CN118154591B - Method, device, equipment and storage medium for detecting intracranial large vessel occlusion point - Google Patents
Method, device, equipment and storage medium for detecting intracranial large vessel occlusion point Download PDFInfo
- Publication number
- CN118154591B CN118154591B CN202410568034.4A CN202410568034A CN118154591B CN 118154591 B CN118154591 B CN 118154591B CN 202410568034 A CN202410568034 A CN 202410568034A CN 118154591 B CN118154591 B CN 118154591B
- Authority
- CN
- China
- Prior art keywords
- intracranial
- angle
- image
- occlusion
- vessel
- 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.)
- Active
Links
- 238000007917 intracranial administration Methods 0.000 title claims abstract description 193
- 238000000034 method Methods 0.000 title claims abstract description 38
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 47
- 230000000903 blocking effect Effects 0.000 claims description 58
- 238000001514 detection method Methods 0.000 claims description 38
- 239000013598 vector Substances 0.000 claims description 21
- 238000004590 computer program Methods 0.000 claims description 14
- 238000012216 screening Methods 0.000 claims description 11
- 210000003657 middle cerebral artery Anatomy 0.000 claims description 9
- 210000003388 posterior cerebral artery Anatomy 0.000 claims description 9
- 210000001841 basilar artery Anatomy 0.000 claims description 8
- 210000001715 carotid artery Anatomy 0.000 claims description 8
- 210000002385 vertebral artery Anatomy 0.000 claims description 8
- 210000001367 artery Anatomy 0.000 claims description 6
- 238000003709 image segmentation Methods 0.000 claims description 4
- 210000004129 prosencephalon Anatomy 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000002792 vascular Effects 0.000 claims description 2
- 206010053648 Vascular occlusion Diseases 0.000 claims 1
- 208000021331 vascular occlusion disease Diseases 0.000 claims 1
- 238000012549 training Methods 0.000 description 13
- 210000002551 anterior cerebral artery Anatomy 0.000 description 6
- 210000003625 skull Anatomy 0.000 description 4
- 208000032382 Ischaemic stroke Diseases 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 201000006474 Brain Ischemia Diseases 0.000 description 2
- 206010008120 Cerebral ischaemia Diseases 0.000 description 2
- 208000031481 Pathologic Constriction Diseases 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 206010008118 cerebral infarction Diseases 0.000 description 2
- 239000002872 contrast media Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 208000037804 stenosis Diseases 0.000 description 2
- 230000036262 stenosis Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- ZCYVEMRRCGMTRW-UHFFFAOYSA-N 7553-56-2 Chemical compound [I] ZCYVEMRRCGMTRW-UHFFFAOYSA-N 0.000 description 1
- 206010058178 Aortic occlusion Diseases 0.000 description 1
- 208000002381 Brain Hypoxia Diseases 0.000 description 1
- 208000006011 Stroke Diseases 0.000 description 1
- 208000007536 Thrombosis Diseases 0.000 description 1
- 150000001450 anions Chemical class 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 150000001768 cations Chemical class 0.000 description 1
- 230000002490 cerebral effect Effects 0.000 description 1
- 230000004087 circulation Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000007954 hypoxia Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000010253 intravenous injection Methods 0.000 description 1
- 229910052740 iodine Inorganic materials 0.000 description 1
- 239000011630 iodine Substances 0.000 description 1
- 239000004973 liquid crystal related substance Substances 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
- 230000001575 pathological effect Effects 0.000 description 1
- 230000010412 perfusion Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000002537 thrombolytic effect Effects 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Geometry (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
Abstract
The application relates to a method, a device, equipment and a storage medium for detecting intracranial large vessel occlusion points, which are characterized in that a plurality of main angle MIP images and a plurality of auxiliary angle MIP images are obtained by carrying out a plurality of main angle projections on an intracranial vessel tree 3D image, a plurality of predicted occlusion point coordinates are obtained based on each MIP image, the predicted occlusion points on the main angle MIP image and the corresponding plurality of auxiliary angle MIP images are combined for each main angle to obtain corresponding accurate occlusion point coordinates and occlusion point squares, and on the intracranial vessel tree 3D image, all the occlusion point squares corresponding to the vessel category on two main angles specified according to the vessel section category extend into a cuboid along the projection angles and obtain an intersecting cuboid, and the intracranial large vessel occlusion points are obtained according to the three-dimensional centroid coordinates of the intersecting cuboid. The method can be used for accurately and reliably detecting the intracranial large blood vessel occlusion point based on CTA data.
Description
Technical Field
The application relates to the technical field of medical image analysis, in particular to a method, a device, equipment and a storage medium for detecting an intracranial large vessel occlusion point.
Background
Intracranial large vessel occlusion (LARGE VESSEL Occlusions, LVO) refers to a pathological condition of cerebral ischemia and hypoxia caused by stenosis or occlusion of main vessels such as intracranial carotid artery, middle cerebral artery, anterior cerebral artery, vertebral artery, basilar artery or posterior cerebral artery. Intracranial macrovascular occlusion is a common cause of ischemic stroke, accounting for over about 40% of all ischemic strokes. Is also an important cause of recurrence of ischemic stroke, and the recurrence risk of the stroke is 3.6% -22.0%.
The intracranial CTA image is obtained by intravenous injection of iodine-containing contrast agent, continuous original data acquisition is carried out in peak period of filling of contrast agent in target blood vessel of a subject by using spiral CT, and then target blood vessel is reconstructed by using post-processing function of a computer to form three-dimensional displayed blood vessel image. The method has an important effect on screening of large vessel occlusion, can rapidly and accurately detect the position, degree, range and side branch circulation condition of intracranial large artery occlusion, and provides important information for clinical diagnosis and treatment. It can also help select patients suitable for intravascular treatment, such as thrombolysis, stents, etc., to restore blood flow and alleviate cerebral ischemia. And can evaluate the effect of intravascular treatment, such as thrombus removal, vascular patency, cerebral tissue perfusion, etc., to guide subsequent treatment and prognosis.
However, the judgment of intracranial macrovascular occlusion is a challenging task, requiring a combination of imaging and clinical factors. Some techniques use a determination based on changes in blood vessel density, and it is thought that areas of reduced blood vessel density may be sites of occlusion. However, this technique has certain limitations in that the cause of the reduced vessel density may also be stenosis rather than occlusion, or artifacts due to vessel bending, bifurcation, overlapping, etc. Moreover, this technique cannot determine the specific location of the occlusion, can only approximate the extent of the occlusion, and cannot provide accurate anatomical information.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, apparatus, device, and storage medium for detecting an intracranial large vessel occlusion point, which can improve the accuracy and reliability of determination of the intracranial large vessel occlusion.
A method of intracranial large vessel occlusion point detection, the method comprising:
acquiring intracranial CTA image data, and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
performing multiple main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, and rotating each main angle vector for multiple times in a random direction according to a preset angle to obtain a plurality of sub-angle MIP images corresponding to each main angle MIP image;
Inputting each MIP image into a trained two-dimensional occlusion point detection model to obtain predicted occlusion point coordinates in each MIP image;
Aiming at each main angle, combining the main angle MIP image and the predicted blocking points on the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points to obtain accurate blocking point coordinates on each main angle;
constructing and obtaining a square block point by taking each accurate block point coordinate as a center according to a preset side length;
According to the category of the intracranial vessel segment, extending all the squares of the occlusion points corresponding to the vessel category on the two main angles designated by the category into cuboid along the projection angle of the squares, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
In one embodiment, the generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data comprises:
Dividing the intracranial CTA image data by adopting a trained image segmentation model to obtain an intracranial 3D vessel tree binary image;
Multiplying the intracranial CTA image data with the intracranial 3D vessel tree binary image to obtain the intracranial vessel tree 3D image.
In one embodiment, the 3D image of the intracranial vessel tree is projected at a plurality of different angles in an anatomical coordinate system to obtain a MIP image;
Projecting the intracranial vessel tree 3D image in the anatomical coordinate system towards four main angle projection directions;
the four principal angle projection vectors are respectively: (1, -1, 0), (1, 1, 0), (0, 1, 1), (0, 1, -1).
In one embodiment, for each main angle, calculating the euclidean distance between two predicted occlusion points, and combining the predicted occlusion points on the main angle MIP image and the corresponding multiple sub-angle MIP images to obtain the accurate occlusion point coordinates corresponding to each main angle includes:
sequencing the predicted occlusion point coordinates on the main angle MIP image and the corresponding multiple auxiliary angle MIP images from high to low according to the confidence level aiming at each main angle;
Sequentially calculating Euclidean distances between the predicted occlusion point coordinates and all the predicted occlusion point coordinates higher than the confidence coefficient according to the confidence coefficient from high to low, and screening out the predicted occlusion point coordinates if the calculated result is smaller than a preset value;
And screening out the residual predicted blocking point coordinates, namely the accurate blocking point coordinates corresponding to the main angle.
In one embodiment, when the occlusion square is constructed with each accurate occlusion point coordinate as a center according to a preset side length: dividing the intracranial arterial vessel segment into two main categories, and constructing the occlusion point square by adopting the corresponding side length according to the main category of the arterial segment where the accurate occlusion point coordinate is located.
In one embodiment, the extending all the squares of the occlusion points corresponding to the blood vessel category along the projection angle of the squares of the two main angles specified by the category into a cuboid on the 3D image of the intracranial blood vessel tree according to the category of the intracranial blood vessel segment comprises:
when the blood vessel segment is classified as an intracranial carotid artery or a forebrain artery, the vector of the specified projection direction is (1, -1, 0), (1, 1, 0) two main angles;
when the blood vessel segment is classified into one of the vertebral artery, basilar artery, middle cerebral artery and posterior cerebral artery, the vector specifying the projection direction is (0, 1, 1), (0, 1, -1) two principal angles.
In one embodiment, the obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid includes:
and removing the intersected cuboids with the shortest side length smaller than a preset value from the intersected cuboids, wherein the three-dimensional centroid coordinates of the remaining intersected cuboids are intracranial large vessel occlusion points on the intracranial CTA image data.
The application also provides an intracranial large vessel occlusion point detection device, which comprises:
the intracranial vessel tree 3D image generation module is used for acquiring intracranial CTA image data and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
the multi-angle MIP image generation module is used for carrying out multi-main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, respectively rotating each main angle vector to a random direction for a plurality of times according to a preset angle to obtain a plurality of auxiliary angle MIP images corresponding to each main angle MIP image;
The predicted blocking point coordinate obtaining module is used for inputting each MIP image into the trained two-dimensional blocking point detection model to obtain the predicted blocking point coordinate in each MIP image;
the accurate blocking point coordinate obtaining module is used for combining the main angle MIP images and the predicted blocking points on the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points for each main angle to obtain the accurate blocking point coordinate corresponding to each main angle;
The block point square obtaining module is used for obtaining block point squares by taking each accurate block point coordinate as a center and constructing according to a preset side length;
the intracranial large blood vessel occlusion point obtaining module is used for extending all the occlusion point squares corresponding to the blood vessel category on two main angles designated by the category into cuboid along the projection angle of the squares according to the category of the intracranial blood vessel segment on the 3D image of the intracranial blood vessel tree, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large blood vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
A computer device comprising a memory storing a computer program and a processor which when executing the computer program performs the steps of:
acquiring intracranial CTA image data, and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
performing multiple main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, and rotating each main angle vector for multiple times in a random direction according to a preset angle to obtain a plurality of sub-angle MIP images corresponding to each main angle MIP image;
Inputting each MIP image into a trained two-dimensional occlusion point detection model to obtain predicted occlusion point coordinates in each MIP image;
Aiming at each main angle, combining the main angle MIP image and the predicted blocking points on the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points to obtain accurate blocking point coordinates on each main angle;
constructing and obtaining a square block point by taking each accurate block point coordinate as a center according to a preset side length;
According to the category of the intracranial vessel segment, extending all the squares of the occlusion points corresponding to the vessel category on the two main angles designated by the category into cuboid along the projection angle of the squares, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
acquiring intracranial CTA image data, and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
performing multiple main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, and rotating each main angle vector for multiple times in a random direction according to a preset angle to obtain a plurality of sub-angle MIP images corresponding to each main angle MIP image;
Inputting each MIP image into a trained two-dimensional occlusion point detection model to obtain predicted occlusion point coordinates in each MIP image;
Aiming at each main angle, combining the main angle MIP image and the predicted blocking points on the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points to obtain accurate blocking point coordinates on each main angle;
constructing and obtaining a square block point by taking each accurate block point coordinate as a center according to a preset side length;
According to the category of the intracranial vessel segment, extending all the squares of the occlusion points corresponding to the vessel category on the two main angles designated by the category into cuboid along the projection angle of the squares, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
According to the intracranial large vessel occlusion point detection method, the device, the equipment and the storage medium, the intracranial large vessel occlusion point 3D image is generated based on intracranial CTA image data, a plurality of corresponding main angle MIP images are obtained through a plurality of main angle projections, a plurality of auxiliary angle MIP images are obtained through a plurality of times of rotation in random directions according to each main angle vector, each MIP image is input into a trained two-dimensional occlusion point detection model to obtain predicted occlusion point coordinates in each MIP image, the main angle MIP image and the predicted occlusion points on the corresponding plurality of auxiliary angle MIP images are combined for each main angle to obtain accurate occlusion point coordinates corresponding to each main angle, and an occlusion point square is correspondingly constructed, according to the category of the intracranial vessel segment, all the occlusion point squares corresponding to the vessel category on the two main angles designated by the category extend into cuboid along the projection angle, each cuboid is intersected, and the intracranial large vessel occlusion point on the intracranial CTA image data is obtained according to the three-dimensional centroid of the cuboid. By adopting the method, the intracranial large vessel occlusion point on the intracranial CTA image data can be accurately and reliably detected.
Drawings
FIG. 1 is a flow chart of a method for detecting an intracranial large vessel occlusion in an embodiment;
FIG. 2 is a graph showing MIP display results at different angles after the MCA vessel segment occlusion is rotated along the L-R axis according to one embodiment;
FIG. 3 is a graph showing MIP presentation results at various angles after ICA segment occlusion is rotated along the S-I axis in one embodiment;
FIG. 4 is a schematic representation of the results of a 3D presentation of BA vessel segment occlusion in one embodiment;
FIG. 5 is a schematic view showing the results of 3D display of ACA vessel segment occlusion in one embodiment
FIG. 6 is a block diagram of an embodiment of an intracranial large vessel occlusion point detection device;
Fig. 7 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
Aiming at the problems of low accuracy and low reliability of the intracranial large blood vessel occlusion point detection method in the prior art, as shown in fig. 1, the application provides the intracranial large blood vessel occlusion point detection method, which comprises the following steps:
Step S100, acquiring intracranial CTA image data, and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data.
Step S110, performing multiple principal angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding principal angle MIP images, and rotating each principal angle vector to a random direction for multiple times according to a preset angle to obtain a plurality of sub-angle MIP images corresponding to each principal angle MIP image.
Step S120, inputting each MIP image into the trained two-dimensional occlusion point detection model to obtain the predicted occlusion point coordinates in each MIP image.
Step S130, for each main angle, calculating the Euclidean distance between two predicted blocking points, and combining the predicted blocking points on the main angle MIP image and the corresponding multiple auxiliary angle MIP images to obtain the accurate blocking point coordinates corresponding to each main angle.
And step S140, constructing and obtaining a square of the blocking point according to a preset side length by taking the coordinates of each accurate blocking point as the center.
And S150, extending all the squares of the occlusion points corresponding to the blood vessel category on the two main angles designated by the category into cuboid along the projection angle according to the category of the intracranial blood vessel section on the 3D image of the intracranial blood vessel tree, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large blood vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
In the embodiment, the method based on AI vessel tree segmentation and multi-angle MIP, namely maximum density projection detection is adopted, so that the accuracy is ensured, the reliability is improved, and the calculation time is saved. The method avoids the problems that when the traditional MIP method is adopted, the original CTA image is required to be subjected to skull removal, then the image after skull removal is subjected to MIP treatment, the skull removal accuracy requirement is high, if the skull removal is inaccurate, the distortion of a blood vessel image is caused, and the LVO detection accuracy is affected. Meanwhile, the method has little requirement on calculation time and cost, does not need powerful hardware and algorithm support, and can realize quick LVO detection.
In step S100, related intracranial CTA image data is acquired based on the target, and the intracranial CTA image data is segmented by using a trained image segmentation model, so as to obtain an intracranial 3D vessel tree binary image. And multiplying the intracranial CTA image data with the intracranial 3D vessel tree binary image to obtain the intracranial vessel tree 3D image.
In this embodiment, the generation of the intracranial 3D vessel tree binary image based on the intracranial CTA image data may be generated by a method including threshold segmentation, region growing, or morphological processing, in addition to the image segmentation model.
In this embodiment, three-dimensional intracranial vessel tree 3D images are projected with maximum intensity in different directions, respectively, to obtain two-dimensional MIP images. And LVO detection is performed based on the two-dimensional MIP image.
In step S110, when the intracranial vessel tree 3D image is projected at different angles in the anatomical coordinate system to obtain the MIP image, the intracranial vessel tree 3D image is projected at four principal angles, and the vectors of the projection directions are respectively: (1, -1, 0), (1, 1, 0), (0, 1, 1), (0, 1, -1).
It should be noted that, according to the specific application scenario of the method, 3 different angles, or more, of projections may be performed on the 3D image of the intracranial vessel tree, and a corresponding number of MIP images with a main angle may be obtained.
In this embodiment, the anatomical coordinate system has an L (left) direction as an x-axis positive direction, an R (right) direction as an x-axis negative direction, an a (anti) direction as a y-axis positive direction, a P (post) direction as a y-axis negative direction, an S (super) direction as a z-axis positive direction, and an I (refer) direction as a z-axis negative direction.
Considering the problem that the main angle MIP images can possibly detect the number of false anions and false cations caused by blood vessel occlusion, in the embodiment, a plurality of corresponding auxiliary angles are selected randomly according to each main angle MIP image to assist in determining LVO, so that the diversity of data is ensured, and the accuracy of LVO detection is improved.
Specifically, each main angle vector is randomly rotated by a certain angle. This operation was repeated three times, with angles of rotation of 5 degrees, 10 degrees, 15 degrees, respectively. Thus, a total of 16 MIP images with 4 principal angles and 12 random secondary angles can be obtained from each intracranial vessel tree 3D image.
Here, in the case of generating a plurality of sub-angles based on the main angle, the angle of each rotation is a fixed preset angle, which may be set to 5 ° or 10 ° or the like, and the direction of each rotation is random.
Similarly, according to a specific application scenario of the method, each main angle can generate any number of auxiliary angles according to the customized angle, so as to meet the requirement of auxiliary determination of LVO in the scenario.
In step S120, a trained two-dimensional occlusion point detection model is used to predict the occlusion points in each MIP image, so as to obtain a plurality of predicted occlusion point coordinates. Here, the predicted occlusion point coordinates are the center point coordinates of the LVO.
In this embodiment, for each main angle, a plurality of predicted occlusion points obtained by the main angle and the sub angle are combined to determine an accurate occlusion point.
Specifically, for each main angle, by calculating the euclidean distance between two predicted occlusion points, combining the predicted occlusion points on the main angle MIP image and the corresponding multiple auxiliary angle MIP images, and obtaining accurate occlusion point coordinates corresponding to each main angle includes: and sequencing the predicted blocking point coordinates on the main angle MIP image and the corresponding multiple auxiliary angle MIP images according to the confidence coefficient, sequentially calculating Euclidean distances between the predicted blocking point coordinates and all the predicted blocking point coordinates higher than the confidence coefficient according to the confidence coefficient from high to low, screening out the predicted blocking point coordinates if the calculated result is smaller than a preset value, and screening out the residual predicted blocking point coordinates after screening out, namely the accurate blocking point coordinates corresponding to the main angle.
In this embodiment, when the predicted occlusion point coordinates are ordered, the predicted occlusion point coordinates corresponding to the main angle MIP image are ordered first, and then the predicted occlusion point coordinates corresponding to the auxiliary angle are arranged in parallel behind all the predicted occlusion point coordinates according to the confidence level.
In this embodiment, when the predicted occlusion points are combined, the euclidean distance between the predicted occlusion point with the highest confidence after sorting and other occlusion points is considered, and when the euclidean distance between the predicted occlusion points is greater than a certain preset value, it is indicated that another predicted occlusion point is not the same occlusion point with the highest confidence in the main angle, so that the predicted occlusion point is reserved. When the distance between the two is smaller than a certain preset value, the predicted occlusion point is most likely to be the same as the predicted occlusion point with the highest confidence, so that the predicted occlusion point is removed. The last remaining point is the exact occlusion point corresponding to the principal angle.
Further, since the common blood vessel diameters of the blood vessel sections are different, different preset values are adopted for judging whether to screen out according to the type of the blood vessel section where the predicted occlusion point is located. The classes of intracranial arterial vessel segments include: intracranial carotid artery, anterior cerebral artery, vertebral artery, basilar artery, middle cerebral artery and posterior cerebral artery. When the predicted occlusion points to be screened are positioned on the intracranial carotid artery, the vertebral artery and the basilar artery, the same preset value is adopted to judge whether to screen out. When the predicted occlusion points to be screened are positioned on the anterior cerebral artery, the middle cerebral artery and the posterior cerebral artery, judging whether to screen out or not by adopting another same preset value.
Specifically, when the predicted occlusion point to be screened is located on the intracranial carotid artery, vertebral artery and basilar artery, the preset value is 25 mm, and screening judgment is performed. When the predicted occlusion point to be screened is positioned on the anterior cerebral artery, the middle cerebral artery and the posterior cerebral artery, the preset value is 15 mm, and screening judgment is carried out.
Here, when the preset value for comparison is set for the screening determination, the adjustment may be performed according to the application scenario, and is not limited to the above two specific values, and 25 mm and 15 mm are only preferred values.
In step S130, when the occlusion square is constructed with each accurate occlusion point coordinate as a center and according to a preset side length: and the common blood vessel diameters of the blood vessel sections are different, so that the intracranial arterial blood vessel sections are classified into two main categories, and the occlusion point square is constructed by adopting the corresponding preset side length according to the main category of the arterial section where the accurate occlusion point coordinate is located.
In one embodiment, when the artery segment where the accurate occlusion point coordinates are located is an intracranial carotid artery, a vertebral artery, and a basilar artery, the square is constructed according to a preset side length of 12.5 mm. When the artery segment where the accurate occlusion point coordinate is located is on the anterior cerebral artery, the middle cerebral artery and the posterior cerebral artery, square construction is carried out according to the preset side length of 7.5 mm.
It should be noted that the specific values of the side length of the square of the blocking point can be adjusted according to the application scenario, and are not limited to the two specific values, and 12.5 mm and 7.5 mm are only preferred values.
In step S140, for different vessel segment categories, the final intracranial aortic occlusion point position is determined using all the exact occlusion point coordinates at the two principal angles specified for the category.
Specifically, according to the category of the intracranial vessel segment, on the intracranial vessel tree 3D image, extending all the squares of the occlusion points corresponding to the vessel category on the two main angles specified by the category into a cuboid along the projection angle thereof comprises: when the vessel segment class is an intracranial carotid artery or a forebrain artery, the vector specifying the projection direction is (1, -1, 0), (1, 1, 0) two principal angles. When the blood vessel segment is one of the vertebral artery, basilar artery, middle cerebral artery and posterior cerebral artery, the vector of the projection direction is designated as (0, 1, 1), (0, 1, -1) two principal angles.
Specifically, obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid comprises: and removing the intersected cuboids with the shortest side length smaller than the preset value, wherein the three-dimensional centroid coordinates of the remaining intersected cuboids are intracranial large vessel occlusion points on intracranial CTA image data.
Furthermore, on the intracranial vessel tree 3D image, extending all LVO square frames of the vessel segment class in two main angles taken by each vessel segment class into cuboid along the projection angle, and intersecting all the cuboids to obtain small intersecting cuboid. And calculating the shortest side length of all the intersecting cuboids, if the shortest side length is larger than a certain value, reserving the intersecting cuboids, and otherwise, discarding the intersecting cuboids. And calculating the three-dimensional centroid coordinates of all the reserved intersecting cuboids to obtain the occlusion center points of all LVOs of the intracranial vessel tree 3D image.
Specifically, when judging whether the intersecting cuboid needs to be reserved according to the shortest side length of the intersecting cuboid, the same judgment is performed by adopting different preset values according to the blood vessel section type. When the vessel segment categories are intracranial carotid, vertebral, and basilar, then the preset value is set to 6.25 millimeters. When the blood vessel segment categories are anterior, middle and posterior cerebral arteries, the preset value is set to 3.75 millimeters.
It should be noted that, the preset value is set to 6.25 mm and 3.75 mm as the preferred scheme, and other suitable data can be selected according to different application scenarios under other application scenarios.
In this embodiment, the two-dimensional occlusion point detection model adopts yolo model, and when training is performed, a corresponding intracranial vessel tree 3D training image is generated by using the intracranial CTA training image for training, and the three-dimensional coordinates of the occlusion center point of each LVO are marked on the 3D training image. And generating a plurality of MIP training images by using the intracranial vessel tree 3D training image according to the step S110, respectively projecting three-dimensional coordinates of the blocking center points marked on the 3D training image onto the corresponding MIP training images according to the projection angles of the MIP training images to obtain two-dimensional coordinates of the blocking center points in each MIP training image, and generating a LVO square frame by taking the two-dimensional coordinates as the center, wherein the length of the side length of the square is consistent with that in the step S130.
Further, training is performed according to the MIP training image and the corresponding square label frame two-dimensional occlusion point detection model, and the output is the center point coordinates of the LVO square frame, the blood vessel segment type where the LVO square frame is located and the confidence of the LVO square frame.
As shown in fig. 2 and 3, multi-angle MIP images marked with two-dimensional intracranial large vessel occlusion points on different vessel segment categories, and as shown in fig. 4 and 5, the 3D display results of the intracranial large vessel occlusion points detected on the different vessel segment categories by the method are shown.
In the method for detecting the intracranial large blood vessel occlusion point, for the problem of intracranial CTA image large blood vessel occlusion detection, the method based on AI blood vessel tree segmentation and multi-angle MIP target detection is adopted, the accuracy is ensured, the output is a 3D occlusion detection result, and the reasoning time is saved. The method of three-dimensional extension intersection combination of the multi-angle vessel tree MIP images is adopted to detect LVO, and three-dimensional LVO detection results can be generated by detecting a plurality of two-dimensional images. Furthermore, a multi-angle selection mode of the vessel tree MIP image with four main angles and three random auxiliary angles is adopted. The number of detection false negative and false positive caused by the occlusion of the blood vessel at the main angle is reduced. And when the detection model is trained, the random selection of three auxiliary angles also ensures the data diversity. For each primary angle, its final LVO detection box is ordered according to its secondary angle and repeated boxes pruned according to confidence. For each vessel segment, the two most clear main angles are adopted for three-dimensional extending, intersecting and combining of the LVO detection frames. And judging whether the position is reserved as an LVO detection result according to the length of the shortest side of the intersected cuboid.
It should be understood that, although the steps in the flowchart of fig. 1 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 strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in fig. 1 may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor do the order in which the sub-steps or stages are performed necessarily performed in sequence, but may be performed alternately or alternately with at least a portion of other steps or sub-steps of other steps.
In one embodiment, as shown in FIG. 6, there is provided an intracranial large vessel occlusion point detection device, comprising: an intracranial vessel tree 3D image generation module 200, a multi-angle MIP image generation module 210, a predicted occlusion point coordinate acquisition module 220, an accurate occlusion point coordinate acquisition module 230, an occlusion point square acquisition module 240, and an intracranial large vessel occlusion point acquisition module 250, wherein:
the intracranial vessel tree 3D image generation module 200 is used for acquiring intracranial CTA image data and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
The multi-angle MIP image generation module 210 is configured to perform multiple main angle projection on the intracranial vessel tree 3D image to obtain multiple corresponding main angle MIP images, and rotate each main angle vector toward a random direction multiple times according to a preset angle to obtain multiple sub-angle MIP images corresponding to each main angle MIP image;
The predicted occlusion point coordinate obtaining module 220 is configured to input each MIP image into a trained two-dimensional occlusion point detection model, so as to obtain predicted occlusion point coordinates in each MIP image;
the accurate blocking point coordinate obtaining module 230 is configured to, for each main angle, combine the predicted blocking points on the main angle MIP image and the corresponding multiple auxiliary angle MIP images by calculating the euclidean distance between the two predicted blocking points, so as to obtain an accurate blocking point coordinate corresponding to each main angle;
The block point square obtaining module 240 is configured to construct and obtain a block point square according to a preset side length with each accurate block point coordinate as a center;
The intracranial large vessel occlusion point obtaining module 250 is configured to extend, on the intracranial vessel tree 3D image, all the occlusion point squares corresponding to the vessel category on two main angles specified by the category into a cuboid along the projection angle thereof according to the category of the intracranial vessel segment, and obtain an intersecting cuboid by intersecting the cuboids, and obtain the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
Specific limitations regarding the intracranial large vessel occlusion point detection device can be found in the above description of the method for detecting an intracranial large vessel occlusion point, and are not described in detail herein. The various modules in the intracranial large vessel occlusion point detection device described above can be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a terminal, and the internal structure of which may be as shown in fig. 7. The computer device includes a processor, a memory, a network 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 includes a non-volatile 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 the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method for detecting an intracranial large vessel occlusion point. 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, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structure shown in FIG. 7 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:
acquiring intracranial CTA image data, and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
performing multiple main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, and rotating each main angle vector for multiple times in a random direction according to a preset angle to obtain a plurality of sub-angle MIP images corresponding to each main angle MIP image;
Inputting each MIP image into a trained two-dimensional occlusion point detection model to obtain predicted occlusion point coordinates in each MIP image;
Aiming at each main angle, combining the main angle MIP image and the predicted blocking points on the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points to obtain accurate blocking point coordinates on each main angle;
constructing and obtaining a square block point by taking each accurate block point coordinate as a center according to a preset side length;
According to the category of the intracranial vessel segment, extending all the squares of the occlusion points corresponding to the vessel category on the two main angles designated by the category into cuboid along the projection angle of the squares, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring intracranial CTA image data, and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
performing multiple main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, and rotating each main angle vector for multiple times in a random direction according to a preset angle to obtain a plurality of sub-angle MIP images corresponding to each main angle MIP image;
Inputting each MIP image into a trained two-dimensional occlusion point detection model to obtain predicted occlusion point coordinates in each MIP image;
Aiming at each main angle, combining the main angle MIP image and the predicted blocking points on the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points to obtain accurate blocking point coordinates on each main angle;
constructing and obtaining a square block point by taking each accurate block point coordinate as a center according to a preset side length;
According to the category of the intracranial vessel segment, extending all the squares of the occlusion points corresponding to the vessel category on the two main angles designated by the category into cuboid along the projection angle of the squares, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.
Claims (9)
1. A method for detecting an intracranial large vessel occlusion point, the method comprising:
acquiring intracranial CTA image data, and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
performing multiple main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, and rotating each main angle vector for multiple times in a random direction according to a preset angle to obtain a plurality of sub-angle MIP images corresponding to each main angle MIP image;
Inputting each MIP image into a trained two-dimensional occlusion point detection model to obtain predicted occlusion point coordinates in each MIP image;
Combining the predicted blocking points on the main angle MIP images and the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points for each main angle to obtain accurate blocking point coordinates corresponding to each main angle, specifically, sequencing the predicted blocking point coordinates on the main angle MIP images and the corresponding multiple auxiliary angle MIP images according to the confidence level from high to low for each main angle, sequentially calculating the Euclidean distance between the predicted blocking point coordinates and all the predicted blocking point coordinates higher than the confidence level according to the confidence level from high to low, and screening out the predicted blocking point coordinates if the calculated result is smaller than a preset value, wherein the residual predicted blocking point coordinates after screening out are the accurate blocking point coordinates corresponding to the main angle;
constructing and obtaining a square block point by taking each accurate block point coordinate as a center according to a preset side length;
According to the category of the intracranial vessel segment, extending all the squares of the occlusion points corresponding to the vessel category on the two main angles designated by the category into cuboid along the projection angle of the squares, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
2. The method of claim 1, wherein generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data comprises:
Dividing the intracranial CTA image data by adopting a trained image segmentation model to obtain an intracranial 3D vessel tree binary image;
Multiplying the intracranial CTA image data with the intracranial 3D vessel tree binary image to obtain the intracranial vessel tree 3D image.
3. The method for detecting the intracranial large vessel occlusion point according to claim 1, wherein the 3D image of the intracranial vessel tree is projected at a plurality of different angles in an anatomical coordinate system to obtain an MIP image;
Projecting the intracranial vessel tree 3D image in the anatomical coordinate system towards four main angle projection directions;
the four principal angle projection vectors are respectively: (1, -1, 0), (1, 1, 0), (0, 1, 1), (0, 1, -1).
4. The method for detecting an intracranial large vessel occlusion point according to claim 3, wherein when the occlusion point square is constructed with each of the accurate occlusion point coordinates as a center and a predetermined side length: dividing the intracranial arterial vessel segment into two main categories, and constructing the occlusion point square by adopting the corresponding side length according to the main category of the arterial segment where the accurate occlusion point coordinate is located.
5. The method for detecting a large intracranial vascular occlusion point according to claim 3, wherein extending all the occlusion point squares corresponding to the blood vessel category in two main angles designated by the category into a cuboid along the projection angle thereof on the 3D image of the intracranial vascular tree according to the category of the intracranial blood vessel segment comprises:
when the blood vessel segment is classified as an intracranial carotid artery or a forebrain artery, the vector of the specified projection direction is (1, -1, 0), (1, 1, 0) two main angles;
when the blood vessel segment is classified into one of the vertebral artery, basilar artery, middle cerebral artery and posterior cerebral artery, the vector specifying the projection direction is (0, 1, 1), (0, 1, -1) two principal angles.
6. The method of claim 5, wherein obtaining the intracranial macrovascular occlusion point on the intracranial CTA image data from the three-dimensional centroid coordinates of the intersecting cuboid comprises:
and removing the intersected cuboids with the shortest side length smaller than a preset value from the intersected cuboids, wherein the three-dimensional centroid coordinates of the remaining intersected cuboids are intracranial large vessel occlusion points on the intracranial CTA image data.
7. An intracranial macrovascular occlusion point detection device, the device comprising:
the intracranial vessel tree 3D image generation module is used for acquiring intracranial CTA image data and generating a corresponding intracranial vessel tree 3D image based on the intracranial CTA image data;
the multi-angle MIP image generation module is used for carrying out multi-main angle projection on the intracranial vessel tree 3D image to obtain a plurality of corresponding main angle MIP images, respectively rotating each main angle vector to a random direction for a plurality of times according to a preset angle to obtain a plurality of auxiliary angle MIP images corresponding to each main angle MIP image;
The predicted blocking point coordinate obtaining module is used for inputting each MIP image into the trained two-dimensional blocking point detection model to obtain the predicted blocking point coordinate in each MIP image;
The accurate blocking point coordinate obtaining module is used for combining the predicted blocking points on the main angle MIP image and the corresponding multiple auxiliary angle MIP images by calculating the Euclidean distance between the two predicted blocking points for each main angle to obtain the accurate blocking point coordinate corresponding to each main angle;
The block point square obtaining module is used for obtaining block point squares by taking each accurate block point coordinate as a center and constructing according to a preset side length;
the intracranial large blood vessel occlusion point obtaining module is used for extending all the occlusion point squares corresponding to the blood vessel category on two main angles designated by the category into cuboid along the projection angle of the squares according to the category of the intracranial blood vessel segment on the 3D image of the intracranial blood vessel tree, intersecting the cuboids to obtain an intersecting cuboid, and obtaining the intracranial large blood vessel occlusion point on the intracranial CTA image data according to the three-dimensional centroid coordinates of the intersecting cuboid.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410568034.4A CN118154591B (en) | 2024-05-09 | 2024-05-09 | Method, device, equipment and storage medium for detecting intracranial large vessel occlusion point |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410568034.4A CN118154591B (en) | 2024-05-09 | 2024-05-09 | Method, device, equipment and storage medium for detecting intracranial large vessel occlusion point |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118154591A CN118154591A (en) | 2024-06-07 |
| CN118154591B true CN118154591B (en) | 2024-08-02 |
Family
ID=91295060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410568034.4A Active CN118154591B (en) | 2024-05-09 | 2024-05-09 | Method, device, equipment and storage medium for detecting intracranial large vessel occlusion point |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118154591B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115251967A (en) * | 2022-07-19 | 2022-11-01 | 同心智医科技(北京)有限公司 | Method, device, terminal and storage medium for establishing large vessel occlusion recognition model |
| CN115880219A (en) * | 2021-09-29 | 2023-03-31 | 西门子医疗有限公司 | Probability tree tracking and large vessel occlusion detection in medical imaging |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114170134B (en) * | 2021-11-03 | 2023-03-10 | 杭州脉流科技有限公司 | Stenosis assessment method and device based on intracranial DSA image |
| CN114155232A (en) * | 2021-12-08 | 2022-03-08 | 中国科学院深圳先进技术研究院 | Intracranial hemorrhage area detection method and device, computer equipment and storage medium |
| CN114764788B (en) * | 2022-03-29 | 2022-12-16 | 首都医科大学附属北京天坛医院 | Intracranial arterial stenosis detection method and system |
| US11967079B1 (en) * | 2023-02-03 | 2024-04-23 | Qure.Ai Technologies Private Limited | System and method for automatically detecting large vessel occlusion on a computational tomography angiogram |
-
2024
- 2024-05-09 CN CN202410568034.4A patent/CN118154591B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115880219A (en) * | 2021-09-29 | 2023-03-31 | 西门子医疗有限公司 | Probability tree tracking and large vessel occlusion detection in medical imaging |
| CN115251967A (en) * | 2022-07-19 | 2022-11-01 | 同心智医科技(北京)有限公司 | Method, device, terminal and storage medium for establishing large vessel occlusion recognition model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118154591A (en) | 2024-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zeng et al. | Automatic diagnosis based on spatial information fusion feature for intracranial aneurysm | |
| US20230139405A1 (en) | Stenosis assessment method and device based on intracranial DSA imaging | |
| CN111161241B (en) | Liver image identification method, electronic equipment and storage medium | |
| CN111080573B (en) | Rib image detection method, computer device and storage medium | |
| US20210374452A1 (en) | Method and device for image processing, and elecrtonic equipment | |
| CN110717961B (en) | Multi-modal image reconstruction method and device, computer equipment and storage medium | |
| CN110400626B (en) | Image detection method, image detection device, computer equipment and storage medium | |
| CN116681716B (en) | Method, device, equipment and storage medium for dividing intracranial vascular region of interest | |
| CN115082348A (en) | Image reconstruction method and device | |
| US8306354B2 (en) | Image processing apparatus, method, and program | |
| Elaraby et al. | A generalized entropy-based two-phase threshold algorithm for noisy medical image edge detection | |
| CN114092475A (en) | Focal length determining method, image labeling method, device and computer equipment | |
| CN110751187A (en) | Training method of abnormal area image generation network and related product | |
| KR102216455B1 (en) | Apparatus and method for measuring left atrium thickness of heart | |
| WO2021032325A1 (en) | Updating boundary segmentations | |
| Reddy et al. | Anatomical landmark detection using deep appearance-context network | |
| CN118154591B (en) | Method, device, equipment and storage medium for detecting intracranial large vessel occlusion point | |
| CN110738664B (en) | Image positioning method and device, computer equipment and storage medium | |
| CN117036530B (en) | Cross-modal data-based coronary artery fractional flow reserve prediction method and device | |
| US20210342638A1 (en) | X-ray image synthesis from ct images for training nodule detection systems | |
| US10373350B2 (en) | Apparatus and method for CT data reconstruction based on motion compensation | |
| Agafonova et al. | Segmentation of meningiomas in MRI of the brain using deep learning methods | |
| CN116491968A (en) | Acute ischemic stroke collateral circulation assessment method based on single-phase CTA | |
| CN118154590B (en) | Method, device, equipment and storage medium for detecting intracranial large vessel occlusion region | |
| CN114373216A (en) | Eye movement tracking method, device, equipment and storage medium for anterior segment OCTA |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |