Blood vessel segmentation method, device, computer equipment and storage medium

Abstract

The present application relates to a blood vessel segmentation method, apparatus, computer equipment, storage medium and computer program product. The method comprises: obtaining a blood vessel segmentation result and a first arteriovenous segmentation result of a tissue of interest in a medical image, and determining the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result, and then processing the adhesion area in the medical image based on the arteriovenous adhesion position to obtain a first processing result, and determining the target arteriovenous segmentation result of the tissue of interest based on the first processing result and the first arteriovenous segmentation result. In this way, the adhesion area in the medical image is optimized based on the arteriovenous adhesion position of the tissue of interest to obtain the first processing result, and then the final target arteriovenous segmentation result of the tissue of interest is determined based on the first processing result and the first arteriovenous segmentation result, thereby improving the accuracy of the final target arteriovenous segmentation result.

Description

Technical Field

[0001] The present application relates to the technical field of medical image segmentation, and in particular to a blood vessel segmentation method, apparatus, computer equipment, storage medium and computer program product. Background Art

[0002] Segmenting the blood vessels of tissues in medical images helps doctors analyze tissue diseases and plan surgeries, thereby avoiding damage to blood vessels and bleeding and reducing treatment risks for patients.

[0003] In traditional technology, region growing technology is used, for example, arteriovenous blood vessels are taken as the starting point from their respective trunks, and region growing is performed using the continuity of grayscale and the first-come-first-served rule. Among them, first-come-first-served means that if the pixel point in the blood vessel is first merged into the arterial blood vessel area, then the pixel point is regarded as a pixel point in the arterial blood vessel, and if the pixel point is first merged into the venous blood vessel area, then the pixel point is regarded as a pixel point in the venous blood vessel.

[0004] However, the grayscale boundaries of arteries and veins at the adhesion sites are weak or even nonexistent, which can easily lead to regional growing errors, thus causing arteriovenous segmentation errors. Summary of the invention

[0005] Based on this, it is necessary to provide a blood vessel segmentation method, apparatus, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of arteriovenous segmentation results in response to the above technical problems.

[0006] In a first aspect, the present application provides a blood vessel segmentation method. The method comprises:

[0007] Obtaining a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image;

[0008] Determining the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result;

[0009] Based on the arteriovenous adhesion position, processing the adhesion area in the medical image to obtain a first processing result;

[0010] Based on the first processing result and the first artery and vein segmentation result, a target artery and vein segmentation result of the tissue of interest is determined.

[0011] In one embodiment, the processing of the adhesion region in the medical image based on the arteriovenous adhesion position to obtain a first processing result includes:

[0012] Cropping the medical image based on the arteriovenous adhesion position to obtain a local image corresponding to the adhesion area;

[0013] The local image is input into a first model to obtain the first processing result corresponding to the local image.

[0014] In one embodiment, determining the target artery and vein segmentation result of the tissue of interest based on the first processing result and the first artery and vein segmentation result includes:

[0015] Using the first processing result to replace the area corresponding to the local image in the first artery and vein segmentation result, to obtain a second artery and vein segmentation result;

[0016] A connected domain in the second artery and vein segmentation result is obtained, and a target artery and vein segmentation result of the tissue of interest is determined according to the connected domain in the second artery and vein segmentation result.

[0017] In one embodiment, determining the target artery and vein segmentation result of the tissue of interest according to the connected domain in the second artery and vein segmentation result includes:

[0018] If a first connected domain of the first type exists in the second artery and vein segmentation result, and the first connected domain is connected to multiple second connected domains of the second type, the type of the first connected domain is modified to the second type to obtain a target artery and vein segmentation result of the tissue of interest.

[0019] In one of the embodiments, if there is a third connected domain that is not connected to the maximum connected domain in the target artery and vein segmentation result, the third connected domain is removed.

[0020] In one of the embodiments, the shortest distance from the foreground point to the background point on the blood vessel segmentation result is determined to obtain a blood vessel distance map;

[0021] Determine the shortest distance between the foreground point and the background point on the first artery and vein segmentation result to obtain an artery and vein distance map;

[0022] The blood vessel distance map and the arteriovenous distance map are input into the second model to obtain the arteriovenous adhesion position of the tissue of interest.

[0023] In one embodiment, the medical image is input into a third model to obtain a segmented image, wherein the segmented image includes the tissue of interest;

[0024] Based on the medical image and the segmented image, a blood vessel segmentation result of the tissue of interest and the first artery and vein segmentation result are obtained.

[0025] In a second aspect, the present application also provides a blood vessel segmentation device. The device comprises:

[0026] An acquisition module, used for acquiring a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image;

[0027] A first determination module is used to determine the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result;

[0028] A processing module, configured to process the adhesion region in the medical image based on the arteriovenous adhesion position to obtain a first processing result;

[0029] The second determination module is used to determine the target artery and vein segmentation result of the tissue of interest based on the first processing result and the first artery and vein segmentation result.

[0030] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of any of the above methods when executing the computer program:

[0031] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of any of the above methods when executed by a processor.

[0032] In a fifth aspect, the present application further provides a computer program product, wherein the computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0033] The above-mentioned blood vessel segmentation method, apparatus, computer equipment, storage medium and computer program product obtain the blood vessel segmentation result and the first arteriovenous segmentation result of the tissue of interest in the medical image, and determine the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result, and then process the adhesion area in the medical image based on the arteriovenous adhesion position to obtain a first processing result, and determine the target arteriovenous segmentation result of the tissue of interest based on the first processing result and the first arteriovenous segmentation result. In the conventional technology, the first arteriovenous segmentation result, that is, the initial arteriovenous segmentation result, is directly used as the final arteriovenous segmentation result, and the first arteriovenous segmentation result has the situation of incorrect segmentation of the arteriovenous adhesion area. In this embodiment, based on the arteriovenous adhesion position of the tissue of interest, the adhesion area in the medical image is optimized to obtain the first processing result. Since the first processing result is the result obtained after re-optimization of the adhesion area, the first processing result corresponding to the adhesion area is more accurate than the first arteriovenous segmentation result corresponding to the adhesion area, and then based on the first processing result and the first arteriovenous segmentation result, the final target arteriovenous segmentation result of the tissue of interest is determined, thereby improving the accuracy of the final target arteriovenous segmentation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of a blood vessel segmentation method provided in an embodiment of the present application;

[0035] Figure 2 It is a flowchart of a method for obtaining a first processing result provided in an embodiment of the present application;

[0036] Figure 3 It is a schematic diagram of the structure of a neural network provided in an embodiment of the present application;

[0037] Figure 4 It is a flowchart of a method for determining target artery and vein segmentation results provided in an embodiment of the present application;

[0038] Figure 5 It is a flow chart of a method for determining the position of arteriovenous adhesion provided in an embodiment of the present application;

[0039] Figure 6 is a schematic diagram of a blood vessel distance map provided in an embodiment of the present application;

[0040] Figure 7 This is a schematic diagram of locating the position of arteriovenous adhesion provided in an embodiment of the present application;

[0041] Figure 8 It is a flowchart of a method for obtaining a blood vessel segmentation result and a first artery and vein segmentation result provided in an embodiment of the present application;

[0042] Fig. 9 is a structural schematic diagram of a blood vessel segmentation device provided in an embodiment of the present application;

[0043] Fig.10 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0045] In one embodiment, Figure 1 As shown, Figure 1 : is a flow chart of a blood vessel segmentation method provided in an embodiment of the present application. The method can be applied to a computer device. The method includes the following steps:

[0046] S101 , obtaining a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image.

[0047] In this embodiment, a region growing method may be used to obtain a first arteriovenous segmentation result of a tissue of interest in a medical image, or a neural network may be used to obtain a first arteriovenous segmentation result of a tissue of interest. The first arteriovenous segmentation result refers to an initial arteriovenous segmentation result obtained. The blood vessel segmentation result may be a blood vessel mask, and the first arteriovenous segmentation result may be a segmented initial arteriovenous mask.

[0048] Medical images can be images acquired by an imaging device such as magnetic resonance imaging (MR), computed tomography (CT), single photon emission computed tomography (SPECT), positron emission tomography (PET), ultrasound (US) and X-ray (XR).

[0049] The tissue of interest may be lung, heart, liver, etc. A deep learning method may be used to segment the blood vessel segmentation result and the first arteriovenous segmentation result of the tissue of interest, and the first arteriovenous segmentation result may be a result obtained based on the blood vessel segmentation result and the medical image.

[0050] S102 : Determine the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result.

[0051] In this embodiment, based on the blood vessel segmentation result and the first artery and vein segmentation result, a deep learning method can be used to determine the arteriovenous adhesion position of the tissue of interest, wherein the arteriovenous adhesion position refers to the adhesion position of the arteriovenous blood vessels.

[0052] S103: Based on the arteriovenous adhesion position, the adhesion area in the medical image is processed to obtain a first processing result.

[0053] In this embodiment, the first processing result may be a local mask obtained by optimizing the local image corresponding to the adhesion area in the medical image.

[0054] Among them, based on the position of arteriovenous adhesion, a deep learning method can be used to process the adhesion area in the medical image to obtain a first processing result.

[0055] S104 : Determine a target artery and vein segmentation result of the tissue of interest based on the first processing result and the first artery and vein segmentation result.

[0056] Wherein, based on the first processing result and the first arteriovenous segmentation result, determining the target arteriovenous segmentation result of the tissue of interest can be achieved in the following manner:

[0057] The first processing result is used to replace the area corresponding to the local image in the first arteriovenous segmentation result to obtain a second arteriovenous segmentation result, and the second arteriovenous segmentation result is used as the target arteriovenous segmentation result of the tissue of interest.

[0058] The blood vessel segmentation method provided in this embodiment obtains the blood vessel segmentation result and the first arteriovenous segmentation result of the tissue of interest in the medical image, and determines the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result, and then processes the adhesion area in the medical image based on the arteriovenous adhesion position to obtain a first processing result, and determines the target arteriovenous segmentation result of the tissue of interest based on the first processing result and the first arteriovenous segmentation result. In the conventional technology, the first arteriovenous segmentation result, that is, the initial arteriovenous segmentation result, is directly used as the final arteriovenous segmentation result, and the first arteriovenous segmentation result has the situation of incorrect segmentation of the arteriovenous adhesion area. In this embodiment, based on the arteriovenous adhesion position of the tissue of interest, the adhesion area in the medical image is optimized to obtain the first processing result. Since the first processing result is the result obtained after re-optimization of the adhesion area, the first processing result corresponding to the adhesion area is more accurate than the first arteriovenous segmentation result corresponding to the adhesion area, and then based on the first processing result and the first arteriovenous segmentation result, the final target arteriovenous segmentation result of the tissue of interest is determined, thereby improving the accuracy of the final target arteriovenous segmentation result.

[0059] Reference Figure 2 , Figure 2 1 is a flow chart of a method for obtaining a first processing result provided in an embodiment of the present application. This embodiment relates to an optional implementation method of how to process the adhesion area in a medical image based on the arteriovenous adhesion position to obtain a first processing result. Based on the above embodiment, the above S103 may include the following steps:

[0060] S201 . Crop the medical image based on the position of arteriovenous adhesion to obtain a local image corresponding to the adhesion area.

[0061] There may be one or more arteriovenous adhesion positions. If multiple arteriovenous adhesion positions are determined, each arteriovenous adhesion position corresponds to an adhesion region. The adhesion region may be a region with the arteriovenous adhesion position point as the center point and a length and width of corresponding preset values, so as to determine the specific position of the adhesion region. After determining the specific position of the adhesion region, the adhesion region is cropped out from the medical image to obtain a local image corresponding to the adhesion region.

[0062] S202: Input the local image into the first model to obtain a first processing result corresponding to the local image.

[0063] Among them, the first model can be a segmentation model based on convolutional neural networks (CNNs), such as VB-Net neural networks, UNet, VNet, FCN, SegNet and DeepLab networks that have been trained to convergence. Among them, the FCN network is a semantic segmentation (Fully Convolutional Networks for Semantic Segmentation) network based on a fully convolutional network. The SegNet network is a deep network proposed by Cambridge to solve the image semantic segmentation problem of autonomous driving or intelligent robots. The SegNet network is based on FCN and is a semantic segmentation network obtained by modifying the VGG-16 network. There are two versions of SegNet, namely SegNet and Bayesian SegNet. The DeepLab network is a method that combines deep convolutional neural networks and probabilistic graph models. Among them, the network structure of the VB-Net neural network is as follows: Figure 3 As shown, Figure 3 VB-Net is a schematic diagram of a neural network structure provided in an embodiment of the present application. The main difference of VB-Net is that the bottleneck layer (bottleneck structure) is used to replace the conventional convolutional layer in upsampling and downsampling. The use of VB-Net neural network can reduce the number of model parameters and speed up model training.

[0064] In the stage of training the initial network corresponding to the first model, the parameters of the fourth model described below can be used as the initialization parameters of the initial network to perform transfer learning to obtain the first model trained to convergence. In the training stage, a strategy of balancing positive and negative samples can be adopted to randomly select samples from the candidate adhesion positions and non-adhesion positions of the initial medical samples with a probability of 1:1 for learning to obtain the first model.

[0065] In this embodiment, the medical image is cropped based on the position of arteriovenous adhesion to obtain a local image corresponding to the adhesion area, and the local image is input into the first model to obtain a first processing result corresponding to the local image. Since the first model is a trained model, it can accurately segment the arteries and veins in the adhesion area of ​​the local image. Therefore, inputting the local image into the first model can obtain an accurate first processing result, and then based on the accurate first processing result and the first arteriovenous segmentation result, a more accurate target arteriovenous segmentation result can be determined.

[0066] Reference Figure 4 , Figure 4 1 is a flow chart of a method for determining a target arteriovenous segmentation result provided in an embodiment of the present application. This embodiment relates to an optional implementation method of how to determine a target arteriovenous segmentation result of a tissue of interest based on a first processing result and a first arteriovenous segmentation result. Based on the above embodiment, the above S104 may include the following steps:

[0067] S401: Use the first processing result to replace the area corresponding to the local image in the first artery and vein segmentation result to obtain a second artery and vein segmentation result.

[0068] Since the first processing result is an optimization result of the first arteriovenous segmentation result corresponding to the adhesion area, the arteriovenous segmentation in the first processing result is more accurate. Therefore, the first processing result is used to replace the area corresponding to the local image in the first arteriovenous segmentation result, thereby optimizing the area corresponding to the local image in the first arteriovenous segmentation result, so that the second arteriovenous segmentation result is more accurate.

[0069] S402: Acquire a connected domain in the second artery and vein segmentation result, and determine a target artery and vein segmentation result of the tissue of interest according to the connected domain in the second artery and vein segmentation result.

[0070] In this embodiment, the connected domain in the second arteriovenous segmentation result includes a venous connected domain and an arterial connected domain. If the first connected domain of the first type in the second arteriovenous segmentation result is not connected to multiple second connected domains of the second type, the second arteriovenous segmentation result can be directly used as the target arteriovenous segmentation result of the tissue of interest. The first type and the second type are different types. For example, if the first type is an arterial type, the second type is a venous type; if the first type is a venous type, the second type is an arterial type.

[0071] The method provided in this embodiment obtains a second arteriovenous segmentation result by replacing the area corresponding to the local image in the first arteriovenous segmentation result with the first processing result, thereby optimizing the area corresponding to the local image in the first arteriovenous segmentation result, making the optimized second arteriovenous segmentation result more accurate, and further making the target arteriovenous segmentation result determined based on the connected domain in the accurate second arteriovenous segmentation result more accurate.

[0072] In some embodiments, determining the target artery and vein segmentation result of the tissue of interest according to the connected domain in the second artery and vein segmentation result in S402 above can be achieved in the following manner:

[0073] If there is a first connected domain of the first type in the second arteriovenous segmentation result, and the first connected domain is connected to multiple second connected domains of the second type, the type of the first connected domain is modified to the second type to obtain a target arteriovenous segmentation result of the tissue of interest.

[0074] For example, the second arteriovenous segmentation result includes 10 connected domains, connected domains 1 to 5 are connected domains of the vein type, and connected domains 6 to 10 are connected domains of the artery type. If connected domain 1 is connected to connected domains 6 to 10, the type of connected domain 1 is changed from the vein type to the artery type, thereby correcting the incorrectly segmented connected domains in the second arteriovenous segmentation result, further improving the accuracy of the final target arteriovenous segmentation result.

[0075] In some embodiments, the following steps may also be included:

[0076] If there is a third connected domain that is not connected to the maximum connected domain in the target artery and vein segmentation result, the third connected domain is removed.

[0077] The maximum connected domain includes the maximum arterial connected domain and the maximum venous connected domain in the target arteriovenous segmentation result.

[0078] In this embodiment, if there is a third connected domain that is not connected to the maximum connected domain in the target arteriovenous segmentation result, the third connected domain is removed. In combination with the above example, if the connected domain 10 is not connected to both the maximum arterial connected domain and the maximum venous connected domain, it means that the connected domain 10 is an incorrectly segmented connected domain. Therefore, in this embodiment, the connected domain 10 is removed to remove the incorrectly segmented connected domain, so that the target arteriovenous segmentation result after removing the third connected domain that is not connected to the maximum connected domain is more accurate.

[0079] Reference Figure 5 , Figure 5 : is a flow chart of a method for determining the location of arteriovenous adhesions provided in an embodiment of the present application. This embodiment relates to an optional implementation method of how to determine the location of arteriovenous adhesions of a tissue of interest based on a blood vessel segmentation result and a first arteriovenous segmentation result. Based on the above embodiment, the above S102 may include the following steps:

[0080] S501 , determining the shortest distance from the foreground point to the background point on the blood vessel segmentation result, and obtaining a blood vessel distance map.

[0081] Among them, the foreground point refers to the three-dimensional voxel point of the blood vessel in the blood vessel segmentation result. The blood vessel distance map can be obtained by the distance transformation method, which is the shortest distance from the foreground point to the background point. Figure 6 As shown, Figure 6 It is a schematic diagram of a blood vessel distance map provided in an embodiment of the present application.

[0082] S502: Determine the shortest distance from the foreground point to the background point on the first artery and vein segmentation result to obtain an artery and vein distance map.

[0083] The foreground points in this step may be arteriovenous points on the arteriovenous mask, and the shortest distance from the foreground points to the background points on the first arteriovenous segmentation result may be calculated by distance transformation to obtain an arteriovenous distance map.

[0084] S503: Input the blood vessel distance map and the arteriovenous distance map into the second model to obtain the arteriovenous adhesion position of the tissue of interest.

[0085] Among them, the second model can be a segmentation model based on a convolutional neural network, such as a VB-Net neural network, UNet, VNet, FCN, SegNet, and DeepLab network, which has been trained to convergence. In the stage of training the initial network corresponding to the second model, the initialization parameters of the initial network can be the parameters of the fourth model described below for outputting the vascular segmentation results. The parameters of the fourth model are used as the initialization parameters of the initial network to perform transfer learning and obtain the second model trained to convergence. The input of the second model is two channels, one of which is used to input a vascular distance map, and the other channel is used to input an arteriovenous distance map. The physical resolution is 0.5 mm.

[0086] like Figure 7 As shown, Figure 7 Schematic diagram of the location of arteriovenous adhesion provided in the embodiment of the present application. Figure 7 As shown, Figure 7 The cross-shaped position shown in the figure is the confirmed location of arteriovenous adhesion.

[0087] In this embodiment, a blood vessel distance map is obtained by determining the shortest distance from the foreground point to the background point on the blood vessel segmentation result, and an arteriovenous distance map is obtained by determining the shortest distance from the foreground point to the background point on the first arteriovenous segmentation result. Then, the blood vessel distance map and the arteriovenous distance map are input into the second model to obtain the arteriovenous adhesion position of the tissue of interest, so that the arteriovenous adhesion position of the tissue of interest can be accurately determined, and the adhesion area in the medical image is optimized based on the arteriovenous adhesion position to obtain the first processing result, and then based on the first processing result and the first arteriovenous segmentation result, the final target arteriovenous segmentation result of the tissue of interest is determined, thereby improving the accuracy of the final target arteriovenous segmentation result.

[0088] Reference Figure 8 , Figure 8 1 is a flowchart of a method for obtaining a blood vessel segmentation result and a first arteriovenous segmentation result provided in an embodiment of the present application. This embodiment relates to an optional implementation method of how to obtain a blood vessel segmentation result and a first arteriovenous segmentation result of a tissue of interest in a medical image. Based on the above embodiment, the above S101 may include the following steps:

[0089] S801. Input a medical image into a third model to obtain a segmented image, wherein the segmented image includes a tissue of interest.

[0090] Wherein, S801, inputting the medical image into the third model to obtain the segmented image can be achieved in the following manner:

[0091] According to the preset window width and window position, the image data of the medical image is normalized to obtain a normalized medical image; the normalized medical image is input into the third model to obtain a segmented image.

[0092] Alternatively, the medical image is not normalized and is directly input into the third model to obtain a segmented image.

[0093] It should be noted that the third model can be a segmentation model based on a convolutional neural network such as a VB-Net neural network, UNet, VNet, FCN, SegNet, and DeepLab network that has been trained to convergence. When the third model is used to segment the segmented image of the tissue of interest, the physical sampling resolution can be 5 mm.

[0094] The image data in the embodiment of the present application refers to the image pixel value. The window position is the center position of the window, the window width refers to the width of the window, the window position is 40, the window width is 800, and the image data of the medical image is normalized according to the window width and the window position. Among them, when the window position is 40 and the window width is 800, when the image pixel value is normalized, the pixel value less than -360 is adjusted to 0, and the pixel value greater than 440 is adjusted to 1.

[0095] S802 : Obtain a blood vessel segmentation result of a tissue of interest and a first artery and vein segmentation result based on the medical image and the segmented image.

[0096] Wherein, S802, obtaining a blood vessel segmentation result and a first artery and vein segmentation result based on the medical image and the segmented image, can be implemented in the following manner:

[0097] The medical image and the segmented image are input into the fourth model to obtain a blood vessel segmentation result; the medical image and the blood vessel segmentation result are input into the fifth model to obtain a first arteriovenous segmentation result.

[0098] It should be noted that the fourth model and the fifth model can be segmentation models based on convolutional neural networks such as VB-Net neural network, UNet, VNet, FCN, SegNet and DeepLab network that have been trained to convergence. When the fourth model is used to obtain the blood vessel segmentation result, the physical sampling resolution can be 0.5mm.

[0099] Among them, the fifth model can be a model obtained after training a preset initial network. During the training stage, the initial parameters of the initial network can select the network parameters of the fourth model for transfer learning. The network optimizer, loss function, etc. are the same as the fourth model.

[0100] In this embodiment, by inputting the medical image into the third model, a segmented image including the tissue of interest is obtained, and based on the medical image and the segmented image, a blood vessel segmentation result and a first artery and vein segmentation result of the tissue of interest are obtained, thereby ensuring the accuracy of the obtained blood vessel segmentation result and the first artery and vein segmentation result.

[0101] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0102] Based on the same inventive concept, the embodiment of the present application also provides a blood vessel segmentation device for implementing the blood vessel segmentation method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more blood vessel segmentation device embodiments provided below can refer to the limitations on the blood vessel segmentation method above, and will not be repeated here.

[0103] In one embodiment, Fig. 9 As shown, Fig. 9 is a schematic diagram of the structure of a blood vessel segmentation device provided in an embodiment of the present application. The device 900 is deployed in a computer device, and the device 900 includes:

[0104] The acquisition module 901 is used to acquire a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image.

[0105] The first determination module 902 is used to determine the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result.

[0106] The processing module 903 is used to process the adhesion area in the medical image based on the arteriovenous adhesion position to obtain a first processing result.

[0107] The second determination module 904 is used to determine the target artery and vein segmentation result of the tissue of interest based on the first processing result and the first artery and vein segmentation result.

[0108] The blood vessel segmentation device provided in this embodiment obtains the blood vessel segmentation result and the first arteriovenous segmentation result of the tissue of interest in the medical image, and determines the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result, and then processes the adhesion area in the medical image based on the arteriovenous adhesion position to obtain a first processing result, and determines the target arteriovenous segmentation result of the tissue of interest based on the first processing result and the first arteriovenous segmentation result. In the conventional technology, the first arteriovenous segmentation result, that is, the initial arteriovenous segmentation result, is directly used as the final arteriovenous segmentation result, and the first arteriovenous segmentation result has the situation of incorrect segmentation of the adhesion area of ​​the arteriovenous. In this embodiment, based on the arteriovenous adhesion position of the tissue of interest, the adhesion area in the medical image is optimized to obtain the first processing result. Since the first processing result is the result obtained after re-optimization of the adhesion area, the first processing result corresponding to the adhesion area is more accurate than the first arteriovenous segmentation result corresponding to the adhesion area, and then based on the first processing result and the first arteriovenous segmentation result, the final target arteriovenous segmentation result of the tissue of interest is determined, thereby improving the accuracy of the final target arteriovenous segmentation result.

[0109] In some embodiments, the processing module 903 includes:

[0110] A cropping unit, used for cropping the medical image based on the arteriovenous adhesion position to obtain a local image corresponding to the adhesion area;

[0111] The first obtaining unit is used to input the local image into the first model to obtain a first processing result corresponding to the local image.

[0112] In some embodiments, the second determining module 904 includes:

[0113] A second obtaining unit is used to replace the area corresponding to the local image in the first arteriovenous segmentation result with the first processing result to obtain a second arteriovenous segmentation result;

[0114] The first determination unit is used to obtain a connected domain in the second artery and vein segmentation result, and determine a target artery and vein segmentation result of the tissue of interest according to the connected domain in the second artery and vein segmentation result.

[0115] In some embodiments, the first determination unit is specifically used to modify the type of the first connected domain to the second type if there is a first connected domain of the first type in the second artery and vein segmentation result, and the first connected domain is connected to multiple second connected domains of the second type, so as to obtain the target artery and vein segmentation result of the tissue of interest.

[0116] In some embodiments, the apparatus 900 may further include:

[0117] The removal module is used to remove the third connected domain if there is a third connected domain that is not connected to the maximum connected domain in the target artery and vein segmentation result.

[0118] In some embodiments, the first determining module 902 includes:

[0119] A second determination unit is used to determine the shortest distance from the foreground point to the background point on the blood vessel segmentation result to obtain a blood vessel distance map;

[0120] A third determining unit is used to determine the shortest distance from the foreground point to the background point on the first artery and vein segmentation result to obtain an artery and vein distance map;

[0121] The third obtaining unit is used to input the blood vessel distance map and the arteriovenous distance map into the second model to obtain the arteriovenous adhesion position of the tissue of interest.

[0122] In some embodiments, the acquisition module 901 includes:

[0123] A segmentation unit, used for inputting the medical image into the third model to obtain a segmented image, wherein the segmented image includes the tissue of interest;

[0124] The fourth obtaining unit is used to obtain a blood vessel segmentation result and a first artery and vein segmentation result based on the medical image and the segmented image.

[0125] Each module in the above-mentioned blood vessel segmentation device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0126] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig.10 As shown, Fig.10 1 is an internal structure diagram of a computer device in an embodiment. The computer device includes a processor, a memory and a network interface connected through a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store medical image data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a blood vessel segmentation method is implemented.

[0127] Those skilled in the art will understand that Fig.10The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0128] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0129] Obtaining a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image;

[0130] Determining the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result;

[0131] Based on the arteriovenous adhesion position, processing the adhesion area in the medical image to obtain a first processing result;

[0132] Based on the first processing result and the first artery and vein segmentation result, a target artery and vein segmentation result of the tissue of interest is determined.

[0133] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0134] Cropping the medical image based on the arteriovenous adhesion position to obtain a local image corresponding to the adhesion area;

[0135] The local image is input into a first model to obtain the first processing result corresponding to the local image.

[0136] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0137] Using the first processing result to replace the area corresponding to the local image in the first artery and vein segmentation result, to obtain a second artery and vein segmentation result;

[0138] A connected domain in the second artery and vein segmentation result is obtained, and a target artery and vein segmentation result of the tissue of interest is determined according to the connected domain in the second artery and vein segmentation result.

[0139] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0140] If a first connected domain of the first type exists in the second artery and vein segmentation result, and the first connected domain is connected to multiple second connected domains of the second type, the type of the first connected domain is modified to the second type to obtain a target artery and vein segmentation result of the tissue of interest.

[0141] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0142] If there is a third connected domain that is not connected to the largest connected domain in the target artery and vein segmentation result, the third connected domain is removed.

[0143] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0144] Determine the shortest distance from the foreground point to the background point on the blood vessel segmentation result to obtain a blood vessel distance map;

[0145] Determine the shortest distance between the foreground point and the background point on the first artery and vein segmentation result to obtain an artery and vein distance map;

[0146] The blood vessel distance map and the arteriovenous distance map are input into the second model to obtain the arteriovenous adhesion position of the tissue of interest.

[0147] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0148] Inputting the medical image into a third model to obtain a segmented image, wherein the segmented image includes the tissue of interest;

[0149] Based on the medical image and the segmented image, a blood vessel segmentation result of the tissue of interest and the first artery and vein segmentation result are obtained.

[0150] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0151] Obtaining a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image;

[0152] Determining the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result;

[0153] Based on the arteriovenous adhesion position, processing the adhesion area in the medical image to obtain a first processing result;

[0154] Based on the first processing result and the first artery and vein segmentation result, a target artery and vein segmentation result of the tissue of interest is determined.

[0155] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0156] Cropping the medical image based on the arteriovenous adhesion position to obtain a local image corresponding to the adhesion area;

[0157] The local image is input into a first model to obtain the first processing result corresponding to the local image.

[0158] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0159] Using the first processing result to replace the area corresponding to the local image in the first artery and vein segmentation result, to obtain a second artery and vein segmentation result;

[0160] A connected domain in the second artery and vein segmentation result is obtained, and a target artery and vein segmentation result of the tissue of interest is determined according to the connected domain in the second artery and vein segmentation result.

[0161] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0162] If a first connected domain of the first type exists in the second artery and vein segmentation result, and the first connected domain is connected to multiple second connected domains of the second type, the type of the first connected domain is modified to the second type to obtain a target artery and vein segmentation result of the tissue of interest.

[0163] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0164] If there is a third connected domain that is not connected to the maximum connected domain in the target artery and vein segmentation result, the third connected domain is removed.

[0165] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0166] Determine the shortest distance from the foreground point to the background point on the blood vessel segmentation result to obtain a blood vessel distance map;

[0167] Determine the shortest distance between the foreground point and the background point on the first artery and vein segmentation result to obtain an artery and vein distance map;

[0168] The blood vessel distance map and the arteriovenous distance map are input into the second model to obtain the arteriovenous adhesion position of the tissue of interest.

[0169] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0170] Inputting the medical image into a third model to obtain a segmented image, wherein the segmented image includes the tissue of interest;

[0171] Based on the medical image and the segmented image, a blood vessel segmentation result of the tissue of interest and the first artery and vein segmentation result are obtained.

[0172] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0173] Obtaining a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image;

[0174] Determining the arteriovenous adhesion position of the tissue of interest based on the blood vessel segmentation result and the first arteriovenous segmentation result;

[0175] Based on the arteriovenous adhesion position, processing the adhesion area in the medical image to obtain a first processing result;

[0176] Based on the first processing result and the first artery and vein segmentation result, a target artery and vein segmentation result of the tissue of interest is determined.

[0177] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0178] Cropping the medical image based on the arteriovenous adhesion position to obtain a local image corresponding to the adhesion area;

[0179] The local image is input into a first model to obtain the first processing result corresponding to the local image.

[0180] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0181] Using the first processing result to replace the area corresponding to the local image in the first artery and vein segmentation result, to obtain a second artery and vein segmentation result;

[0182] A connected domain in the second artery and vein segmentation result is obtained, and a target artery and vein segmentation result of the tissue of interest is determined according to the connected domain in the second artery and vein segmentation result.

[0183] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0184] If a first connected domain of the first type exists in the second artery and vein segmentation result, and the first connected domain is connected to multiple second connected domains of the second type, the type of the first connected domain is modified to the second type to obtain a target artery and vein segmentation result of the tissue of interest.

[0185] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0186] If there is a third connected domain that is not connected to the maximum connected domain in the target artery and vein segmentation result, the third connected domain is removed.

[0187] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0188] Determine the shortest distance from the foreground point to the background point on the blood vessel segmentation result to obtain a blood vessel distance map;

[0189] Determine the shortest distance between the foreground point and the background point on the first artery and vein segmentation result to obtain an artery and vein distance map;

[0190] The blood vessel distance map and the arteriovenous distance map are input into the second model to obtain the arteriovenous adhesion position of the tissue of interest.

[0191] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0192] Inputting the medical image into a third model to obtain a segmented image, wherein the segmented image includes the tissue of interest;

[0193] Based on the medical image and the segmented image, a blood vessel segmentation result of the tissue of interest and the first artery and vein segmentation result are obtained.

[0194] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0195] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0196] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0197] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A blood vessel segmentation method, characterized in that: The method comprises: Obtaining a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image; Determine the shortest distance from the foreground point to the background point on the blood vessel segmentation result to obtain a blood vessel distance map; determine the shortest distance from the foreground point to the background point on the first artery and vein segmentation result to obtain an arteriovenous distance map; input the blood vessel distance map and the arteriovenous distance map into a second model to obtain the arteriovenous adhesion position of the tissue of interest; Based on the arteriovenous adhesion position, processing the adhesion area in the medical image to obtain a first processing result; Based on the first processing result and the first artery and vein segmentation result, a target artery and vein segmentation result of the tissue of interest is determined.

2. The method according to claim 1, characterized in that The step of processing the adhesion region in the medical image based on the arteriovenous adhesion position to obtain a first processing result includes: Cropping the medical image based on the arteriovenous adhesion position to obtain a local image corresponding to the adhesion area; The local image is input into a first model to obtain the first processing result corresponding to the local image.

3. The method according to claim 2, characterized in that The determining, based on the first processing result and the first artery and vein segmentation result, a target artery and vein segmentation result of the tissue of interest includes: Using the first processing result to replace the area corresponding to the local image in the first artery and vein segmentation result, to obtain a second artery and vein segmentation result; A connected domain in the second artery and vein segmentation result is obtained, and a target artery and vein segmentation result of the tissue of interest is determined according to the connected domain in the second artery and vein segmentation result.

4. The method according to claim 3, characterized in that The step of determining the target artery and vein segmentation result of the tissue of interest according to the connected domain in the second artery and vein segmentation result includes: If a first connected domain of the first type exists in the second artery and vein segmentation result, and the first connected domain is connected to multiple second connected domains of the second type, the type of the first connected domain is modified to the second type to obtain a target artery and vein segmentation result of the tissue of interest.

5. The method according to claim 4, characterized in that The method further comprises: If there is a third connected domain that is not connected to the maximum connected domain in the target artery and vein segmentation result, the third connected domain is removed.

6. The method according to claim 1 or 2, characterized in that: The step of obtaining a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image includes: Inputting the medical image into a third model to obtain a segmented image, wherein the segmented image includes the tissue of interest; Based on the medical image and the segmented image, a blood vessel segmentation result of the tissue of interest and the first artery and vein segmentation result are obtained.

7. The method according to claim 6, characterized in that The step of inputting the medical image into a third model to obtain a segmented image comprises: Normalizing the medical image according to a preset window width and window level to obtain a normalized medical image; The normalized medical image is input into the third model to obtain the segmented image.

8. A blood vessel segmentation device, characterized in that: The device comprises: An acquisition module, used for acquiring a blood vessel segmentation result and a first artery and vein segmentation result of a tissue of interest in a medical image; The first determination module is used to determine the shortest distance from the foreground point to the background point on the blood vessel segmentation result to obtain a blood vessel distance map; determine the shortest distance from the foreground point to the background point on the first arteriovenous segmentation result to obtain an arteriovenous distance map; input the blood vessel distance map and the arteriovenous distance map into the second model to obtain the arteriovenous adhesion position of the tissue of interest; A processing module, configured to process the adhesion region in the medical image based on the arteriovenous adhesion position to obtain a first processing result; The second determination module is used to determine the target artery and vein segmentation result of the tissue of interest based on the first processing result and the first artery and vein segmentation result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images