CN114445412A - Blood vessel segmentation method, device and storage medium - Google Patents
Blood vessel segmentation method, device and storage medium Download PDFInfo
- Publication number
- CN114445412A CN114445412A CN202210111422.0A CN202210111422A CN114445412A CN 114445412 A CN114445412 A CN 114445412A CN 202210111422 A CN202210111422 A CN 202210111422A CN 114445412 A CN114445412 A CN 114445412A
- Authority
- CN
- China
- Prior art keywords
- blood vessel
- image
- path
- arteriovenous
- segmentation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 210000004204 blood vessel Anatomy 0.000 title claims abstract description 460
- 230000011218 segmentation Effects 0.000 title claims abstract description 143
- 238000000034 method Methods 0.000 title claims abstract description 71
- 230000004927 fusion Effects 0.000 claims abstract description 33
- 210000003462 vein Anatomy 0.000 claims description 75
- 210000001367 artery Anatomy 0.000 claims description 59
- 210000000056 organ Anatomy 0.000 claims description 49
- 238000005070 sampling Methods 0.000 claims description 21
- 238000004590 computer program Methods 0.000 claims description 19
- 238000010845 search algorithm Methods 0.000 claims description 14
- 238000010586 diagram Methods 0.000 claims description 9
- 238000009795 derivation Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 5
- 238000013135 deep learning Methods 0.000 abstract description 3
- 238000003062 neural network model Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 8
- 230000009466 transformation Effects 0.000 description 8
- 238000012549 training Methods 0.000 description 7
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 201000005202 lung cancer Diseases 0.000 description 6
- 208000020816 lung neoplasm Diseases 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 210000004072 lung Anatomy 0.000 description 5
- 238000010606 normalization Methods 0.000 description 5
- 230000002093 peripheral effect Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 206010028980 Neoplasm Diseases 0.000 description 3
- 238000002059 diagnostic imaging Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012163 sequencing technique Methods 0.000 description 3
- 201000011510 cancer Diseases 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 238000013170 computed tomography imaging Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 210000001147 pulmonary artery Anatomy 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- 238000001356 surgical procedure Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 229910021389 graphene Inorganic materials 0.000 description 1
- 210000002216 heart Anatomy 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 210000003734 kidney Anatomy 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 210000004185 liver Anatomy 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000496 pancreas Anatomy 0.000 description 1
- 230000002685 pulmonary effect Effects 0.000 description 1
- 210000003492 pulmonary vein Anatomy 0.000 description 1
- 238000002271 resection Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
Abstract
The present application relates to a blood vessel segmentation method, apparatus and storage medium. The method comprises the following steps: obtaining a blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model; inputting the initial medical image and the blood vessel mask image into a preset regression model to obtain an arteriovenous distance map; the arteriovenous distance map is used for representing the distance between each arterial blood vessel point and an arterial blood vessel boundary in the initial medical image and the distance between each venous blood vessel point and a venous blood vessel boundary in the initial medical image; and obtaining a blood vessel segmentation image corresponding to the initial medical image according to the blood vessel mask image, the arteriovenous distance image and a preset fusion model. By adopting the method, the blood vessel mask image and the blood vessel arteriovenous distance image can be extracted by utilizing deep learning, the blood vessel can be effectively segmented from the initial medical image, and then characteristic fusion is carried out on the basis of the arteriovenous distance image and the initial medical image to obtain a more accurate arteriovenous segmentation result.
Description
Technical Field
The present application relates to the field of image processing technologies, and in particular, to a blood vessel segmentation method, device, and storage medium.
Background
The incidence of cancer, such as lung cancer, has increased year by year in recent years. For lung cancer treatment, a lesion resection operation is generally required, and therefore preoperative planning before surgery is particularly important in the surgical process. For surgical planning of lung cancer, organs such as lung need to be reconstructed three-dimensionally, and a critical part in the three-dimensional reconstruction of each organ is to accurately segment blood vessels in each organ.
In the conventional technology, a threshold separation method is generally adopted for the segmentation of blood vessels in each organ, such as separating arteriovenous blood vessel boundaries by using a distance map of gray boundary features or using morphological directional thickness, and further obtaining a blood vessel segmentation image.
However, the current blood vessel segmentation method has the problem of low segmentation accuracy.
Disclosure of Invention
In view of the above, it is necessary to provide a blood vessel segmentation method, a device and a storage medium capable of accurately segmenting arteriovenous blood vessels in view of the above technical problems.
In a first aspect, the present application provides a vessel segmentation method. The method comprises the following steps:
obtaining a blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model;
inputting the initial medical image and the blood vessel mask image into a preset regression model to obtain an arteriovenous distance map; the arteriovenous distance map is used for representing the distance between each arterial blood vessel point and an arterial blood vessel boundary in the initial medical image and the distance between each venous blood vessel point and a venous blood vessel boundary in the initial medical image;
and obtaining a blood vessel segmentation image corresponding to the initial medical image according to the blood vessel mask image, the arteriovenous distance image and a preset fusion model.
In one embodiment, obtaining a blood vessel segmentation image corresponding to an initial medical image according to a blood vessel mask image, an arteriovenous distance image and a preset fusion model includes:
determining a target path of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance image;
and inputting the blood vessel mask image, the arteriovenous distance image and the target path of each blood vessel into a preset fusion model to obtain a blood vessel segmentation image corresponding to the initial medical image.
In one embodiment, determining the target path of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance map comprises:
determining the starting point of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance image;
determining the termination point of each blood vessel in the blood vessel mask image according to the blood vessel mask image;
normalizing each distance value in the arteriovenous distance map to obtain a processed distance map;
determining paths between the starting points of all blood vessels and the termination points of all blood vessels according to the processed distance map and the shortest path search algorithm;
and determining the target path of each blood vessel according to the path between the starting point of each blood vessel and the ending point of each blood vessel.
In one embodiment, the starting points of the respective vessels include the starting point of a venous vessel and the starting point of an arterial vessel; determining a path between the starting point of each blood vessel and the ending point of each blood vessel according to the processed distance map and the shortest path search algorithm, wherein the path comprises the following steps:
for the termination point of each blood vessel, adopting a shortest path search algorithm to search a path between the termination point of each blood vessel and the starting point of the vein blood vessel and determine the path as a first path of each blood vessel;
and the number of the first and second groups,
and for the termination point of each blood vessel, searching a path between the starting point of the artery blood vessel and the termination point of each blood vessel by adopting a shortest path searching algorithm and determining the path as a second path of each blood vessel.
In one embodiment, determining the target path of each blood vessel from the path between the start point of each blood vessel and the end point of each blood vessel comprises:
the first path of each blood vessel is derived to obtain a first slope corresponding to the first path;
conducting derivation on the second path of each blood vessel to obtain a second slope corresponding to the second path;
and determining the target path of each blood vessel by using the first slope, the second slope and a preset slope threshold.
In one embodiment, the segmentation model comprises a first network model and a second network model; obtaining a blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model, wherein the obtaining of the blood vessel mask image comprises the following steps:
inputting the initial medical image into a first network model for rough segmentation to obtain a segmented image of an organ in the initial medical image;
determining the position of an organ in the initial medical image according to the segmentation image of the organ, and acquiring a minimum circumscribed cube of the organ;
inputting the minimum circumscribed cube into a second network model for fine segmentation to obtain a blood vessel mask image; wherein the sampling resolution of the second network model is less than the sampling resolution of the first network model.
In one embodiment, determining the starting point of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance map comprises:
processing the blood vessel mask image by utilizing the segmented image of the organ to obtain a main blood vessel of an arterial blood vessel and a main blood vessel of a venous blood vessel;
determining a connected domain corresponding to a main vessel of the artery and vein in the arteriovenous distance map according to the main vessel of the artery and vein, and determining a point with the maximum distance value in the connected domain corresponding to the main vessel of the artery and vein as a starting point of the artery and vein;
determining a connected domain corresponding to the main vessel of the vein vessel in the arteriovenous distance map according to the main vessel of the vein vessel, and determining a point with the maximum distance value in the connected domain corresponding to the main vessel of the vein vessel as a starting point of the vein vessel.
In one embodiment, the blood vessel segmentation method further includes:
determining a bifurcation point of each blood vessel in the blood vessel segmentation image according to a shortest path search method;
and (3) according to a preset segmentation grading rule, segmenting and grading the blood vessel sections between adjacent bifurcation points in each blood vessel to obtain a structure diagram of each blood vessel.
In a second aspect, the present application further provides a vessel segmentation apparatus. The device includes:
the segmentation module is used for obtaining a blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model;
the regression module is used for inputting the initial medical image and the blood vessel mask image into a preset regression model to obtain an arteriovenous distance map; the arteriovenous distance map is used for representing the distance between each arterial blood vessel point and an arterial blood vessel boundary in the initial medical image and the distance between each venous blood vessel point and a venous blood vessel boundary in the initial medical image;
and the fusion module is used for obtaining a blood vessel segmentation image corresponding to the initial medical image according to the blood vessel mask image, the arteriovenous distance image and a preset fusion model.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the steps of the method in any of the embodiments of the first aspect when the computer program is executed.
In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method according to any of the above-mentioned embodiments of the first aspect.
In a fifth aspect, the present application further provides a computer program product. The computer program product comprising a computer program that, when executed by a processor, performs the steps of the method according to any of the embodiments of the first aspect described above.
According to the blood vessel segmentation method, the device and the storage medium, the blood vessel mask image in the initial medical image is obtained according to the initial medical image and the preset segmentation model, the initial medical image and the blood vessel mask image are input into the preset regression model to obtain the arteriovenous distance image, the blood vessel segmentation image corresponding to the initial medical image is obtained according to the blood vessel mask image, the arteriovenous distance image and the preset fusion model, the blood vessel mask image and the blood vessel arteriovenous distance image can be extracted by utilizing depth learning, blood vessels can be effectively segmented from the background of the initial medical image, and then feature fusion is carried out on the basis of the arteriovenous distance image and the initial medical image to obtain a relatively accurate artery and vein segmentation result.
Drawings
FIG. 1 is a diagram of an embodiment of a vessel segmentation method;
FIG. 2 is a flow diagram illustrating a method for vessel segmentation in one embodiment;
FIG. 3 is a flow chart illustrating a method for vessel segmentation in another embodiment;
FIG. 4 is a flow chart illustrating a method for vessel segmentation in another embodiment;
FIG. 5 is a flow chart illustrating a method for vessel segmentation in another embodiment;
FIG. 6 is a flow chart illustrating a method for vessel segmentation in another embodiment;
FIG. 7 is a flow chart illustrating a method for vessel segmentation in another embodiment;
FIG. 8 is a flow chart illustrating a method for vessel segmentation in another embodiment;
fig. 9 is a block diagram showing a structure of a blood vessel segmentation apparatus according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The incidence of cancer has increased year by year in recent years, with lung cancer being the first factor in cancers that lead to death in men. The treatment of lung cancer generally requires a lesion excision operation, the five-year survival rate after operation is closely related to the quality of the operation, and preoperative planning before the operation is particularly important in the operation process. In the surgical planning of lung cancer, organ tissues and the like of lung need to be reconstructed in three dimensions by means of image reconstruction, so as to relate to surgical approaches, surgical treatment schemes and the like. In three-dimensional reconstruction, the more critical part is the reconstruction of the blood vessel. The spatial position relationship between the blood vessel and the operation path, the spatial proximity relationship between the blood vessel and the focus, the operation scheme of the blood vessel ligation in the operation and the like are all closely related to the success or failure of the operation.
The pulmonary vessel tree has rich branches, and after entering from the lung portal, the arterial trunk and the venous trunk are branched and interpenetrated for many times to walk to the lung margin. Because the blood vessel tree has long walking path, rich branches and many branch variations, the complete and accurate pulmonary blood vessel is difficult to be segmented by adopting the traditional algorithm based on filtering or region growing. And deep learning can adapt to various vessel tree forms due to good feature extraction and expression capability. However, there are many adhesion situations of different forms in the main part of the pulmonary artery and vein in the process of interpenetration walking, and the condition of segmentation and classification errors is easily caused by using the gray scale enhancement inconsistency of the increase and decrease data obtained by scanning the target object by the medical imaging equipment, and directly segmenting the artery and vein blood vessels of the initial image obtained by scanning the medical imaging equipment by using the deep learning neural network.
Based on this, the embodiment of the present application provides a blood vessel segmentation method, which may be applied to a computer device as shown in fig. 1, where the computer device may be a terminal. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, the initial medical image obtained by scanning the target object by the medical imaging equipment is input into a preset segmentation model to obtain the blood vessel mask image. And obtaining an arteriovenous distance image according to the blood vessel mask image and the initial medical image, and fusing the blood vessel mask image and the arteriovenous distance image to obtain a blood vessel segmentation image corresponding to the initial medical image. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
It should be noted that, in the blood vessel segmentation method provided in the embodiment of the present application, an execution subject may be a Computed Tomography (CT) imaging apparatus, and the CT imaging apparatus may be implemented as part or all of a computer device by software, hardware, or a combination of software and hardware. In the following method embodiments, the execution subject is a computer device as an example.
Those skilled in the art will appreciate that the architecture shown in fig. 1 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, as shown in fig. 2, a blood vessel segmentation method is provided, which is described by taking the method as an example applied to the computer device in fig. 1, and comprises the following steps:
s202, obtaining a blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model.
The initial medical image is pre-processed DICOM data obtained by scanning a target part of a target object by a CT imaging device. The DICOM data can be preprocessed and converted into volume data, and the volume data is subjected to unified window width and window level normalization. For example, an enhancement window can be used, taking [40,400] as the normalized mean and standard deviation, and when a voxel in the volume data is out of the preset maximum range, the out voxel is adjusted to the nearest maximum range boundary value. For example, with [40,400] as the mean and standard deviation of the normalization, the voxels in the volume data need to be adjusted to-360 to +440, if the voxels are less than-360, the voxels are adjusted to-360, and if the voxels are greater than +440, the voxels are adjusted to + 440. Wherein the target object may comprise a human body or an animal body. The target site may include: lung, heart, liver, kidney, pancreas, etc. The preset segmentation model may include a convolutional neural network model such as a VB-Net neural network model, a V-Net neural network model, and a U-Net neural network model, which is not limited herein. The preset segmentation model can adopt an Adam training optimizer, and the loss function can adopt a DiceLoss loss function, a cross entropy loss function and the like.
Specifically, the initial medical image may be input into a preset segmentation model, and then the blood vessel segmentation may be performed on the initial medical image to obtain a blood vessel mask image in the initial medical image. Or inputting the initial medical image into a preset first network model with a first sampling resolution ratio for rough segmentation to obtain a target part in the determined initial medical image, capturing an image of the target part from the initial medical image, and inputting the image of the target part into a preset second network model with a second sampling resolution ratio again for fine segmentation to obtain a blood vessel mask image. Wherein the first sampling resolution is greater than the second sampling resolution; and the difference value between the first sampling resolution and the second sampling resolution is greater than or equal to a preset difference value threshold value. For example, if the preset difference threshold is 4.95, the first sampling resolution is 5mm, and the second sampling resolution is 0.05mm at most.
S204, inputting the initial medical image and the blood vessel mask image into a preset regression model to obtain an arteriovenous distance map; the arteriovenous distance map is used for representing the distance between each arterial blood vessel point and an arterial blood vessel boundary in the initial medical image and the distance between each venous blood vessel point and a venous blood vessel boundary in the initial medical image.
The preset regression model may include a dual-channel convolutional neural network model such as a VB-Net neural network model, a V-Net neural network model, and a U-Net neural network model, which is not limited herein. The arteriovenous distance map is a distance map of two types of blood vessels including an artery and a vein.
Specifically, an initial medical image and a blood vessel mask image are input into a preset regression model, the blood vessel mask image is used as enhancement information to restrict, characteristic information of a blood vessel and non-blood vessels is extracted, distance transformation is carried out, and an arteriovenous distance map is output.
Optionally, in the training process of the preset regression model, the setting of the initial parameters may be migration learning by using the training parameters of the preset segmentation model, and the Loss function is an L1Loss function; the regression golden standard of the preset regression model is a standard arteriovenous distance map obtained by distance conversion of a standard arteriovenous vessel map.
And S206, obtaining a blood vessel segmentation image corresponding to the initial medical image according to the blood vessel mask image, the arteriovenous distance image and the preset fusion model.
Specifically, after obtaining the blood vessel mask image and the arteriovenous distance map, the blood vessel mask image and the arteriovenous distance map can be input into the fusion model to perform feature extraction of the artery blood vessel and the vein blood vessel, so as to obtain a blood vessel segmentation image corresponding to the fused initial medical image. The blood vessel mask image can provide blood vessel boundary information of the blood vessel indistinguishable branch veins, and the arteriovenous distance image can provide blood vessel branch structure, diameter information and boundary information. The fusion model judges the structures of the artery and the vein according to the structure of the blood vessel branches, determines the respective boundaries of the arteriovenous by utilizing the diameter information and the blood vessel boundary information of the arteriovenous, and obtains the segmentation results of the arteriovenous classification. The fusion model can be a convolution neural network model such as a multi-channel VB-Net neural network model, a V-Net neural network model, a U-Net neural network model and the like.
Optionally, the initialization parameter used in the fusion model training process may be a training parameter of a preset segmentation model, and the preset training parameter of the segmentation model is used for performing transfer learning, so as to complete the training of the fusion model.
According to the blood vessel segmentation method, a blood vessel mask image in an initial medical image is obtained according to the initial medical image and a preset segmentation model, the initial medical image and the blood vessel mask image are input into a preset regression model to obtain an arteriovenous distance map, a blood vessel segmentation image corresponding to the initial medical image is obtained according to the blood vessel mask image, the arteriovenous distance map and a preset fusion model, the blood vessel mask image and the blood vessel arteriovenous distance map can be extracted by utilizing depth learning, a blood vessel can be effectively segmented from the background of the initial medical image, and then feature fusion is carried out on the basis of the arteriovenous distance map and the initial medical image to obtain a relatively accurate artery and vein segmentation result.
The above embodiment explains the blood vessel segmentation method, and how to obtain a blood vessel segmentation image corresponding to an initial medical image according to a blood vessel mask image, an arteriovenous distance map and a preset fusion model in the method is further explained by an embodiment. In an embodiment, as shown in fig. 3, obtaining a blood vessel segmentation image corresponding to the initial medical image according to the blood vessel mask image, the arteriovenous distance map, and the preset fusion model includes:
s302, determining the target path of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance image.
Specifically, after a blood vessel mask image is obtained, the blood vessel boundary information of the arteriovenous which is not distinguished by each blood vessel can be determined, and the arteriovenous distance map can provide the blood vessel branch structure, the diameter information and the boundary information of the arteriovenous. And determining the path of each blood vessel according to the blood vessel boundary information of the artery and the vein of the blood vessel mask image, the blood vessel branch structure, the diameter information and the boundary information in the arteriovenous distance image, namely determining the target path of each blood vessel in the blood vessel mask image.
Further, in an embodiment, as shown in fig. 4, determining a target path of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance map includes:
s402, determining the starting point of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance image.
Specifically, after the blood vessel mask map is determined, the main blood vessels of the arterial blood vessels and the main blood vessels of the venous blood vessels can be determined according to the blood vessel mask map. After the main vessels of the arterial blood vessels and the main vessels of the venous blood vessels are determined, due to the fact that the blood vessel mask image and the arteriovenous distance image have the corresponding relation, the connected domain corresponding to the main vessels of the arterial blood vessels in the corresponding arteriovenous distance image can be determined according to the main vessels of the arterial blood vessels in the blood vessel mask image; and determining a connected domain corresponding to the main blood vessel of the vein blood vessel in the corresponding arteriovenous distance map according to the main blood vessel of the vein blood vessel in the blood vessel mask map. Furthermore, the position with the largest diameter in the connected domain corresponding to the main blood vessel of the artery blood vessel and the position with the largest diameter in the connected domain corresponding to the main blood vessel of the vein blood vessel can be correspondingly found out to be used as the starting point of each blood vessel. For example, after obtaining the blood vessel mask image, the location area of the main blood vessel of the artery blood vessel and the location area of the main blood vessel of the vein blood vessel in the blood vessel mask image can be firstly determined according to the blood vessel mask image. Then, according to the position area of the main artery and vein of the artery and vein, finding out the corresponding position area in the arteriovenous distance map as the connected domain of the main artery and vein of the artery and vein. Similarly, according to the position area of the main blood vessel of the vein blood vessel, the corresponding position area in the arteriovenous distance map is found and is used as the connected domain of the main blood vessel of the vein blood vessel. Then, the place with the largest distance value, namely the thickest blood vessel, is found in the connected domain of the main blood vessel of the artery blood vessel as the starting point of the artery blood vessel. Finding the place with the largest distance value, namely the thickest blood vessel, in the connected domain of the main blood vessel of the vein blood vessel as the starting point of the vein blood vessel.
Optionally, before determining the starting point of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance map, a further explanation of the blood vessel mask image in the initial medical image can be obtained according to the initial medical image and a preset segmentation model, in an embodiment, as shown in fig. 6, the segmentation model includes a first network model and a second network model; obtaining a blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model, wherein the obtaining of the blood vessel mask image comprises the following steps:
s602, the initial medical image is input into the first network model for rough segmentation, and a segmented image of an organ in the initial medical image is obtained.
S604, according to the segmented image of the organ, the position of the organ is determined in the initial medical image, and the minimum circumscribed cube of the organ is obtained.
S606, inputting the minimum circumscribed cube into a second network model for fine segmentation to obtain a blood vessel mask image; wherein the sampling resolution of the second network model is less than the sampling resolution of the first network model.
When the rough segmentation is performed, the sampling resolution of the first network model may be set to be 5 mm; for fine segmentation, the second network model may be set to 0.5 mm. Wherein, the first network model and the second network model can be VB-net neural network models.
Specifically, the initial medical image is input into the first network model, segmentation is performed, background information is removed, only foreground information containing the position of the organ is reserved, and the segmented image of the organ in the initial medical image is obtained. After obtaining the segmented image of the organ, the position of the organ in the initial medical image can be determined by using the segmented image of the organ, and the minimum circumscribed cube of the organ is taken based on the position of the organ. And then inputting the minimum circumscribed cube into a second network model, and determining each level of blood vessel in the organ to obtain a blood vessel mask image.
In this embodiment, an initial medical image is input into a first network model for rough segmentation to obtain a segmented image of an organ in the initial medical image, the position of the organ is determined in the initial medical image according to the segmented image of the organ, a minimum circumscribed cube of the organ is obtained, and the minimum circumscribed cube is input into a second network model for fine segmentation to obtain a blood vessel mask image. The position area of the organ is accurately determined in the first network model based on the larger sampling resolution, the target area needing blood vessel segmentation is determined according to the position area of the organ, the image of the target area is input into the second network model based on the smaller sampling resolution, the accurate blood vessel segmentation result in the organ is obtained, and the accuracy of blood vessel segmentation is improved.
Based on the determined blood vessel mask image, further, in an embodiment, as shown in fig. 7, determining a starting point of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance map, includes:
s702, the blood vessel mask image is processed by utilizing the segmentation image of the organ, and the main blood vessel of the arterial blood vessel and the main blood vessel of the venous blood vessel are obtained.
Specifically, if the segmented image of the organ is a mask image of the organ, the region where the segmented image and the blood vessel mask image intersect can be subtracted from the blood vessel mask image to obtain the filtering results, i.e., the main blood vessels of the blood vessels and the main blood vessels of the blood vessels.
S704, determining a connected domain corresponding to the main artery of the artery and vein in the arteriovenous distance map according to the main artery of the artery and vein, and determining a point with the maximum distance value in the connected domain corresponding to the main artery of the artery and vein as a starting point of the artery and vein.
Specifically, according to the position region of the main artery blood vessel of the artery blood vessel in the blood vessel mask image, the connected domain corresponding to the main artery blood vessel of the artery blood vessel in the same position region in the arteriovenous distance image is determined, and the point with the maximum diameter distance value in the connected domain corresponding to the main artery blood vessel of the artery blood vessel is determined as the starting point of the artery blood vessel.
S706, determining a connected domain corresponding to the main vessel of the vein in the arteriovenous distance map according to the main vessel of the vein, and determining a point with the maximum distance value in the connected domain corresponding to the main vessel of the vein as the starting point of the vein.
Specifically, according to the position region of the main blood vessel of the vein blood vessel in the blood vessel mask map, the connected domain corresponding to the main blood vessel of the vein blood vessel in the same position region in the arteriovenous distance map is determined, and the point with the maximum distance value in the connected domain corresponding to the main blood vessel of the vein blood vessel is determined as the starting point of the vein blood vessel.
In this embodiment, the vessel mask image is filtered through the segmented image of the organ to obtain a main region of an arterial blood vessel and a main region of a venous blood vessel of the organ, a connected domain corresponding to the main blood vessel of the arterial blood vessel in the arterial-venous distance map is determined according to the main region of the arterial blood vessel, a connected domain corresponding to the main blood vessel of the venous blood vessel in the venous distance map is determined according to the main region of the venous blood vessel, a point with the largest distance value in the connected domain corresponding to the main blood vessel of the arterial blood vessel is found and determined as a starting point of the arterial blood vessel, and a point with the largest distance value in the connected domain corresponding to the main blood vessel of the venous blood vessel is found and determined as a starting point of the venous blood vessel.
S404, determining the termination point of each blood vessel in the blood vessel mask image according to the blood vessel mask image.
Specifically, extracting the central line of each blood vessel in the blood vessel mask image; and traversing each point on the central line of each blood vessel by using a depth traversal search algorithm, and taking the point with the number of the neighborhood points being 1 as the termination point of each blood vessel. Wherein the termination point is the distal end point of each blood vessel.
S406, normalizing each distance value in the arteriovenous distance map to obtain a processed distance map.
The distance value is the distance from the foreground point of each blood vessel in the arteriovenous distance map to the boundary of the blood vessel. For example, the distance from a foreground point to the boundary of the cross section of the blood vessel where the foreground point is located, i.e. the radius value of the cross section.
Specifically, each distance value in the arteriovenous distance map is input into a normalized scaling transformation formula, wherein the normalized scaling transformation formula is as follows:
wherein D is a distance value from the foreground point of each blood vessel to the corresponding blood vessel boundary, max (D) is a maximum distance value in the distance values, and D' is a value obtained by transforming the distance value from the foreground point of each blood vessel to the corresponding blood vessel boundary through a normalization scaling transformation formula. After the distance values obtained by the distance values through the normalization scaling transformation formula are obtained, a processed distance map can be formed according to the distance values obtained through the normalization scaling transformation formula. The value distribution of the transformed distance map has the following characteristics: the value on the central line of each blood vessel changes from small to large from the starting point of the blood vessel to the ending point of the blood vessel; the same cross-section vessel, from center to boundary, varies in value from small to large. By utilizing the characteristics of the processed distance map,by searching for a path between the start point and each end point of the arteriovenous two categories, the change in the correct path value is a smooth transition from small to large, with a sudden change in value when a path crossing occurs.
And S408, determining a path between the starting point of each blood vessel and the ending point of each blood vessel according to the processed distance map and the shortest path search algorithm.
Specifically, a shortest path search algorithm is adopted to perform path search on the center line of the processed distance map, and a path between the starting point of each blood vessel and the ending point of each blood vessel is obtained. That is, the termination point of each blood vessel may correspond to the start of an artery and the start of a vein, respectively, resulting in two paths.
S410, determining a target path of each blood vessel according to the path between the starting point of each blood vessel and the ending point of each blood vessel.
Specifically, for the termination point of each blood vessel, a shortest path search algorithm is adopted to search a path between the termination point of each blood vessel and the starting point of the vein blood vessel and determine the path as a first path of each blood vessel; and for the termination points of the blood vessels, searching a path between the starting point of the artery blood vessel and the termination point of the blood vessel by adopting a shortest path searching algorithm and determining the path as a second path of the blood vessel. Optionally, the first path and the second path may be differentiated respectively to obtain a first slope of the first path and a second slope of the second path, and the target path may be determined according to the first slope of the first path and the second slope of the second path. Further, the first slope and the second slope may be compared, and the target path may be determined according to the comparison result.
In the present embodiment, since the value distribution of the distance map after the normalized scaling transform has a range from the trunk coarse vessel to the peripheral fine vessel, the value on the centerline of the vessel changes from small to large; the blood vessel with the same section has the characteristic that the value changes from small to large from the center to the boundary. Then, the characteristics of the distance map after normalized scaling transformation can be utilized to search the path from the starting point of the two arteriovenous vessels to each peripheral end point (end point), a correct target path is determined according to the slopes of the first path and the second path, the starting point corresponding to each peripheral end point is accurately determined to be the starting point of the arterial vessel or the starting point of the venous vessel, and further the correct path between the peripheral end point and the target starting point, namely the target path of the arterial vessel or the venous vessel, is determined, so that the classification of the vessel categories is realized.
Further, in one embodiment, as shown in fig. 5, determining the target path of each blood vessel according to the path between the starting point of each blood vessel and the ending point of each blood vessel includes:
s502, derivation is carried out on the first path of each blood vessel to obtain a first slope corresponding to the first path.
Specifically, derivation is performed on each point of the first path of each blood vessel, so that a plurality of first slopes corresponding to the first path can be obtained.
S504, the second path of each blood vessel is derived to obtain a second slope corresponding to the second path.
Specifically, derivation is performed on each point of the second path of each blood vessel, so that a plurality of second slopes corresponding to the second path can be obtained.
S506, determining the target path of each blood vessel according to the first slope, the second slope and a preset slope threshold.
Specifically, each first slope is compared with a preset slope threshold, the absolute value of the first slope larger than the preset slope threshold is taken, and the first slopes are sorted according to the absolute value of each first slope.
And comparing each second slope with a preset slope threshold, taking the absolute value of the second slopes larger than the preset slope threshold, and sequencing according to the absolute value of each second slope.
And according to the sequencing order, comparing the first slope with the maximum absolute value of the first slope with the second slope with the maximum absolute value of the second slope, and taking the path corresponding to the larger slope as the target path. And if the first slope with the maximum absolute value of the first slope is equal to the second slope with the maximum absolute value of the second slope, comparing the first slope with the second maximum absolute value in the sequencing order, and taking the path corresponding to the larger slope as the target path. If the first slope and the second slope with the second largest absolute value are equal, the first slope and the second slope with the third largest absolute value are judged, and the like is performed until the target path is determined. For example, the first slope is 0.1, 0.2, -0.3, 0.4; the second slope is 0.11, -0.21, 0.31 and 0.41; if the preset slope threshold is 0.15, the first slope greater than the preset slope threshold is 0.2 or 0.4, and the second slope greater than the preset slope threshold is 0.31 or 0.41. Sorting the first slope into 0.4 and 0.2 after taking the absolute value; the absolute value of the second slope is taken and then the second slope is ranked as 0.41 and 0.31. At this time, the first slope value and the second slope value of the first sequence are compared, and if 0.41 is greater than 0.4, the second path corresponding to the second slope is the target path.
In this embodiment, a first slope corresponding to a first path is obtained by deriving the first path of each blood vessel, a second slope corresponding to a second path is obtained by deriving the second path of each blood vessel, and a target path of each blood vessel is determined by using the first slope, the second slope, and a preset slope threshold. The correct path, namely the target path, can be accurately determined through the slope value of the path.
S304, inputting the blood vessel mask image, the arteriovenous distance image and the target path of each blood vessel into a preset fusion model to obtain a blood vessel segmentation image corresponding to the initial medical image.
The vessel mask image provides vessel boundary information of vessels indistinguishable from arteriovenous vessels, the target path of each vessel provides branch and trunk structure information of arteriovenous vessels, and the arteriovenous distance map provides rough diameter and boundary information of arteriovenous vessel branch and trunk structures. And inputting the blood vessel mask image, the arteriovenous distance image and the target path of each blood vessel into a preset fusion model to obtain a blood vessel segmentation image corresponding to the initial medical image.
Optionally, the segmentation probability map, the arteriovenous distance map, and the target path of each blood vessel corresponding to the blood vessel mask image may be input to a preset fusion model, so as to obtain a blood vessel segmentation image corresponding to the initial medical image. The segmentation probability map corresponding to the blood vessel mask image is a probability map of which each pixel point is a foreground point obtained when the initial medical image is input into a preset segmentation model for segmentation, and further, the blood vessel mask image can be obtained after the probability map is activated by adopting an activation function.
In the embodiment, the target path of each blood vessel in the blood vessel mask image is determined according to the blood vessel mask image and the arteriovenous distance image, the blood vessel mask image, the arteriovenous distance image and the target path of each blood vessel are input into a preset fusion model to obtain a blood vessel segmentation image corresponding to an initial medical image, the path from each peripheral end point to the arterial trunk starting point and the venous trunk starting point can be searched on the regressed arteriovenous distance image through a shortest path search algorithm, and according to the change characteristics of the distance path weight values after normalized scaling transformation, a reasonable target path can be distinguished, a blood vessel segmentation image with the arterial blood vessel path and the venous blood vessel path is obtained, and an accurate blood vessel segmentation result is obtained. Because the blood vessel generally has a certain width, the algorithm based on the path search has higher fault tolerance and better interpretability than the method based on the threshold separation.
In the above description of the blood vessel segmentation method in the embodiment, after the blood vessel segmentation image is obtained, segmentation and classification may be further performed to form a complete tree-level branch structure. To illustrate this by an embodiment, in one implementation, as shown in fig. 8, the blood vessel segmentation method further includes:
s802, determining the bifurcation point of each blood vessel in the blood vessel segmentation image according to a shortest path search method.
S804, according to the preset segmentation grading rule, the blood vessel sections between adjacent bifurcation points in each blood vessel are segmented and graded to obtain a structure chart of each blood vessel.
Specifically, a shortest path search algorithm is adopted to search the center line of the blood vessel segmentation image to obtain the bifurcation point of each blood vessel. Triggered from the starting point of artery and vein trunk, when meeting the bifurcation point, the classification rule is carried out, the artery and vein vessels are divided into different grades from thick to thin, and the section between every two bifurcation points is taken as different sections to obtain the structure chart of each vessel. The shortest path algorithm includes, but is not limited to, Dijkstra algorithm such as Bellman-ford algorithm, SPFA algorithm, Floyd algorithm, etc.
In this embodiment, the bifurcation point of each blood vessel in the blood vessel segmentation image is determined according to the shortest path search method, and the blood vessel segments between adjacent bifurcation points in each blood vessel are segmented and graded according to the preset segmentation and grading rules to obtain the structure diagram of each blood vessel, so that the structure diagram of the blood vessel with a clear tree-level structure can be obtained.
To facilitate understanding by those skilled in the art, the vessel segmentation method will be further described with an embodiment. In one embodiment, a vessel segmentation method includes:
and S10, inputting the initial medical image into the first network model for rough segmentation to obtain a segmented image of the organ in the initial medical image.
S20, according to the segmentation image of the organ, the position of the organ is determined in the initial medical image, and the minimum circumscribed cube of the organ is obtained.
S30, inputting the minimum circumscribed cube into a second network model for fine segmentation to obtain a blood vessel mask image; wherein the sampling resolution of the second network model is less than the sampling resolution of the first network model.
S40, inputting the initial medical image and the blood vessel mask image into a preset regression model to obtain an arteriovenous distance map; the arteriovenous distance map is used for representing the distance between each arterial blood vessel point and an arterial blood vessel boundary in the initial medical image and the distance between each venous blood vessel point and a venous blood vessel boundary in the initial medical image.
S50, the blood vessel mask image is processed by using the segmented image of the organ, and a main blood vessel of the arterial blood vessel and a main blood vessel of the venous blood vessel are obtained.
S60, determining a connected domain corresponding to the main artery and vein of the artery and vein distance map according to the main artery and vein of the artery and vein, and determining the point with the maximum distance value in the connected domain corresponding to the main artery and vein of the artery and vein as the starting point of the artery and vein.
S70, determining a connected domain corresponding to the main vessel of the vein in the arteriovenous distance map according to the main vessel of the vein, and determining the point with the maximum distance value in the connected domain corresponding to the main vessel of the vein as the starting point of the vein.
And S80, determining the termination point of each blood vessel in the blood vessel mask image according to the blood vessel mask image.
And S90, normalizing each distance value in the arteriovenous distance map to obtain a processed distance map.
And S100, determining a path between the starting point of each blood vessel and the ending point of each blood vessel according to the processed distance map and the shortest path search algorithm.
S110, derivation is carried out on the first path of each blood vessel to obtain a first slope corresponding to the first path.
And S120, deriving the second path of each blood vessel to obtain a second slope corresponding to the second path.
And S130, determining the target path of each blood vessel by using the first slope, the second slope and a preset slope threshold.
And S140, inputting the blood vessel mask image, the arteriovenous distance map and the target path of each blood vessel into a preset fusion model to obtain a blood vessel segmentation image corresponding to the initial medical image.
And S150, determining the bifurcation point of each blood vessel in the blood vessel segmentation image according to a shortest path search method.
And S160, carrying out segmentation grading on the blood vessel sections between the adjacent bifurcation points in each blood vessel according to a preset segmentation grading rule to obtain a structure diagram of each blood vessel.
In this embodiment, a vessel mask image in an initial medical image is obtained according to the initial medical image and a preset segmentation model, the initial medical image and the vessel mask image are input into a preset regression model to obtain an arteriovenous distance map, a vessel segmentation image corresponding to the initial medical image is obtained according to the vessel mask image, the arteriovenous distance map and a preset fusion model, the vessel mask image and the vessel arteriovenous distance map can be extracted by using depth learning, a vessel can be effectively segmented from the background of the initial medical image, and then feature fusion is performed on the basis of the arteriovenous distance map and the initial medical image to obtain a relatively accurate artery and vein segmentation result.
It should be understood that, although the steps in the flowcharts related to the above embodiments are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts related to the above embodiments may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least a part of the steps or stages in other steps.
Based on the same inventive concept, the embodiment of the present application further provides a blood vessel segmentation device for implementing the blood vessel segmentation method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so specific limitations in one or more embodiments of the blood vessel segmentation device provided below can be referred to the limitations on the blood vessel segmentation method in the above description, and details are not repeated here.
In one embodiment, as shown in fig. 9, there is provided a blood vessel segmentation device, including:
a segmentation module 901, configured to obtain a blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model;
a regression module 902, configured to input the initial medical image and the blood vessel mask image into a preset regression model to obtain an arteriovenous distance map; the arteriovenous distance map is used for representing the distance between each arterial blood vessel point and an arterial blood vessel boundary in the initial medical image and the distance between each venous blood vessel point and a venous blood vessel boundary in the initial medical image;
and a fusion module 903, configured to obtain a blood vessel segmentation image corresponding to the initial medical image according to the blood vessel mask image, the arteriovenous distance map, and a preset fusion model.
The apparatus for determining a target object provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
In one embodiment, a fusion module includes:
the first determining unit is used for determining a target path of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance image;
and the second determining unit is used for inputting the blood vessel mask image, the arteriovenous distance map and the target path of each blood vessel into a preset fusion model to obtain a blood vessel segmentation image corresponding to the initial medical image.
In an embodiment, the first determining unit is specifically configured to determine starting points of blood vessels in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance map; determining the termination point of each blood vessel in the blood vessel mask image according to the blood vessel mask image; normalizing each distance value in the arteriovenous distance map to obtain a processed distance map; determining paths between the starting points of all blood vessels and the termination points of all blood vessels according to the processed distance map and the shortest path search algorithm; and determining the target path of each blood vessel according to the path between the starting point of each blood vessel and the ending point of each blood vessel.
In one embodiment, the start points of the respective vessels comprise the start point of a venous vessel and the start point of an arterial vessel; the first determining unit is specifically used for searching a path between the termination point of each blood vessel and the start point of a vein blood vessel by adopting a shortest path searching algorithm for the termination point of each blood vessel and determining the path as a first path of each blood vessel; and for the termination point of each blood vessel, searching a path between the starting point of the artery blood vessel and the termination point of each blood vessel by adopting a shortest path searching algorithm and determining the path as a second path of each blood vessel.
In an embodiment, the first determining unit is specifically configured to derive a first path of each blood vessel to obtain a first slope corresponding to the first path; conducting derivation on the second path of each blood vessel to obtain a second slope corresponding to the second path; and determining the target path of each blood vessel by using the first slope, the second slope and a preset slope threshold.
In one embodiment, the segmentation model comprises a first network model and a second network model; a segmentation module comprising:
the rough segmentation unit is used for inputting the initial medical image into the first network model for rough segmentation to obtain a segmented image of an organ in the initial medical image;
the second determining unit is used for determining the position of the organ in the initial medical image according to the segmented image of the organ and acquiring a minimum circumscribed cube of the organ;
the fine segmentation unit inputs the minimum circumscribed cube into a second network model for fine segmentation to obtain a blood vessel mask image; wherein the sampling resolution of the second network model is less than the sampling resolution of the first network model.
In an embodiment, the first determining unit is specifically configured to process the blood vessel mask image by using a segmented image of an organ, so as to obtain a main blood vessel of an arterial blood vessel and a main blood vessel of a venous blood vessel; determining a connected domain corresponding to a main vessel of the artery and vein in the arteriovenous distance map according to the main vessel of the artery and vein, and determining a point with the maximum distance value in the connected domain corresponding to the main vessel of the artery and vein as a starting point of the artery and vein; determining a connected domain corresponding to the main vessel of the vein vessel in the arteriovenous distance map according to the main vessel of the vein vessel, and determining a point with the maximum distance value in the connected domain corresponding to the main vessel of the vein vessel as a starting point of the vein vessel.
In one embodiment, the vessel segmentation apparatus further comprises:
the determining module is used for determining the bifurcation point of each blood vessel in the blood vessel segmentation image according to a shortest path searching method;
and the grading module is used for grading the blood vessel sections between the adjacent bifurcation points in each blood vessel in a subsection mode according to a preset subsection grading rule to obtain a structural chart of each blood vessel.
The apparatus for determining a target object provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
The modules in the blood vessel segmentation device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, comprising a memory in which a computer program is stored and a processor, which when executing the computer program performs the steps of the method in any of the above embodiments.
In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the method of any of the above embodiments.
In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, performs the steps of the method of any of the above embodiments.
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 for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include a Read-Only Memory (ROM), a magnetic tape, a floppy disk, a flash Memory, an optical Memory, a high-density embedded nonvolatile Memory, a resistive Random Access Memory (ReRAM), a Magnetic Random Access Memory (MRAM), a Ferroelectric Random Access Memory (FRAM), a Phase Change Memory (PCM), a graphene Memory, and the like. Volatile Memory can include Random Access Memory (RAM), external cache Memory, and the like. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), for example. The databases referred to in various embodiments provided herein may include at least one of relational and non-relational databases. The non-relational database may include, but is not limited to, a block chain based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing based data processing logic devices, etc., without limitation.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.
Claims (10)
1. A method of vessel segmentation, the method comprising:
obtaining a blood vessel mask image in an initial medical image according to the initial medical image and a preset segmentation model;
inputting the initial medical image and the blood vessel mask image into a preset regression model to obtain an arteriovenous distance map; the arteriovenous distance map is used for representing the distance between each arterial blood vessel point and an arterial blood vessel boundary in the initial medical image and the distance between each venous blood vessel point and a venous blood vessel boundary in the initial medical image;
and obtaining a blood vessel segmentation image corresponding to the initial medical image according to the blood vessel mask image, the arteriovenous distance image and a preset fusion model.
2. The method according to claim 1, wherein obtaining a vessel segmentation image corresponding to the initial medical image according to the vessel mask image, the arteriovenous distance map and a preset fusion model comprises:
determining a target path of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance image;
and inputting the blood vessel mask image, the arteriovenous distance map and the target path of each blood vessel into a preset fusion model to obtain a blood vessel segmentation image corresponding to the initial medical image.
3. The method of claim 2, wherein determining the target path for each vessel in the vessel mask image from the vessel mask image and the arteriovenous distance map comprises:
determining the starting point of each blood vessel in the blood vessel mask image according to the blood vessel mask image and the arteriovenous distance image;
determining the termination point of each blood vessel in the blood vessel mask image according to the blood vessel mask image;
normalizing each distance value in the arteriovenous distance map to obtain a processed distance map;
determining a path between the starting point of each blood vessel and the ending point of each blood vessel according to the processed distance map and a shortest path search algorithm;
and determining a target path of each blood vessel according to the path between the starting point of each blood vessel and the ending point of each blood vessel.
4. The method of claim 3, wherein the starting points of each of the vessels comprise a starting point of a venous vessel and a starting point of an arterial vessel; the determining a path between the starting point of each blood vessel and the ending point of each blood vessel according to the processed distance map and the shortest path search algorithm comprises:
for the termination point of each blood vessel, adopting the shortest path search algorithm to search a path between the termination point of each blood vessel and the starting point of the vein blood vessel and determine the path as a first path of each blood vessel;
and the number of the first and second groups,
and for the termination point of each blood vessel, searching a path between the starting point of the artery blood vessel and the termination point of each blood vessel by adopting the shortest path searching algorithm, and determining the path as a second path of each blood vessel.
5. The method of claim 4, wherein determining a target path for each of the blood vessels based on a path between a starting point of each of the blood vessels and a terminating point of each of the blood vessels comprises:
performing derivation on the first path of each blood vessel to obtain a first slope corresponding to the first path;
performing derivation on the second path of each blood vessel to obtain a second slope corresponding to the second path;
and determining the target path of each blood vessel by using the first slope, the second slope and a preset slope threshold.
6. The method of claim 3, wherein the segmentation model comprises a first network model and a second network model; the obtaining of the blood vessel mask image in the initial medical image according to the initial medical image and a preset segmentation model comprises:
inputting the initial medical image into the first network model for rough segmentation to obtain a segmented image of an organ in the initial medical image;
according to the segmented image of the organ, determining the position of the organ in the initial medical image, and acquiring a minimum circumscribed cube of the organ;
inputting the minimum circumscribed cube into the second network model for fine segmentation to obtain the blood vessel mask image; wherein a sampling resolution of the second network model is less than a sampling resolution of the first network model.
7. The method of claim 6, wherein determining the starting point of each blood vessel in the blood vessel mask image from the blood vessel mask image and the arteriovenous distance map comprises:
processing the blood vessel mask image by utilizing the segmented image of the organ to obtain a main blood vessel of the arterial blood vessel and a main blood vessel of the venous blood vessel;
determining a connected domain corresponding to the main artery of the artery and vein distance map according to the main artery of the artery and vein distance map, and determining a point with the maximum distance value in the connected domain corresponding to the main artery of the artery and vein distance map as a starting point of the artery and vein;
determining a connected domain corresponding to the main vessel of the vein vessel in the arteriovenous distance map according to the main vessel of the vein vessel, and determining a point with the maximum distance value in the connected domain corresponding to the main vessel of the vein vessel as a starting point of the vein vessel.
8. The method according to any one of claims 1-7, further comprising:
determining a bifurcation point of each blood vessel in the blood vessel segmentation image according to a shortest path search method;
and according to a preset segmentation grading rule, carrying out segmentation grading on the blood vessel segments between adjacent bifurcation points in each blood vessel to obtain a structural diagram of each blood vessel.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 9 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210111422.0A CN114445412A (en) | 2022-01-29 | 2022-01-29 | Blood vessel segmentation method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210111422.0A CN114445412A (en) | 2022-01-29 | 2022-01-29 | Blood vessel segmentation method, device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114445412A true CN114445412A (en) | 2022-05-06 |
Family
ID=81372469
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210111422.0A Pending CN114445412A (en) | 2022-01-29 | 2022-01-29 | Blood vessel segmentation method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114445412A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115953636A (en) * | 2023-03-10 | 2023-04-11 | 南京博视医疗科技有限公司 | Blood vessel grading method, blood vessel tortuosity calculation method and device |
| CN117373070A (en) * | 2023-12-07 | 2024-01-09 | 瀚依科技(杭州)有限公司 | Method and device for labeling blood vessel segments, electronic equipment and storage medium |
-
2022
- 2022-01-29 CN CN202210111422.0A patent/CN114445412A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115953636A (en) * | 2023-03-10 | 2023-04-11 | 南京博视医疗科技有限公司 | Blood vessel grading method, blood vessel tortuosity calculation method and device |
| CN117373070A (en) * | 2023-12-07 | 2024-01-09 | 瀚依科技(杭州)有限公司 | Method and device for labeling blood vessel segments, electronic equipment and storage medium |
| CN117373070B (en) * | 2023-12-07 | 2024-03-12 | 瀚依科技(杭州)有限公司 | Method and device for labeling blood vessel segments, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7086336B2 (en) | Tissue nodule detection method and its model Training method, equipment, equipment, system, and its computer program | |
| Usman et al. | Volumetric lung nodule segmentation using adaptive roi with multi-view residual learning | |
| CN111145206B (en) | Liver image segmentation quality assessment method and device and computer equipment | |
| CN111696089B (en) | Arteriovenous determination method, device, equipment and storage medium | |
| WO2021238438A1 (en) | Tumor image processing method and apparatus, electronic device, and storage medium | |
| CN114445412A (en) | Blood vessel segmentation method, device and storage medium | |
| CN110796670B (en) | Dissection method and device for dissecting interbed artery | |
| CN111768411A (en) | Coronary artery central line extraction method and device, computer equipment and storage medium | |
| CN109360209B (en) | Coronary vessel segmentation method and system | |
| CN111932554A (en) | Pulmonary blood vessel segmentation method, device and storage medium | |
| CN112348818B (en) | Image segmentation method, device, equipment and storage medium | |
| CN112102275B (en) | Pulmonary aortic vessel image extraction method, device, storage medium and electronic equipment | |
| CN112308846A (en) | Blood vessel segmentation method and device and electronic equipment | |
| CN107507212A (en) | Digital brain method for visualizing, device, computing device and storage medium | |
| CN116051738A (en) | Method for reconstructing coronary artery blood vessel model based on CTA image and readable storage medium | |
| CN109658377B (en) | Breast MRI lesion area detection method based on multidimensional information fusion | |
| WO2024021641A1 (en) | Blood vessel segmentation method and apparatus, device, and storage medium | |
| CN117474930A (en) | Combined segmentation method and system for left atrium and left auricle tissue of CT image | |
| US20230222668A1 (en) | Image processing apparatus, image processing method, and recording medium | |
| CN116468103A (en) | Training method, application method and system for lung nodule benign and malignant recognition model | |
| CN113379741B (en) | Retinal blood vessel segmentation method, device and storage medium based on blood vessel characteristics | |
| CN114081625B (en) | Navigation path planning method, system and readable storage medium | |
| CN114998582A (en) | Coronary artery blood vessel segmentation method, device and storage medium | |
| CN115546089A (en) | Medical image segmentation method, pathological image processing method, device and equipment | |
| Tang et al. | SM-RNet: A Scale-aware-based Multi-attention Guided Reverse Network for Pulmonary Nodules Segmentation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |