Blood vessel radius prediction method and device and computer readable storage medium

Abstract

The invention discloses a blood vessel radius prediction method and device and a computer readable storage medium, and can be applied to the technical field of medical image processing, and the method comprises the steps: obtaining a center line of a target blood vessel; obtaining a straightened image corresponding to the target blood vessel based on the center line and the original contrast image of the target blood vessel; dividing the straightened image into a plurality of straightened sub-images in the blood vessel direction in the straightened image; inputting the plurality of data groups into a blood vessel radius prediction model, and predicting to obtain radius information of the target blood vessel; wherein one data set in the plurality of data sets at least comprises one sub-straightened image in the plurality of sub-straightened images. By means of the mode, accurate blood vessel radius information can be obtained through prediction, and therefore the accurate blood vessel contour can be obtained.

Description

Technical Field

[0001] This application relates to the field of medical image processing technology, and particularly to a method, device, and computer-readable storage medium for predicting blood vessel radius. Background Art

[0002] In the diagnosis and treatment of blood vessel diseases, angiography technology plays a crucial role. This technology makes blood vessels clearly visible in images by injecting contrast agents, thereby helping doctors evaluate the condition of blood vessels.

[0003] Traditional angiography analysis mainly relies on image segmentation models. This technology identifies and segments the blood vessel foreground in angiography images, and then processes the foreground to extract the boundary of the blood vessel as a contour. To a certain extent, this method can help doctors with condition assessment and treatment decision-making. However, this method also has certain limitations. For example, the extraction accuracy is relatively low; the contour extraction effect for stenotic blood vessels is poor, and even the contour of stenotic blood vessels cannot be extracted, resulting in an interruption phenomenon in the blood vessel contour. Summary of the Invention

[0004] This application discloses a method, device, and computer-readable storage medium for predicting blood vessel radius, which can predict relatively accurate blood vessel radius information, thereby facilitating the obtaining of accurate blood vessel contours, and the prediction effect for stenotic blood vessels is also accurate.

[0005] In a first aspect, an embodiment of this application provides a method for predicting blood vessel radius. The method includes: obtaining the centerline of a target blood vessel; obtaining a straightened image corresponding to the target blood vessel based on the centerline and the original angiography image of the target blood vessel; dividing the straightened image into multiple sub-straightened images in the blood vessel direction of the straightened image; inputting multiple data groups into a blood vessel radius prediction model to predict the radius information of the target blood vessel; where one data group in the multiple data groups includes at least one sub-straightened image in the multiple sub-straightened images.

[0006] In this technical solution, on the one hand, the radius information of the target blood vessel can be accurately predicted, which is conducive to obtaining an accurate blood vessel contour. On the second hand, when using the image segmentation technology to extract the blood vessel contour, the effect of extracting the contour of the stenotic blood vessel is poor. In contrast, the method adopted in the embodiment of the present application can obtain an accurate blood vessel contour through the predicted radius information of the blood vessel, and the method adopted in the embodiment of the present application has the same accurate prediction effect on the stenotic blood vessel. On the third hand, when using the image segmentation technology to extract the blood vessel contour, the extracted contour is relatively rough and the edge of the blood vessel contour is not smooth. In contrast, the method adopted in the embodiment of the present application can obtain an accurate, clear and smooth-edge blood vessel contour through the predicted radius information of the blood vessel. On the fourth hand, when using the image segmentation technology to extract the blood vessel contour, if the same image segmentation model is used for different types of blood vessels, the segmentation effect will be unsatisfactory. If different image segmentation models are used, higher-performance devices are required to meet the requirements. In contrast, the method adopted in the embodiment of the present application can use the same model (i.e., the blood vessel radius prediction model) to predict different types of blood vessels, which can reduce the requirements for device performance.

[0007] In a possible implementation manner, based on the centerline and the original angiography image of the target blood vessel, a straightened image corresponding to the target blood vessel is obtained, including: based on a plurality of sampling points on the centerline, a plurality of cropped images are intercepted from the original angiography image of the target blood vessel, and different cropped images are centered on different sampling points; based on the plurality of cropped images, a straightened image corresponding to the target blood vessel is obtained.

[0008] In a possible implementation manner, the plurality of sampling points include a first sampling point, and the plurality of cropped images include a first cropped image, and the first cropped image is centered on the first sampling point; based on the plurality of sampling points on the centerline, intercepting a plurality of cropped images from the original angiography image of the target blood vessel, including: based on the tangent line and the normal line of the first sampling point on the centerline, and the size information of the intercepting frame, intercepting the first cropped image from the original angiography image of the target blood vessel; wherein, the intercepting frame is a rectangle, and two adjacent sides of the intercepting frame are respectively parallel to the tangent line and the normal line of the first sampling point on the centerline.

[0009] In this technical solution, it is beneficial to make the straightened image closer to the effect of the real straightened blood vessel, which is further beneficial to improving the accuracy of subsequent prediction of the blood vessel radius.

[0010] In a possible implementation, obtaining a straightened image corresponding to a target blood vessel based on the multiple cropped images includes: splicing the multiple cropped images according to the tangent directions of the respective sampling points corresponding to the multiple cropped images in the center line to obtain a straightened image corresponding to the target blood vessel; or splicing the regional images of the respective cropped images among the multiple cropped images according to the tangent directions of the respective sampling points corresponding to the multiple cropped images in the center line to obtain a straightened image corresponding to the target blood vessel, where the regional image of the first cropped image includes: the pixel information of a row where the first sampling point is located in the first cropped image.

[0011] In this technical solution, after re-cropping the cropped images and then splicing the regional images obtained after the re-cropping, it is beneficial for the straightened image to be closer to the effect of a real straightened blood vessel, and thus beneficial for improving the accuracy of predicting the blood vessel radius in the subsequent process.

[0012] In a possible implementation, the multiple sub-straightened images include a first sub-straightened image and a second sub-straightened image, and the first sub-straightened image and the second sub-straightened image are adjacent sub-straightened images in the straightened image; the blood vessel radius prediction model is used to: predict the radius information of a third blood vessel based on the radius information of a first blood vessel and the radius information of a second blood vessel; where the first blood vessel is the blood vessel in the first sub-straightened image, the second blood vessel is the blood vessel in the second sub-straightened image, and the third blood vessel includes a section of blood vessel located between the first blood vessel and the second blood vessel; the radius information of the target blood vessel includes at least the radius information of the first blood vessel, the radius information of the second blood vessel, and the radius information of the third blood vessel.

[0013] In this technical solution, the radius information of the second blood vessel can be predicted. Based on more adjacent sub-straightened images, sufficient radius information can be predicted, so that the predicted radius information of the target blood vessel can be relatively accurate, which is beneficial for obtaining an accurate blood vessel contour. In addition, it can also make the edge of the blood vessel contour smoother and the fitting effect better.

[0014] In a possible implementation, the multiple data groups include a first data group, and the first data group includes a third sub-straightened image among the multiple sub-straightened images; the first data group further includes one or more of the following: the curvature information of the blood vessel in the third sub-straightened image, the position information of the third sub-straightened image in the straightened image.

[0015] In this technical solution, by inputting the curvature information of the blood vessel in the sub-straightened image into the blood vessel radius prediction model, it is beneficial for improving the prediction accuracy.

[0016] In a possible implementation, the curvature information of the blood vessels in the third sub-straightened image includes: the sum of the vectors of multiple tangent vectors of the blood vessels in the third sub-straightened image in the horizontal axis direction, and the sum of the vectors of the multiple tangent vectors in the vertical axis direction.

[0017] In a second aspect, an embodiment of the present application provides a blood vessel radius prediction device, and the device includes a unit for implementing the method described in the first aspect.

[0018] In a third aspect, an embodiment of the present application provides another blood vessel radius prediction device, including a processor; the processor is configured to execute the method described in the first aspect.

[0019] In an alternative implementation, the blood vessel radius prediction device may further include a memory, and the memory is connected to the processor; the memory is used to store computer programs or instructions; the processor is specifically configured to call the computer programs or instructions from the memory and execute the method described in the first aspect.

[0020] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, and the computer-readable storage medium stores computer programs or computer instructions, and when the computer programs or computer instructions are executed, the computer is caused to execute the method described in the first aspect.

[0021] In a fifth aspect, an embodiment of the present application provides a computer program product including computer programs or instructions, and when the computer programs or instructions run on a computer, the computer is caused to execute the method described in the first aspect. Description of the Drawings

[0022] Figure 1 is a flowchart of a blood vessel radius prediction method provided by an embodiment of the present application;

[0023] Figure 2 is a schematic diagram of the center line of a target blood vessel provided by an embodiment of the present application;

[0024] Figure 3 is a schematic diagram of a blood vessel map provided by an embodiment of the present application;

[0025] Figure 4 is a schematic diagram of intercepting 14 cropped images from an original angiogram provided by an embodiment of the present application;

[0026] Figure 5 is a schematic diagram of stitching 14 cropped images to obtain a straightened image provided by an embodiment of the present application;

[0027] Figure 6 is a schematic diagram of a regional image of a cropped image provided by an embodiment of the present application;

[0028] Figure 7 is a schematic diagram provided by an embodiment of the present application for dividing the straightened image shown in Figure 5 into 7 sub-straightened images;

[0029] Figure 8 is a schematic diagram provided by an embodiment of the present application for intercepting 5 cropped images from the original angiographic image;

[0030] Figure 9 is a schematic diagram provided by an embodiment of the present application for splicing the 5 cropped images shown in Figure 8 vertically to obtain a straightened image of the target blood vessel;

[0031] Figure 10 is a schematic diagram of an actual angiographic image of the target blood vessel provided by an embodiment of the present application;

[0032] Figure 11 is a schematic diagram provided by an embodiment of the present application for intercepting a cropped image from the original angiographic image shown in Figure 10 ;

[0033] Figure 12 is a schematic diagram of a straightened image corresponding to the target blood vessel obtained based on the cropped image provided by an embodiment of the present application;

[0034] Figure 13 is a schematic diagram of a contour image of the target blood vessel provided by an embodiment of the present application;

[0035] Figure 14 is a schematic diagram of the structure of a blood vessel radius prediction device provided by an embodiment of the present application;

[0036] Figure 15 is a schematic diagram of the structure of another blood vessel radius prediction device provided by an embodiment of the present application. Specific embodiments

[0037] It should be understood that the terms "first", "second", etc. involved in the embodiments of the present application are used to distinguish different objects, rather than to describe a specific order. The "at least one" in the embodiments of the present application refers to one or more, and the plurality refers to two or more. The "and / or" in the embodiments of the present application describes the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B may represent the following three situations: A exists alone, A and B exist simultaneously, and B exists alone. Among them, A and B may be singular or plural. The character " / " may represent an "or" relationship between the preceding and following associated objects. In addition, the symbol " / " may also represent a division sign, that is, perform a division operation.

[0038] The "at least one (individual)" or its similar expression in the embodiments of the present application refers to any combination of these items, including any combination of single item (individual) or plural items (individuals). For example, at least one (individual) of a, b, or c can represent the following seven cases: a, b, c, a and b, a and c, b and c, a, b, and c. Among them, each of a, b, and c can be an element or a set containing one or more elements.

[0039] In the embodiments of the present application, the terms "corresponding", "associated, related", "corresponding, relevant", and "mapped" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the concepts or meanings to be expressed are the same.

[0040] The method for predicting the blood vessel radius proposed in the embodiments of the present application will be described below. This method for predicting the blood vessel radius can be executed by an electronic device, or by a device (such as a chip, a chip module, or a processor, etc., which are devices placed inside the electronic device) that is matched with the electronic device. In the embodiments of the present application, the case where the method for predicting the blood vessel radius is executed by an electronic device is taken as an example for illustration. Among them, the electronic device can include a terminal device, a network device, a server, a cluster, etc. The embodiments of the present application do not limit the specific form of the electronic device.

[0041] Please refer to Figure 1 , Figure 1 which is a schematic flow chart of a method for predicting the blood vessel radius provided by the embodiments of the present application. As Figure 1 shown, the method for predicting the blood vessel radius may include but is not limited to the following steps.

[0042] S101: Obtain the center line of the target blood vessel.

[0043] Among them, the target blood vessel is the blood vessel for which the blood vessel radius needs to be predicted.

[0044] The center line of the target blood vessel can be obtained by the electronic device from a contrast image with marking information, and this marking information is used to mark the center line of the target blood vessel. For example, the user can mark the center line of the target blood vessel on the original contrast image of the target blood vessel.

[0045] Alternatively, the electronic device can obtain the centerline of the target blood vessel in the following manner: using image segmentation technology, extract the foreground image of the target blood vessel from the original angiography image of the target blood vessel; process the foreground image to extract the blood vessel boundary as the contour of the target blood vessel; after performing skeletonization processing on the contour, obtain the centerline of the target blood vessel. Among them, processing the foreground image can include but is not limited to the following processing procedures: filtering processing, Otsu segmentation processing, denoising processing. Optionally, the electronic device can also use a Support Vector Machine (SVM) to extract the contours with plaques and thick curve noises from the background image of the target blood vessel. In this way, the accuracy of the extracted blood vessel contour can be improved, and the omission of extracting the blood vessel contour can be avoided.

[0046] Among them, SVM is a machine learning model widely used in statistical classification and regression analysis. Skeletonization processing is the process of converting a two-dimensional or three-dimensional image of a shape or object into its skeleton (or topological medial axis).

[0047] Please refer to Figure 2 , Figure 2 for a schematic diagram of the centerline of the target blood vessel. As Figure 2 shown, the gray-filled part represents the target blood vessel, and the blood line part represents the centerline of the target blood vessel.

[0048] It should be noted that the embodiment of the present application does not limit the method for obtaining the centerline of the target blood vessel; the embodiment of the present application does not limit the format of the original angiography image of the target blood vessel. For example, the format of the original angiography image can be the Digital Imaging and Communications in Medicine (DICOM) format or other formats. Among them, DICOM is an international standard for medical images and related information.

[0049] It should also be noted that the embodiment of the present application does not limit the type and specific morphology of the blood vessels involved. For example, the types of blood vessels mentioned in the embodiment of the present application can include but are not limited to: arteries, veins, capillaries. Another example is that taking the distribution according to human body regions as an example, the blood vessels mentioned in the embodiment of the present application can include but are not limited to: cardiac blood vessels, head blood vessels, neck blood vessels, upper limb blood vessels, chest blood vessels, abdominal blood vessels, pelvic blood vessels, lower limb blood vessels. Among them, cardiac blood vessels can include coronary blood vessels, and coronary blood vessels can include coronary arteries and coronary veins.

[0050] In a possible implementation manner, the electronic device can obtain a blood vessel map and determine the target blood vessel from the blood vessel map; among them, the blood vessel map can include at least one blood vessel. Taking Figure 3Taking the vascular map shown as an example, as Figure 3 shown, the vascular map includes multiple blood vessels, and the target blood vessel can be one of them. Figure 3 In Figure 3 , the target blood vessel is represented by the gray-filled part.

[0051] Optionally, the electronic device can receive user input and determine the target blood vessel based on the user input. For example, the user can input the starting position and the ending position of the target blood vessel, and the electronic device can determine the target blood vessel based on the starting position and the ending position. Alternatively, the electronic device can display the vascular map, and the display interface of the vascular map can further include the identification information of each blood vessel in the vascular map. The user can input the identification information of the target blood vessel so that the electronic device can determine the target blood vessel. Among them, the identification information of the blood vessel is used to uniquely identify a blood vessel in the vascular map. For example, the identification information can be a number.

[0052] Optionally, the electronic device can display the vascular map and determine the target blood vessel based on the trigger operation detected on the vascular map. For example, the trigger operation can be a sliding operation of the user along the target blood vessel on the display interface of the vascular map; or, the trigger operation can be a click operation of the user to select the starting point and the ending point of the target blood vessel on the display interface of the vascular map; or, the trigger operation can be a box selection operation of the user for the target blood vessel on the display interface of the vascular map; or, the trigger operation can be a selection operation of the user for the identification information of the target blood vessel on the display interface of the vascular map. The embodiments of the present application do not limit the manner in which the electronic device determines the target blood vessel.

[0053] S102: Based on the centerline and the original angiography image of the target blood vessel, obtain the straightened image corresponding to the target blood vessel.

[0054] In a possible implementation manner, the electronic device can intercept multiple cropped images from the original angiography image of the target blood vessel based on multiple sampling points on the centerline, and different cropped images are centered on different sampling points. Among them, the number of sampling points is the same as the number of obtained cropped images. Hereinafter, an example in which the number of sampling points is N will be described, and N is a positive integer greater than or equal to 2.

[0055] Taking N = 14 and the original angiography image of the target blood vessel as shown in Figure 3 as an example, the schematic diagram of intercepting 14 cropped images from the original angiography image is shown in Figure 4 shown. As shown in Figure 4 shown, the dots represent the sampling points, and the dashed rectangular frames represent the intercepting frames (or referred to as the first intercepting frames). One sampling point corresponds to one intercepting frame, and one sampling point corresponds to one cropped image. As shown in Figure 4 shown, 14 cropped images are obtained based on 14 sampling points.

[0056] It should be noted that the first cropping frame involved in the embodiments of the present application is used to indicate the area cropped from the original angiography image. In actual applications, the first cropping frame may or may not exist. For example, the electronic device can center on the sampling point and crop the pixels within a certain range, thereby obtaining the cropped image corresponding to the sampling point. In this case, even if there is no first cropping frame, the cropped image can still be obtained.

[0057] Optionally, the number of sampling points (i.e., the value of N), the size of the first cropping frame, and the shape of the first cropping frame can be default settings of the electronic device; or, the number of sampling points, the size of the first cropping frame, and the shape of the first cropping frame can be set or modified by the electronic device according to user operations. The embodiments of the present application do not limit the number of sampling points, the specific size of the first cropping frame, the specific shape of the first cropping frame, and the setting method. For example, the shape of the first cropping frame can also be a square or other shapes.

[0058] In a possible implementation manner, after the electronic device obtains N cropped images, it can obtain the straightened image corresponding to the target blood vessel based on the N cropped images. For example, the electronic device can splice the N cropped images to obtain the straightened image corresponding to the target blood vessel. Taking Figure 4 the 14 cropped images in Figure 5 as an example, the straightened image obtained by splicing the 14 cropped images is as shown in Figure 5 . It can be seen that this straightened image can indeed present the effect of the blood vessel being straightened to a certain extent.

[0059] Or, after the electronic device obtains N cropped images, it can splice the regional images of each cropped image among the N cropped images to obtain the straightened image corresponding to the target blood vessel. Taking Figure 14 the topmost cropped image in Figure 6 as an example, the regional image of this cropped image can be as shown in Figure 6 . In Figure 6 , the area covered by the dotted line represents the position of the regional image in the corresponding cropped image, and this dotted line part can also be called the second cropping frame (or secondary cropping frame). Based on the method shown in

[0060] , the regional images of other cropped images can be obtained, and then all the obtained regional images are spliced to obtain the straightened image of the target blood vessel. The method of splicing the regional images can refer to the method of splicing the cropped images described above, which will not be elaborated here.

[0061] It should be noted that the second intercepting frame used for re-clipping the clipped image is for illustration. In other embodiments, the length of the second intercepting frame may be less than the length of the clipped image. It should also be noted that the second intercepting frame involved in the embodiments of the present application is used to indicate the area clipped from the clipped image. In practical applications, the second intercepting frame may or may not exist.

[0062] Optionally, taking the regional image (such as called the first regional image) of a certain clipped image (such as called the first clipped image) as an example, the first regional image at least includes the pixel information of the middle row in the first clipped image. Assuming that the first clipped image includes 5 rows of pixels, the first regional image may include the 3rd row of pixels in the first clipped image, or include the 2nd - 4th rows of pixels in the first clipped image. The first clipped image may be any one of the N clipped images.

[0063] Optionally, the size and shape of the second intercepting frame may be default - set by the electronic device; or, the size and shape of the second intercepting frame may be set or modified by the electronic device according to user operations. The embodiments of the present application do not limit the specific size, specific shape, and setting method of the second intercepting frame. For example, the shape of the second intercepting frame may also be a square or other shapes.

[0064] S103: In the direction of the blood vessels in the straightened image, divide the straightened image into multiple sub - straightened images.

[0065] In the following, an example is given with the number of sub - straightened images being M, where M is a positive integer greater than or equal to 2.

[0066] Taking Figure 5 the shown straightened image as an example, as Figure 5 can be seen, the direction of the blood vessels in this straightened image is the vertical direction. Assuming M = 7, the electronic device divides the Figure 5 shown straightened image into 7 sub - straightened images in the vertical direction. The schematic diagram can be as Figure 7 shown.

[0067] Optionally, M can be less than N.

[0068] Optionally, the number of sub - straightened images (i.e., the value of M) may be default - set by the electronic device; or, the number of sub - straightened images may be set or modified by the electronic device according to user operations. The embodiments of the present application do not limit the specific number of sub - straightened images and the setting method.

[0069] S104: Input multiple data groups into the blood vessel radius prediction model, and predict the radius information of the target blood vessel; where one data group in the multiple data groups at least includes one sub - straightened image in the multiple sub - straightened images.

[0070] The following takes the number of data groups as K as an example for illustration, where K is an integer greater than or equal to 2. Optionally, the number of data groups can be less than or equal to the number of sub-straightened images, that is, K <= M.

[0071] The sub-straightened images in different data groups are different.

[0072] Based on the K sub-straightened images included in the input K data groups, the blood vessel radius prediction model can initially obtain partial radius information of the target blood vessel; based on the radius information of the front and back points on the blood vessel in the K sub-straightened images, the blood vessel radius prediction model can also predict the radius information of the middle points. Combining these two parts of radius information can obtain the radius information of the target blood vessel. Among them, the radius information of the front and back points can include the radius change trend information of the front and back points.

[0073] Taking any two adjacent sub-straightened images in the straightened image of the target blood vessel as an example, these two sub-straightened images are the first sub-straightened image and the second sub-straightened image respectively, that is, the M sub-straightened images include the first sub-straightened image and the second sub-straightened image. Taking the M sub-straightened images as Figure 7 shown as an example, the first sub-straightened image and the second sub-straightened image can be Figure 7 any two adjacent sub-straightened images in, for example, Figure 7 the two frontmost sub-straightened images in.

[0074] In a possible implementation manner, the blood vessel radius prediction model can predict the radius information of the third blood vessel based on the radius information of the first blood vessel and the radius information of the second blood vessel; among them, the first blood vessel is the blood vessel in the first sub-straightened image, the second blood vessel is the blood vessel in the second sub-straightened image, and the third blood vessel includes a section of blood vessel located between the first blood vessel and the second blood vessel; the radius information of the target blood vessel at least includes the radius information of the first blood vessel, the radius information of the second blood vessel, and the radius information of the third blood vessel.

[0075] In this way, the radius information of the second blood vessel can be predicted. Based on more adjacent sub-straightened images, sufficient radius information can be predicted, which can make the predicted radius information of the target blood vessel more accurate, thus facilitating obtaining an accurate blood vessel contour. In addition, it can also make the edge of the blood vessel contour smoother and the fitting effect better.

[0076] In a possible implementation manner, taking any one of the K data groups (such as called the first data group), and the sub-straightened image included in the first data group is called the third sub-straightened image as an example. The first data group can also include one or more of the following: the curvature information of the blood vessel in the third sub-straightened image, the position information of the third sub-straightened image in the straightened image. By inputting the curvature information of the blood vessel in the sub-straightened image into the blood vessel radius prediction model, it is beneficial to improve the prediction accuracy.

[0077] Optionally, the curvature information of the blood vessel in the third sub-straightened image may include: the sum of the vector components of multiple tangent vectors of the blood vessel in the third sub-straightened image in the horizontal axis direction, and the sum of the vector components of the multiple tangent vectors in the vertical axis direction.

[0078] Optionally, each of the M sub-straightened images may include multiple rows of pixel information, that is, the third sub-straightened image also includes multiple rows of pixel information. The multiple tangent vectors of the blood vessel in the third sub-straightened image may include: the tangent vectors of each row of pixels in the third sub-straightened image on the blood vessel in the third sub-straightened image, and one row of pixels corresponds to one tangent vector. There is a component (i.e., a vector) of a tangent vector in the horizontal axis direction. Similarly, there is also a component (i.e., a vector) of a tangent vector in the vertical axis direction.

[0079] The blood vessel radius prediction model is a deep learning network model, and the blood vessel radius prediction model can be a model based on the Vision Transformer (VIT) network model framework or other deep learning network structures. VIT is a deep learning network structure, and its main core is the self-attention mechanism.

[0080] The blood vessel radius prediction model is obtained after model training based on training data. Optionally, the training data may include: a straightened image training set and the label values corresponding to the straightened image training set. Among them, the straightened image training set includes straightened images of a large number of blood vessels. The label value corresponding to the straightened image of one blood vessel includes the distance from the left and right contours of the blood vessel to the center line of the blood vessel, or the label value corresponding to the straightened image of one blood vessel includes the radius of the left and right contours of each row in the straightened image.

[0081] Optionally, the electronic device may output the predicted radius information in the form of an image or text. Assuming that it is output in the form of an image, the electronic device may output the contour image of the target blood vessel. The specific value of the blood vessel radius may not be displayed in the contour image, but the predicted radius information can be implied through the contour image. Or, the specific value of the blood vessel radius may be displayed in the contour image. Or, when the electronic device detects a trigger operation for displaying the radius on the display interface of the contour image, the specific value of the blood vessel radius may be displayed on the display interface of the contour image. The embodiments of the present application do not limit the manner in which the electronic device outputs the predicted radius information.

[0082] By implementing the embodiments of the present application, on the one hand, the radius information of the midpoint can be accurately predicted, and the quantity of the obtained radius information is sufficient, so that the radius information of the target blood vessel predicted can be relatively accurate, which is conducive to obtaining an accurate blood vessel contour. On the second hand, the method of extracting the blood vessel contour by using the image segmentation technology has a poor effect on extracting the contour of the stenotic blood vessel. In contrast, the method adopted in the embodiments of the present application can obtain an accurate blood vessel contour through predicting the radius information of the blood vessel, and the method adopted in the embodiments of the present application also has an accurate prediction effect on the stenotic blood vessel. On the third hand, the method of extracting the blood vessel contour by using the image segmentation technology extracts a relatively rough contour, and the edge of the blood vessel contour is not smooth. In contrast, the method adopted in the embodiments of the present application can obtain an accurate, clear and smooth-edge blood vessel contour through predicting the radius information of the blood vessel. On the fourth hand, for the method of extracting the blood vessel contour by using the image segmentation technology, if the same image segmentation model is used for different types of blood vessels, the segmentation effect will be unsatisfactory. If different image segmentation models are used, higher-performance devices are required to meet the requirements. In contrast, the method adopted in the embodiments of the present application can use the same model (i.e., the blood vessel radius prediction model) to predict different types of blood vessels, which can reduce the requirements for device performance.

[0083] Optionally, the interval distance between two adjacent sampling points among the foregoing N sampling points is the same, which is conducive to making the straightened image closer to the effect of a real straightened blood vessel, and further conducive to improving the accuracy of subsequent prediction of the blood vessel radius.

[0084] Optionally, the interval distance between two adjacent sampling points among the N sampling points is less than or equal to the first distance, which is conducive to making the straightened image closer to the effect of a real straightened blood vessel, and further conducive to improving the accuracy of subsequent prediction of the blood vessel radius.

[0085] Optionally, N can be greater than or equal to the first value, so that more cropped images can be obtained, which is conducive to making the straightened image closer to the effect of a real straightened blood vessel, and further conducive to improving the accuracy of subsequent prediction of the blood vessel radius.

[0086] Optionally, the interval distance between two adjacent sampling points, the first distance, and the first value can be set by default by the electronic device; or, the interval distance between two adjacent sampling points, the first distance, and the first value can be set or modified by the electronic device according to the user operation. The embodiments of the present application do not limit the specific values and setting methods of the interval distance between two adjacent sampling points, the first distance, and the first value.

[0087] Next, taking one of the N sampling points (for example, called the first sampling point) as an example, another way to intercept a cropped image from the original angiography image of the target blood vessel based on the first sampling point is described. The first sampling point can be any one of the N sampling points.

[0088] In a possible implementation, the N sampling points may include the first sampling point, the N cropped images include the first cropped image, and the first cropped image is centered on the first sampling point. The process by which the electronic device intercepts the first cropped image from the original angiography image of the target blood vessel based on the first sampling point is as follows: The electronic device intercepts the first cropped image from the original angiography image of the target blood vessel based on the tangent and normal of the first sampling point on the center line of the target blood vessel, and the size information of the intercept box (or called the first intercept box); wherein, the intercept box is rectangular, and two adjacent sides of the intercept box are respectively parallel to the tangent and normal of the first sampling point on the center line. In this way, it is beneficial to straighten the image to be closer to the effect of a real straightened blood vessel, and thus beneficial to improve the accuracy of subsequent prediction of the blood vessel radius.

[0089] For example, one of the two adjacent sides of the intercept box is parallel to the tangent of the first sampling point on the center line, and the other side is parallel to the normal of the first sampling point on the center line; or, one of the two adjacent sides of the intercept box is parallel to the normal of the first sampling point on the center line, and the other side is parallel to the tangent of the first sampling point on the center line.

[0090] Taking N = 5, and the original angiography image of the target blood vessel is as Figure 3 shown as an example, the schematic diagram of intercepting 5 cropped images from this original angiography image is as Figure 8 shown. As Figure 8 shown, the dot represents the sampling point, the solid line with an arrow represents the tangent of a sampling point on the center line, the dashed line with an arrow represents the normal of a sampling point on the center line, and the dashed rectangular box represents the intercept box (i.e., the first intercept box). One sampling point corresponds to one intercept box, and one sampling point corresponds to one cropped image. As Figure 8 shown, based on 5 sampling points, 5 cropped images are obtained. Figure 8 Among them, the 5 cropped images are all placed with the tangent direction of their respective sampling points as the vertical direction.

[0091] In a possible implementation, after the electronic device obtains the N cropped images, it can splice the N cropped images according to the tangent directions of the sampling points corresponding to them in the center line to obtain the straightened image corresponding to the target blood vessel.

[0092] Taking Figure 8 the 5 cropped images in Figure 8 as an example, among them, the 5 cropped images are all placed with the tangent directions of their respective sampling points as the vertical direction. Therefore,Figure 8 The five cropped images shown are stitched vertically to obtain the straightened image of the target blood vessel, and this straightened image is as shown in Figure 9 . As can be seen from Figure 9 , this straightened image can indeed present the effect of the blood vessel after being straightened to a certain extent.

[0093] Alternatively, after the electronic device obtains N cropped images, it can stitch the regional images of each cropped image among the N cropped images according to the tangent direction of the sampling points corresponding to each cropped image in the center line, so as to obtain the straightened image corresponding to the target blood vessel. By re-cropping the cropped images and then stitching the regional images obtained after the re-cropping, this is conducive to making the straightened image closer to the effect of the real straightened blood vessel, and thus conducive to improving the accuracy of predicting the blood vessel radius in the subsequent process. Regarding the specific process of stitching the regional images of each cropped image among the N cropped images to obtain the straightened image, reference can be made to the specific description in S102, which will not be elaborated here.

[0094] Taking any one of the N cropped images (such as called the first cropped image), and the sampling point corresponding to the first cropped image in the center line being the first sampling point as an example, the regional image of the first cropped image can at least include: the pixel information of one row where the first sampling point is located in the first cropped image. Among them, the pixel of one row where the first sampling point is located in the first cropped image is the pixel of the middle row in the first cropped image. That is to say, assuming that the first cropped image includes 5 rows of pixels, the first regional image can include the 3rd row of pixels of the first cropped image, or include the 2nd - 4th rows of pixels of the first cropped image.

[0095] Next, the blood vessel radius prediction method proposed in this embodiment will be described in combination with the actual angiogram image of the target blood vessel.

[0096] Figure 10 is a schematic diagram of the actual angiogram image of the target blood vessel, Figure 11 is from Figure 10 a schematic diagram of intercepting cropped images from the original angiogram image shown. The specific intercepting method can be referred to the previous description, which will not be elaborated here. Figure 11 In, the rectangular frame represents the intercepting frame (i.e., the first intercepting frame), and the solid lines with arrows in the rectangular frame respectively represent the tangent and the normal of a sampling point on the center line, and the intersection point of the tangent and the normal represents the sampling point. It should be noted that for the sake of making the schematic process clearer and more concise, Figure 11 only 5 cropped images are shown for interception. Actually, more cropped images can be intercepted from the original angiogram image shown in Figure 11 .

[0097] The electronic device starts from Figure 11After cropping the original angiographic image as shown to obtain a cropped image, the straightened image corresponding to the target blood vessel obtained based on the cropped image can be as Figure 12 shown. Refer to Figure 12 It can be seen that the obtained straightened image presents the effect of a truly straightened blood vessel, which is beneficial to improving the accuracy of subsequent prediction of the blood vessel radius.

[0098] Among them, the straightened image can be obtained through the following process: The electronic device splices the obtained cropped images according to the tangent directions of the respective sampling points of the cropped images on the center line, or splices the regional images of the obtained cropped images to obtain the straightened image. The specific process can refer to the previous description and will not be elaborated here.

[0099] After the electronic device obtains the straightened image corresponding to the target blood vessel, by executing Figure 1 the S103 - S104 in the corresponding embodiment, the radius information of the target blood vessel can be obtained. Taking the output of the radius information of the target blood vessel in the form of an image as an example, the electronic device can output the contour image of the target blood vessel. The schematic diagram of the contour image of the target blood vessel can be as Figure 13 shown. For easy observation, the contour of the target blood vessel is marked with a wavy solid line in Figure 13 and it can be seen from Figure 13 that the contour image can more intuitively present the radius information of the target blood vessel.

[0100] Please refer to Figure 14 , Figure 14 which is a schematic structural diagram of a blood vessel radius prediction device provided in an embodiment of the present application. As Figure 14 shown, the blood vessel radius prediction device 140 includes an acquisition unit 1401, an image processing unit 1402, and a model prediction unit 1403. The blood vessel radius prediction device 140 can execute the relevant steps of the electronic device in the foregoing method embodiment.

[0101] The acquisition unit 1401 is configured to acquire the center line of the target blood vessel;

[0102] The image processing unit 1402 is configured to obtain the straightened image corresponding to the target blood vessel based on the center line and the original angiographic image of the target blood vessel; and divide the straightened image into a plurality of sub - straightened images in the blood vessel direction of the straightened image;

[0103] The model prediction unit 1403 is configured to input a plurality of data groups into the blood vessel radius prediction model to predict and obtain the radius information of the target blood vessel; wherein, one data group in the plurality of data groups includes at least one sub - straightened image among the plurality of sub - straightened images.

[0104] In a possible implementation, when the image processing unit 1402 is used to obtain a straightened image corresponding to the target blood vessel based on the center line and the original angiographic image of the target blood vessel, it is specifically used for: based on a plurality of sampling points on the center line, intercepting a plurality of cropped images from the original angiographic image of the target blood vessel, with different cropped images centered on different sampling points; and obtaining a straightened image corresponding to the target blood vessel based on the plurality of cropped images.

[0105] In a possible implementation, the plurality of sampling points includes a first sampling point, and the plurality of cropped images includes a first cropped image, with the first cropped image centered on the first sampling point; when the image processing unit 1402 is used to intercept a plurality of cropped images from the original angiographic image of the target blood vessel based on the plurality of sampling points on the center line, it is specifically used for: intercepting the first cropped image from the original angiographic image of the target blood vessel based on the tangent line and normal line of the first sampling point on the center line, and the size information of the interception frame; wherein, the interception frame is a rectangle, and two adjacent sides of the interception frame are respectively parallel to the tangent line and normal line of the first sampling point on the center line.

[0106] In a possible implementation, when the image processing unit 1402 is used to obtain a straightened image corresponding to the target blood vessel based on the plurality of cropped images, it is specifically used for: splicing the plurality of cropped images in the tangent direction of the respective sampling points corresponding to them in the center line to obtain a straightened image corresponding to the target blood vessel; or splicing the regional images of the respective cropped images in the plurality of cropped images in the tangent direction of the respective sampling points corresponding to them in the center line to obtain a straightened image corresponding to the target blood vessel, wherein, the regional image of the first cropped image includes: the pixel information of a row where the first sampling point is located in the first cropped image.

[0107] In a possible implementation, the plurality of sub-straightened images includes a first sub-straightened image and a second sub-straightened image, and the first sub-straightened image and the second sub-straightened image are adjacent sub-straightened images in the straightened image; the blood vessel radius prediction model is used for: predicting the radius information of a third blood vessel based on the radius information of a first blood vessel and the radius information of a second blood vessel; wherein, the first blood vessel is the blood vessel in the first sub-straightened image, the second blood vessel is the blood vessel in the second sub-straightened image, and the third blood vessel includes a section of blood vessel located between the first blood vessel and the second blood vessel; the radius information of the target blood vessel at least includes the radius information of the first blood vessel, the radius information of the second blood vessel, and the radius information of the third blood vessel.

[0108] In a possible implementation, the plurality of data groups includes a first data group, and the first data group includes a third sub-straightened image among the plurality of sub-straightened images; the first data group further includes one or more of the following: the curvature information of the blood vessel in the third sub-straightened image, and the position information of the third sub-straightened image in the straightened image.

[0109] In a possible implementation, the curvature information of the blood vessels in the third sub-straightened image includes: the sum of the vectors of multiple tangent vectors of the blood vessels in the third sub-straightened image in the horizontal axis direction, and the sum of the vectors of the multiple tangent vectors in the vertical axis direction.

[0110] Specifically, in this case, the operations performed by the acquisition unit 1401, the image processing unit 1402, and the model prediction unit 1403 can refer to Figures 1 - 13 the introduction of the electronic device in the corresponding embodiment.

[0111] Please refer to Figure 15 , Figure 15 which is another blood vessel radius prediction device 150 provided by the embodiments of the present application. It can be used to implement the functions of the electronic device in the above method embodiments. The blood vessel radius prediction device 150 may include a processor 1501. Optionally, the blood vessel radius prediction device may further include a memory 1502, and the memory 1502 is connected to the processor 1501. Among them, the processor 1501 and the memory 1502 may be connected through a bus 1503 or other means. The bus is Figure 15 shown in thick lines in Figure 15 For the connection manners between other components, only a schematic illustration is provided and is not limited thereto. The bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity,

[0112] In the embodiments of the present application, the coupling is an indirect coupling or communication connection between devices, units, or modules, which may be electrical, mechanical, or other forms, and is used for information interaction between devices, units, or modules. In the embodiments of the present application, the specific connection medium between the above-mentioned processor 1501 and the memory 1502 is not limited.

[0113] The memory 1502 may include a read-only memory and a random access memory, and provide instructions and data to the processor 1501. A part of the memory 1502 may further include a non-volatile random access memory.

[0114] The processor 1501 may be a Central Processing Unit (CPU), and the processor 1501 may also be 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. The general-purpose processor may be a microprocessor. Optionally, the processor 1501 may also be any conventional processor, etc.

[0115] In one example, when the electronic device adopts Figure 15 the form shown, Figure 15 the processor in it may execute the method performed by the electronic device in any of the above method embodiments.

[0116] In an alternative embodiment, a memory 1502 is used to store computer programs or instructions; a processor 1501 is used to call the computer programs or instructions stored in the memory 1502 for Figures 1 - 13 performing the steps executed by the electronic device in the corresponding embodiment.

[0117] Specifically, Figure 14 the functions / implementation processes of the acquisition unit 1401, the image processing unit 1402, and the model prediction unit 1403 in it can all be implemented by Figure 15 the processor 1501 in it calling the computer programs or instructions stored in the memory 1502.

[0118] In the embodiments of the present application, the method provided in the embodiments of the present application can be implemented by running a computer program (including program code) capable of executing the steps involved in the above method on a general computing device such as a computer including processing elements and storage elements such as a CPU, a Random Access Memory (RAM), and a Read-Only Memory (ROM). The computer program may be recorded on a computer-readable recording medium, loaded into the above computing device through the computer-readable recording medium, and run therein.

[0119] Based on the same inventive concept, the principle of solving problems and the beneficial effects of the blood vessel radius prediction device 150 provided in the embodiments of the present application are similar to the principle of solving problems and the beneficial effects of the electronic device in the method embodiments of the present application. The principle and beneficial effects of the method implementation can be referred to. For the sake of brevity, they will not be described here again.

[0120] The embodiments of the present application further provide a computer-readable storage medium, in which a computer program or computer instructions are stored, and the computer program or computer instructions are adapted to be loaded and executed by a computer to perform the method provided by the above method embodiments.

[0121] The embodiments of the present application further provide a computer program product including a computer program or instructions. When the computer program or instructions run on a computer, the computer is enabled to perform the method provided by the above method embodiments.

[0122] Regarding each device and product described in the above embodiments, each module / unit included therein can be a software module / unit, a hardware module / unit, or can be partially a software module / unit and partially a hardware module / unit. For example, for each device and product applied to or integrated into a chip, each module / unit included therein can be implemented in a hardware manner such as a circuit. Or, at least some modules / units can be implemented in a software program manner, and the software program runs on a processor integrated inside the chip, and the remaining (if any) part of the modules / units can be implemented in a hardware manner such as a circuit; for each device and product applied to or integrated into a chip module, each module / unit included therein can be implemented in a hardware manner such as a circuit. Different modules / units can be located in the same component (such as a chip, a circuit module, etc.) or different components of the chip module. Or, at least some modules / units can be implemented in a software program manner, and the software program runs on a processor integrated inside the chip module, and the remaining (if any) part of the modules / units can be implemented in a hardware manner such as a circuit; for each device and product applied to or integrated into an electronic device, each module / unit included therein can be implemented in a hardware manner such as a circuit. Different modules / units can be located in the same component (such as a chip, a circuit module, etc.) or different components inside the electronic device. Or, at least some modules / units can be implemented in a software program manner, and the software program runs on a processor integrated inside the electronic device, and the remaining (if any) part of the modules / units can be implemented in a hardware manner such as a circuit.

[0123] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, some steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present application.

[0124] In the above embodiments, the descriptions of the respective embodiments each have their own emphasis, and any plurality of embodiments can be combined for use. For parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0125] The steps in the method of the embodiments of the present application can be adjusted, combined, and deleted according to actual needs.

[0126] The modules in the device of the embodiments of the present application can be combined, divided, and deleted according to actual needs.

[0127] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by program instructions and related hardware. The program instructions can be stored in a computer-readable storage medium, and the computer-readable storage medium can include: a flash drive, ROM, RAM, a magnetic disk, or an optical disc, etc.

[0128] What is disclosed above is only one embodiment of the present application, and it is only a part of the embodiments of the present application. It cannot be used to limit the scope of the rights of the present application.

Claims

1. A method for predicting blood vessel radius, characterized in that, The method includes: Obtaining the centerline of the target blood vessel; Based on the centerline and the original angiographic image of the target blood vessel, obtaining a straightened image corresponding to the target blood vessel; Dividing the straightened image into a plurality of sub-straightened images in the blood vessel direction in the straightened image; Inputting a plurality of data groups into a blood vessel radius prediction model to predict the radius information of the target blood vessel; wherein, one data group in the plurality of data groups includes at least one sub-straightened image among the plurality of sub-straightened images.

2. The method according to claim 1, wherein The obtaining a straightened image corresponding to the target blood vessel based on the centerline and the original angiographic image of the target blood vessel includes: Based on a plurality of sampling points on the centerline, intercepting a plurality of cropped images from the original angiographic image of the target blood vessel, with different cropped images centered on different sampling points; Based on the plurality of cropped images, obtaining a straightened image corresponding to the target blood vessel.

3. The method according to claim 2, wherein The plurality of sampling points include a first sampling point, and the plurality of cropped images include a first cropped image, with the first cropped image centered on the first sampling point; The intercepting a plurality of cropped images from the original angiographic image of the target blood vessel based on the plurality of sampling points on the centerline includes: Based on the tangent line, normal line of the first sampling point on the centerline, and the size information of the intercepting frame, intercepting the first cropped image from the original angiographic image of the target blood vessel; wherein, the intercepting frame is a rectangle, and two adjacent sides of the intercepting frame are respectively parallel to the tangent line and the normal line of the first sampling point on the centerline.

4. The method according to claim 3, characterized in that, The obtaining a straightened image corresponding to the target blood vessel based on the plurality of cropped images includes: Stitching the plurality of cropped images according to the tangent direction of the sampling points corresponding to them in the centerline to obtain a straightened image corresponding to the target blood vessel; or, Stitching the regional images of each of the plurality of cropped images according to the tangent direction of the sampling points corresponding to them in the centerline to obtain a straightened image corresponding to the target blood vessel, wherein the regional image of the first cropped image includes: the pixel information of the row where the first sampling point is located in the first cropped image.

5. The method according to any one of claims 1 to 4, characterized in that The plurality of sub-straightened images include a first sub-straightened image and a second sub-straightened image, and the first sub-straightened image and the second sub-straightened image are adjacent sub-straightened images in the straightened image; The blood vessel radius prediction model is used for: predicting the radius information of a third blood vessel based on the radius information of a first blood vessel and the radius information of a second blood vessel; Wherein, the first blood vessel is the blood vessel in the first sub-straightened image, the second blood vessel is the blood vessel in the second sub-straightened image, and the third blood vessel includes a section of blood vessel located between the first blood vessel and the second blood vessel; the radius information of the target blood vessel includes at least the radius information of the first blood vessel, the radius information of the second blood vessel, and the radius information of the third blood vessel.

6. The method according to any one of claims 1 to 4, characterized in that The plurality of data groups include a first data group, and the first data group includes a third sub-straightened image among the plurality of sub-straightened images; The first data set further includes one or more of the following: curvature information of blood vessels in the third sub-straightened image, and position information of the third sub-straightened image in the straightened image.

7. The method according to claim 6, wherein The curvature information of blood vessels in the third sub-straightened image includes: the sum of vectors of multiple tangent vectors of blood vessels in the third sub-straightened image in the horizontal axis direction, and the sum of vectors of the multiple tangent vectors in the vertical axis direction.

8. A blood vessel radius prediction device, characterized in that, It includes a unit for executing the method according to any one of claims 1-7.

9. A blood vessel radius prediction device, characterized in that, It includes a memory and a processor, the memory is connected to the processor; wherein, the memory is used to store computer programs or instructions, and the processor is used to execute the computer programs or instructions in the memory. When the computer programs or instructions are executed by the processor, the device is caused to execute the method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program or computer instructions. When the computer program or computer instructions are executed by a processor, the computer is caused to execute the method according to any one of claims 1-7.

Images