardial segmentation results. However, the implementation of this solution requires accurate coronary vessel segmentation, which is extremely difficult to achieve and difficult to ensure accuracy. As a result, the accuracy of myocardial bridge detection is also difficult to ensure. Moreover, merely determining whether there is adhesion does not ensure that all conditions for determining the presence of a myocardial bridge are met, which can easily lead to overdiagnosis or missed diagnosis, and the effect is not ideal in actual use.

[0005] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0006] The main purpose of the present invention is to provide a myocardial image detection device, system and storage medium, aiming to solve the technical problem that the prior art cannot accurately detect whether a myocardial bridge exists based on myocardial images.

[0007] To achieve the above object, the present invention provides a myocardial image detection device, the myocardial image detection device comprising: a processor, a memory, and a myocardial image detection program stored in the memory and executable on the processor, the myocardial image detection program being configured to implement the following steps:

[0008] Segmenting the target myocardial image to obtain a myocardial segmentation image, and generating a myocardial outer ring image based on the myocardial segmentation image;

[0009] Sampling the myocardial outer ring image to obtain a sampled image, and extracting a plurality of coronary artery centerline key points from the sampled image;

[0010] intercepting a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points;

[0011] The candidate region image is subjected to feature analysis through a preset classification model to determine a myocardial bridge detection result corresponding to the candidate region image.

[0012] Optionally, the myocardial image detection program is configured to implement the following steps:

[0013] Segmenting the target myocardial image by using a preset segmentation model to obtain a myocardial segmentation image;

[0014] dilating the myocardial segmentation image to obtain a masked myocardial image;

[0015] Subtracting the masked myocardial image from the myocardial segmentation image to obtain a myocardial ring image;

[0016] The myocardial annular image is attached to the target myocardial image to obtain a myocardial outer annular image.

[0017] Optionally, the myocardial image detection program is configured to implement the following steps:

[0018] Connect the key points of each coronary artery centerline to obtain the key point direction vector corresponding to each coronary artery centerline key point;

[0019] Determine the perpendicular line from the key points of each coronary artery centerline to the myocardial surface;

[0020] Generate an angle variation curve according to the vertical line and the key point direction vector;

[0021] A candidate region image is intercepted from the myocardial outer ring image according to the angle variation curve.

[0022] Optionally, the myocardial image detection program is configured to implement the following steps:

[0023] Calculate the distance from each key point of the coronary artery centerline to all points on the myocardial surface, and take the point on the myocardial surface with the smallest distance as the target image point;

[0024] A perpendicular line from each key point of the coronary artery centerline to the myocardial surface is generated according to the target image point.

[0025] Optionally, the myocardial image detection program is configured to implement the following steps:

[0026] Obtaining a curve change gradient corresponding to the angle change curve;

[0027] The portion of the angle change curve corresponding to a curve change gradient greater than or equal to a preset gradient threshold is taken as a candidate curve region;

[0028] A candidate region image is intercepted from the myocardial outer ring image according to the candidate curve region.

[0029] Optionally, the myocardial image detection program is configured to implement the following steps:

[0030] Constructing a candidate region centerline according to the coronary artery centerline key points contained in the candidate region image;

[0031] Extracting texture features of the myocardium and the coronary artery based on the candidate region image and the center line of the candidate region to obtain texture feature data;

[0032] The texture feature data is feature classified using a preset classification model to obtain a myocardial bridge detection result corresponding to the candidate region image.

[0033] Optionally, the myocardial image detection program is configured to implement the following steps:

[0034] Dividing the myocardial outer ring image into regions according to a preset region division rule;

[0035] Performing local sampling based on the regions divided in the myocardial outer ring image to obtain a sampling image;

[0036] The key points of the sampled image are extracted using a preset key point extraction model to obtain a plurality of coronary artery centerline key points.

[0037] Optionally, the myocardial image detection program is configured to implement the following steps:

[0038] Generate a myocardial bridge detection report according to the candidate area image and the myocardial bridge detection result corresponding to the candidate area image;

[0039] The myocardial bridge detection report is displayed.

[0040] In addition, to achieve the above-mentioned purpose, the present invention also proposes a myocardial image detection system, which includes: an image segmentation module, a center extraction module, a region determination module, and an image analysis module;

[0041] The image segmentation module is used to segment the target myocardial image to obtain a myocardial segmentation image, and generate a myocardial outer ring image based on the myocardial segmentation image;

[0042] The center extraction module is used to sample the myocardial outer ring image to obtain a sampled image, and extract a plurality of coronary centerline key points from the sampled image;

[0043] The region determination module is used to intercept a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points;

[0044] The image analysis module is used to perform feature analysis on the candidate region image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate region image.

[0045] In addition, to achieve the above-mentioned purpose, the present invention further proposes a computer-readable storage medium, on which a myocardial image detection program is stored, and when the myocardial image detection program is executed, the following steps are implemented:

[0046] Segmenting the target myocardial image to obtain a myocardial segmentation image, and generating a myocardial outer ring image based on the myocardial segmentation image;

[0047] Sampling the myocardial outer ring image to obtain a sampled image, and extracting a plurality of coronary artery centerline key points from the sampled image;

[0048] intercepting a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points;

[0049] The candidate region image is subjected to feature analysis through a preset classification model to determine a myocardial bridge detection result corresponding to the candidate region image.

[0050] The present invention obtains a myocardial segmentation image by segmenting the target myocardial image, and generates a myocardial outer ring image based on the myocardial segmentation image; samples the myocardial outer ring image to obtain a sampled image, and extracts multiple coronary centerline key points from the sampled image; intercepts a candidate area image from the myocardial outer ring image based on multiple coronary centerline key points; and performs feature analysis on the candidate area image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate area image. Since the area suspected of having a myocardial bridge is determined based on the coronary centerline key points, the candidate area image is intercepted from the myocardial outer ring image, and then the candidate area image is further feature analyzed through a preset classification model, it can be determined whether each candidate area has a myocardial bridge, and no manual participation is required, and no coronary segmentation is required, and the detection accuracy is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the structure of an electronic device in a hardware operating environment involved in an embodiment of the present invention;

[0052] Figure 2 It is a schematic diagram of the flow chart of the first embodiment of the myocardial image detection device of the present invention;

[0053] Figure 3 It is a schematic diagram of the flow chart of the second embodiment of the myocardial image detection device of the present invention;

[0054] Figure 4 It is a structural block diagram of the first embodiment of the myocardial image detection system of the present invention.

[0055] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0056] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0057] Reference Figure 1 , Figure 1 The figure is a schematic diagram of the structure of a myocardial image detection device in the hardware operating environment involved in the embodiment of the present invention.

[0058] like Figure 1 As shown, the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0059] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the electronic device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0060] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a myocardial image detection program.

[0061] exist Figure 1In the electronic device shown, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the electronic device of the present invention can be set in a myocardial image detection device, and the electronic device calls the myocardial image detection program stored in the memory 1005 through the processor 1001.

[0062] The embodiment of the present invention provides a myocardial image detection device, referring to Figure 2 , Figure 2 The figure is a flow chart of a first embodiment of a myocardial image detection device according to the present invention.

[0063] This embodiment provides a myocardial image detection device, which includes: a processor, a memory, and a myocardial image detection program stored in the memory and executable on the processor, wherein the myocardial image detection program is configured to implement the following steps:

[0064] Step S10: Segment the target myocardial image to obtain a myocardial segmentation image, and generate a myocardial outer ring image based on the myocardial segmentation image.

[0065] It should be noted that the target myocardial image may be a myocardial image that needs to be detected to determine whether a myocardial bridge exists, and the target myocardial image may be a coronary CT angiography image (CCTA image) acquired by medical equipment. The myocardial outer ring image may be an image used to represent the myocardial contour.

[0066] In actual use, segmenting the target myocardial image to obtain the myocardial segmentation image may be performed by segmenting the myocardium in the target myocardial image using a preset myocardial segmentation tool, thereby obtaining the myocardial segmentation image. Generating the myocardial outer ring image based on the myocardial segmentation image may be performed by depicting the contour of the myocardium according to the myocardial segmentation result in the myocardial segmentation image, thereby obtaining the myocardial outer ring image.

[0067] Furthermore, in order to ensure the accuracy of myocardial segmentation and the validity of the myocardial outer ring image, step S10 in this embodiment may include:

[0068] Segmenting the target myocardial image by using a preset segmentation model to obtain a myocardial segmentation image;

[0069] dilating the myocardial segmentation image to obtain a masked myocardial image;

[0070] Subtracting the masked myocardial image from the myocardial segmentation image to obtain a myocardial ring image;

[0071] The myocardial annular image is attached to the target myocardial image to obtain a myocardial outer annular image.

[0072] It should be noted that the preset segmentation model can be a model obtained after training the initial segmentation model, and the initial segmentation model can be a deep neural network model. When training the initial segmentation model, a large number of CCTA images with artificially marked myocardial bridge areas can be input into the initial segmentation model, and each output will calculate the error between the predicted result and the artificially marked result, and feed it back to the initial segmentation model. With the update and iteration of the data, the weight of the network in the initial segmentation model is continuously adjusted to achieve error convergence, and when the error convergence is achieved, the adjusted initial segmentation model at this time is used as the preset segmentation model.

[0073] It can be understood that the preset segmentation model obtained after training with a large number of samples has extremely high accuracy when performing myocardial segmentation, and its segmentation efficiency is extremely high. Therefore, segmenting the target myocardial image through the preset segmentation model and obtaining the myocardial segmentation image can ensure the accuracy of myocardial segmentation and can greatly increase the execution efficiency of myocardial segmentation.

[0074] It should be noted that the myocardial segmentation image is expanded to obtain the masked myocardial image, which may be to expand the myocardium in the myocardial segmentation image according to a preset ratio, and use the expanded myocardial segmentation image as the masked myocardial image. The masked myocardial image is subtracted from the myocardial segmentation image to obtain the myocardial annulus image, which may be to subtract the myocardium in the masked myocardial image from the myocardial segmentation image to obtain a torus representing the myocardial surface contour, thereby obtaining the myocardial annulus image. The myocardial annulus image is attached to the target myocardial image to obtain the myocardial outer annulus image, which may be to cover the torus in the myocardial annulus image with the target myocardial image, cover the myocardial area in the target myocardial image, and use the covered target myocardial image as the myocardial outer annulus image.

[0075] It is understandable that since the myocardial ring image only contains a ring body of the myocardial surface contour, it is difficult to identify the myocardial bridge based on this alone. It is necessary to combine the coronary blood vessels and other information in the target myocardial image to identify whether the myocardial bridge exists. In order to facilitate subsequent analysis, the myocardial ring image can be attached to the target myocardial image to obtain the myocardial outer ring image.

[0076] Step S20: sampling the myocardial outer ring image to obtain a sampled image, and extracting a plurality of coronary artery centerline key points from the sampled image.

[0077] It should be noted that the key point of the coronary centerline may be a key point on the coronary centerline within the myocardial outer ring. Sampling the myocardial outer ring image to obtain a sampled image may be sampling a region where a coronary vessel exists around the myocardium in the myocardial outer ring image, and then using the sampled image as the sampled image.

[0078] Further, in order to ensure the accuracy of the key points of the coronary centerline, step S20 in this embodiment may include:

[0079] Dividing the myocardial outer ring image into regions according to a preset region division rule;

[0080] Performing local sampling based on the regions divided in the myocardial outer ring image to obtain a sampling image;

[0081] The key points of the sampled image are extracted using a preset key point extraction model to obtain a plurality of coronary artery centerline key points.

[0082] It should be noted that there may be many areas with coronary blood vessels around the myocardium in the outer ring image of the myocardium. In this case, the outer ring image of the myocardium can be divided into regions according to preset region division rules, and the area with coronary blood vessels around the myocardium can be divided into multiple sub-regions. Then, each sub-region is locally sampled to obtain multiple sampled images, and then key points are extracted from each sampled image separately through a preset key point extraction model, and the extracted coronary centerline key points are summarized to obtain multiple coronary centerline key points.

[0083] Among them, the preset key point extraction model can be a pre-trained coronary centerline key point regression network model for extracting coronary artery centerline key points, wherein the coronary centerline key point regression network model can be a deep neural network model.

[0084] In actual use, when training the coronary centerline key point regression network model, a large number of myocardial outer ring images with manually marked centerlines can be used to iteratively train the network based on error feedback.

[0085] Step S30: intercepting a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points.

[0086] It should be noted that, extracting a candidate area image from the myocardial outer ring image based on multiple coronary centerline key points can be accomplished by connecting the coronary centerline key points to obtain the coronary centerline, determining the area in the myocardial outer ring image where a myocardial bridge is suspected to exist based on the coronary centerline, and extracting the image of the area where a myocardial bridge is suspected to exist, thereby obtaining a candidate area image.

[0087] Step S40: performing feature analysis on the candidate region image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate region image.

[0088] It should be noted that the preset classification model can be a pre-trained myocardial bridge classification network, wherein the myocardial bridge classification network can be a deep neural network. The preset classification model can perform feature classification on the input texture feature data to determine whether a myocardial bridge exists, and when a myocardial bridge exists, it can output the classification of the myocardial bridge, such as: deep myocardial bridge, superficial myocardial bridge. After determining the candidate area, the image features in the image of the candidate area can also be extracted, and the image features can be further analyzed by the preset classification model to determine whether a myocardial bridge exists, and when a myocardial bridge exists, the category of the myocardial bridge can be further determined, thereby obtaining a myocardial bridge detection result.

[0089] Furthermore, in order to ensure the accuracy of the myocardial bridge classification result, step S40 in this embodiment may include:

[0090] Constructing a candidate region centerline according to the coronary artery centerline key points contained in the candidate region image;

[0091] Extracting texture features of the myocardium and the coronary artery based on the candidate region image and the center line of the candidate region to obtain texture feature data;

[0092] The texture feature data is feature classified using a preset classification model to obtain a myocardial bridge detection result corresponding to the candidate region image.

[0093] It should be noted that constructing the candidate region centerline based on the coronary artery centerline key points contained in the candidate region image can be achieved by connecting multiple coronary artery centerline key points contained in the candidate region image and using the connecting line as the candidate region centerline.

[0094] In actual use, when extracting the texture features of the myocardium and coronary arteries in the candidate area image, the position data of the center line of the candidate area can provide position constraints. The texture feature data extracted in this way can better represent the relationship between the myocardium and the coronary arteries, that is, whether the center line of the coronary artery is embedded in the myocardium. When feature classification is performed based on this texture feature data, the accuracy of feature classification can be effectively improved.

[0095] Further, in order to facilitate the user to clearly understand whether there is a myocardial bridge in the target myocardial image, after step S40 in this embodiment, the following steps may be further included:

[0096] Generate a myocardial bridge detection report according to the candidate area image and the myocardial bridge detection result corresponding to the candidate area image;

[0097] The myocardial bridge detection report is displayed.

[0098] In actual use, generating a myocardial bridge detection report based on the candidate area image and the myocardial bridge detection results corresponding to the candidate area image can be to obtain a preset report template, and then fill the candidate area image and the myocardial bridge detection results corresponding to the candidate area image into the preset report template to generate a myocardial bridge detection report.

[0099] In a specific implementation, displaying the myocardial bridge detection report can be by displaying the myocardial bridge detection report in a display module in the myocardial image detection device or on a display device connected to the myocardial image detection device. Of course, the myocardial bridge detection report can also be sent to the user's user terminal for display.

[0100] This embodiment obtains a myocardial segmentation image by segmenting the target myocardial image, and generates a myocardial outer ring image based on the myocardial segmentation image; samples the myocardial outer ring image to obtain a sampled image, and extracts multiple coronary centerline key points from the sampled image; intercepts a candidate area image from the myocardial outer ring image based on multiple coronary centerline key points; and performs feature analysis on the candidate area image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate area image. Since the area suspected of having a myocardial bridge is determined based on the coronary centerline key points, the candidate area image is intercepted from the myocardial outer ring image, and then the candidate area image is further feature analyzed through a preset classification model, it can be determined whether each candidate area has a myocardial bridge. No manual participation is required, and no coronary segmentation is required, and the detection accuracy is high.

[0101] refer to Figure 3 , Figure 3 It is a flow chart of the second embodiment of a myocardial image detection device of the present invention.

[0102] Based on the above first embodiment, in this embodiment, step S30 may include:

[0103] Step S301: Connect the key points of the coronary centerline to obtain the key point direction vector corresponding to each key point of the coronary centerline.

[0104] It should be noted that connecting the key points of the coronary centerline to obtain the key point direction vector corresponding to each coronary centerline key point can be achieved by connecting the position of each coronary centerline key point with the next coronary centerline key point in a preset direction order, and using the vector obtained after the connection as the key point direction vector corresponding to the coronary centerline key point, thereby obtaining the key point direction vector corresponding to each coronary centerline key point. The preset direction order can be set by the manager of the myocardial image detection device according to actual needs, for example, the preset direction order is set to be from top to bottom.

[0105] Step S302: Determine the perpendicular line from each key point of the coronary artery centerline to the myocardial surface.

[0106] It should be noted that the perpendicular line from each key point of the coronary artery centerline to the myocardial surface can be determined by determining the foot of the perpendicular corresponding to each key point of the coronary artery centerline on the myocardial surface, that is, the target image point, and then connecting the key point of the coronary artery centerline with the target image point, and using the connected line segment as the perpendicular line from the key point of the coronary artery centerline to the myocardial surface.

[0107] Further, in order to accurately determine the vertical foot corresponding to each key point of the coronary centerline on the myocardial surface, step S302 in this embodiment may include:

[0108] Calculate the distance from each key point of the coronary artery centerline to all points on the myocardial surface, and take the point on the myocardial surface with the smallest distance as the target image point;

[0109] A perpendicular line from each key point of the coronary artery centerline to the myocardial surface is generated according to the target image point.

[0110] It should be noted that the distance from the key point of the coronary centerline to all points on the myocardial surface can be calculated by selecting multiple points on the myocardial surface as candidate points, and then calculating the distance from the key point of the coronary centerline to each candidate point using the distance formula between two points.

[0111] It can be understood that if the distance corresponding to a candidate point is the smallest, it means that the candidate point is the foot of the corresponding perpendicular point of the coronary centerline key point. Therefore, the point with the smallest distance from the myocardial surface can be used as the target image point.

[0112] It should be noted that generating a perpendicular line from the key point of the coronary artery centerline to the myocardial surface based on the target image point can be performed by connecting the key point of the coronary artery centerline with the target image point, and using the connected line segment as the perpendicular line from the key point of the coronary artery centerline to the myocardial surface.

[0113] Step S303: generating an angle variation curve according to the vertical line and the key point direction vector.

[0114] It should be noted that generating the angle variation curve according to the vertical line and the key point direction vector can be by obtaining the vertical line and the key point direction vector corresponding to each coronary artery centerline key point, calculating the angle between the key point direction vector and the vertical line, thereby obtaining the angle corresponding to each coronary artery centerline key point, and constructing the angle variation curve according to the angle corresponding to each coronary artery centerline key point according to the spatial order of each coronary artery centerline key point.

[0115] Step S304: intercepting a candidate region image from the myocardial outer ring image according to the angle variation curve.

[0116] It should be noted that, extracting the candidate area image from the myocardial outer ring image according to the angle change curve can be performed by determining the faster changing part of the angle change curve as the target part, taking the coronary artery centerline key point corresponding to the target part as the target coronary artery centerline key point, extracting the image of the area corresponding to the target coronary artery centerline key point in the myocardial outer ring image, and using the extracted image as the candidate area image.

[0117] Furthermore, in order to quickly determine the candidate region image, step S304 in this embodiment may include:

[0118] Obtaining a curve change gradient corresponding to the angle change curve;

[0119] The portion of the angle change curve corresponding to a curve change gradient greater than or equal to a preset gradient threshold is taken as a candidate curve region;

[0120] A candidate region image is intercepted from the myocardial outer ring image according to the candidate curve region.

[0121] It should be noted that the speed of change of the curve can be represented by the gradient of the curve. The gradient of the curve change can be the gradient of each curve part of the angle change curve.

[0122] In actual use, obtaining the curve change gradient corresponding to the angle change curve may be to calculate the gradient of the angle change curve using a preset gradient calculation formula or a preset gradient calculation tool, and use the calculated gradient as the curve change gradient.

[0123] In a specific implementation, if the curve change gradient corresponding to a certain part is greater than or equal to a preset gradient threshold, it means that the change speed of this part of the angle change curve is too fast, so this part can be used as a candidate curve area.

[0124] In actual use, the candidate area image can be intercepted from the myocardial outer ring image based on the candidate curve area. The coronary artery centerline key point corresponding to the candidate curve area can be used as the target coronary artery centerline key point, and the image of the area corresponding to the target coronary artery centerline key point in the myocardial outer ring image can be intercepted, and the intercepted image can be used as the candidate area image.

[0125] This embodiment connects the key points of the coronary centerline to obtain the key point direction vector corresponding to each key point of the coronary centerline; determines the vertical line from each key point of the coronary centerline to the myocardial surface; generates an angle change curve according to the vertical line and the key point direction vector; and intercepts the candidate area image from the myocardial outer ring image according to the angle change curve. Since the angle change curve is generated according to the vertical line and the key point direction vector corresponding to each key point of the coronary centerline, the area suspected of having a myocardial bridge can be quickly screened out according to the change law of the angle change curve, thereby improving the execution efficiency of myocardial bridge detection.

[0126] In addition, an embodiment of the present invention further provides a storage medium, on which a myocardial image detection program is stored. When the myocardial image detection program is executed by a processor, the following steps are implemented:

[0127] Segmenting the target myocardial image to obtain a myocardial segmentation image, and generating a myocardial outer ring image based on the myocardial segmentation image;

[0128] Sampling the myocardial outer ring image to obtain a sampled image, and extracting a plurality of coronary artery centerline key points from the sampled image;

[0129] intercepting a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points;

[0130] The candidate region image is subjected to feature analysis through a preset classification model to determine a myocardial bridge detection result corresponding to the candidate region image.

[0131] Reference Figure 4 , Figure 4 It is a structural block diagram of the first embodiment of the myocardial image detection system of the present invention.

[0132] like Figure 4 As shown, the myocardial image detection system proposed in the embodiment of the present invention includes: an image segmentation module 10, a center extraction module 20, a region determination module 30, and an image analysis module 40;

[0133] The image segmentation module 10 is used to segment the target myocardial image to obtain a myocardial segmentation image, and generate a myocardial outer ring image based on the myocardial segmentation image;

[0134] The center extraction module 20 is used to sample the myocardial outer ring image to obtain a sampled image, and extract a plurality of coronary artery centerline key points from the sampled image;

[0135] The region determination module 30 is used to intercept a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points;

[0136] The image analysis module 40 is used to perform feature analysis on the candidate region image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate region image.

[0137] This embodiment obtains a myocardial segmentation image by segmenting the target myocardial image, and generates a myocardial outer ring image based on the myocardial segmentation image; samples the myocardial outer ring image to obtain a sampled image, and extracts multiple coronary centerline key points from the sampled image; intercepts a candidate area image from the myocardial outer ring image based on multiple coronary centerline key points; and performs feature analysis on the candidate area image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate area image. Since the area suspected of having a myocardial bridge is determined based on the coronary centerline key points, the candidate area image is intercepted from the myocardial outer ring image, and then the candidate area image is further feature analyzed through a preset classification model, it can be determined whether each candidate area has a myocardial bridge. No manual participation is required, and no coronary segmentation is required, and the detection accuracy is high.

[0138] Furthermore, the image segmentation module 10 is also used to segment the target myocardial image through a preset segmentation model to obtain a myocardial segmentation image; to dilate the myocardial segmentation image to obtain a masked myocardial image; to subtract the masked myocardial image from the myocardial segmentation image to obtain a myocardial ring image; and to attach the myocardial ring image to the target myocardial image to obtain a myocardial outer ring image.

[0139] Furthermore, the region determination module 30 is also used to connect each coronary centerline key point to obtain a key point direction vector corresponding to each coronary centerline key point; determine a perpendicular line from each coronary centerline key point to the myocardial surface; generate an angle change curve based on the perpendicular line and the key point direction vector; and intercept a candidate region image from the myocardial outer ring image based on the angle change curve.

[0140] Furthermore, the area determination module 30 is also used to calculate the distance from each key point of the coronary centerline to all points on the myocardial surface, and take the point with the smallest distance corresponding to the myocardial surface as the target image point; and generate a vertical line from each key point of the coronary centerline to the myocardial surface based on the target image point.

[0141] Furthermore, the area determination module 30 is also used to obtain the curve change gradient corresponding to the angle change curve; take the part of the angle change curve whose corresponding curve change gradient is greater than or equal to a preset gradient threshold as a candidate curve area; and intercept the candidate area image from the myocardial outer ring image according to the candidate curve area.

[0142] Furthermore, the image analysis module 40 is also used to construct a candidate area centerline based on the coronary artery centerline key points contained in the candidate area image; extract texture features of the myocardium and the coronary artery based on the candidate area image and the candidate area centerline to obtain texture feature data; and perform feature classification on the texture feature data through a preset classification model to obtain a myocardial bridge detection result corresponding to the candidate area image.

[0143] Furthermore, the center extraction module 20 is also used to divide the myocardial outer ring image into regions according to preset region division rules; perform local sampling based on the regions divided in the myocardial outer ring image to obtain a sampled image; and extract key points from the sampled image using a preset key point extraction model to obtain multiple coronary centerline key points.

[0144] Furthermore, the image analysis module 40 is also used to generate a myocardial bridge detection report according to the candidate area image and the myocardial bridge detection result corresponding to the candidate area image; and display the myocardial bridge detection report.

[0145] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0146] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0147] In addition, it should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0148] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0149] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0150] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A myocardial image detection device, characterized in that: The myocardial image detection device comprises: a processor, a memory, and a myocardial image detection program stored in the memory and executable on the processor, wherein the myocardial image detection program is configured to implement the following steps: Segmenting the target myocardial image to obtain a myocardial segmentation image, and generating a myocardial outer ring image based on the myocardial segmentation image; Sampling the myocardial outer ring image to obtain a sampled image, and extracting a plurality of coronary artery centerline key points from the sampled image; intercepting a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points; Performing feature analysis on the candidate region image through a preset classification model to determine a myocardial bridge detection result corresponding to the candidate region image; Wherein, the step of intercepting the candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points includes: Connect the key points of each coronary artery centerline to obtain the key point direction vector corresponding to each coronary artery centerline key point; Determine the perpendicular line from the key points of each coronary artery centerline to the myocardial surface; Generate an angle variation curve according to the vertical line and the key point direction vector; Obtaining a curve change gradient corresponding to the angle change curve; The portion of the angle change curve corresponding to a curve change gradient greater than or equal to a preset gradient threshold is taken as a candidate curve region; A candidate region image is intercepted from the myocardial outer ring image according to the candidate curve region.

2. The myocardial image detection device according to claim 1, characterized in that: The myocardial image detection program is configured to implement the following steps: Segmenting the target myocardial image by using a preset segmentation model to obtain a myocardial segmentation image; dilating the myocardial segmentation image to obtain a masked myocardial image; Subtracting the masked myocardial image from the myocardial segmentation image to obtain a myocardial ring image; The myocardial annular image is attached to the target myocardial image to obtain a myocardial outer annular image.

3. The myocardial image detection device according to claim 2, characterized in that: The myocardial image detection program is configured to implement the following steps: Calculate the distance from each key point of the coronary artery centerline to all points on the myocardial surface, and take the point on the myocardial surface with the smallest distance as the target image point; A perpendicular line from each key point of the coronary artery centerline to the myocardial surface is generated according to the target image point.

4. The myocardial image detection device according to claim 1, characterized in that: The myocardial image detection program is configured to implement the following steps: Constructing a candidate region centerline according to the coronary artery centerline key points contained in the candidate region image; Extracting texture features of the myocardium and the coronary artery based on the candidate region image and the center line of the candidate region to obtain texture feature data; The texture feature data is feature classified using a preset classification model to obtain a myocardial bridge detection result corresponding to the candidate region image.

5. The myocardial image detection device according to claim 1, characterized in that: The myocardial image detection program is configured to implement the following steps: Dividing the myocardial outer ring image into regions according to a preset region division rule; Performing local sampling based on the regions divided in the myocardial outer ring image to obtain a sampling image; The key points of the sampled image are extracted using a preset key point extraction model to obtain a plurality of coronary artery centerline key points.

6. The myocardial image detection device according to any one of claims 1 to 5, characterized in that: The myocardial image detection program is configured to implement the following steps: Generate a myocardial bridge detection report according to the candidate area image and the myocardial bridge detection result corresponding to the candidate area image; The myocardial bridge detection report is displayed.

7. A myocardial image detection system, characterized in that: The myocardial image detection system comprises: an image segmentation module, a center extraction module, a region determination module, and an image analysis module; The image segmentation module is used to segment the target myocardial image to obtain a myocardial segmentation image, and generate a myocardial outer ring image based on the myocardial segmentation image; The center extraction module is used to sample the myocardial outer ring image to obtain a sampled image, and extract a plurality of coronary centerline key points from the sampled image; The region determination module is used to intercept a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points; The image analysis module is used to perform feature analysis on the candidate region image through a preset classification model to determine the myocardial bridge detection result corresponding to the candidate region image; The region determination module is further used to connect the key points of the coronary centerline to obtain the key point direction vector corresponding to each coronary centerline key point; determine the perpendicular line from each coronary centerline key point to the myocardial surface; generate an angle change curve according to the perpendicular line and the key point direction vector; obtain the curve change gradient corresponding to the angle change curve; use the part of the angle change curve whose corresponding curve change gradient is greater than or equal to a preset gradient threshold as a candidate curve region; and intercept a candidate region image from the myocardial outer ring image according to the candidate curve region.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a myocardial image detection program, and the myocardial image detection program implements the following steps when executed: Segmenting the target myocardial image to obtain a myocardial segmentation image, and generating a myocardial outer ring image based on the myocardial segmentation image; Sampling the myocardial outer ring image to obtain a sampled image, and extracting a plurality of coronary artery centerline key points from the sampled image; intercepting a candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points; Performing feature analysis on the candidate region image through a preset classification model to determine a myocardial bridge detection result corresponding to the candidate region image; Wherein, the step of intercepting the candidate region image from the myocardial outer ring image based on the multiple coronary centerline key points includes: Connect the key points of each coronary artery centerline to obtain the key point direction vector corresponding to each coronary artery centerline key point; Determine the perpendicular line from the key points of each coronary artery centerline to the myocardial surface; Generate an angle variation curve according to the vertical line and the key point direction vector; Obtaining a curve change gradient corresponding to the angle change curve; The portion of the angle change curve corresponding to a curve change gradient greater than or equal to a preset gradient threshold is taken as a candidate curve region; A candidate region image is intercepted from the myocardial outer ring image according to the candidate curve region.

Images