CN116485810A - Carotid artery segmentation method, device and equipment based on magnetic resonance image - Google Patents
Carotid artery segmentation method, device and equipment based on magnetic resonance image Download PDFInfo
- Publication number
- CN116485810A CN116485810A CN202310308272.7A CN202310308272A CN116485810A CN 116485810 A CN116485810 A CN 116485810A CN 202310308272 A CN202310308272 A CN 202310308272A CN 116485810 A CN116485810 A CN 116485810A
- Authority
- CN
- China
- Prior art keywords
- image data
- dimensional image
- magnetic resonance
- carotid artery
- 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
- 230000011218 segmentation Effects 0.000 title claims abstract description 163
- 210000001715 carotid artery Anatomy 0.000 title claims abstract description 115
- 238000000034 method Methods 0.000 title claims abstract description 56
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 50
- 238000012549 training Methods 0.000 claims description 25
- 238000012937 correction Methods 0.000 claims description 16
- 230000000877 morphologic effect Effects 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 7
- 238000002595 magnetic resonance imaging Methods 0.000 claims description 5
- 238000005260 corrosion Methods 0.000 claims description 4
- 230000007797 corrosion Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 3
- 230000002792 vascular Effects 0.000 abstract description 6
- 230000000694 effects Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 201000001320 Atherosclerosis Diseases 0.000 description 2
- 208000007536 Thrombosis Diseases 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 2
- HVYWMOMLDIMFJA-DPAQBDIFSA-N cholesterol Chemical compound C1C=C2C[C@@H](O)CC[C@]2(C)[C@@H]2[C@@H]1[C@@H]1CC[C@H]([C@H](C)CCCC(C)C)[C@@]1(C)CC2 HVYWMOMLDIMFJA-DPAQBDIFSA-N 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 208000014882 Carotid artery disease Diseases 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 210000003484 anatomy Anatomy 0.000 description 1
- 230000003143 atherosclerotic effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000036770 blood supply Effects 0.000 description 1
- 229910052791 calcium Inorganic materials 0.000 description 1
- 239000011575 calcium Substances 0.000 description 1
- 208000037876 carotid Atherosclerosis Diseases 0.000 description 1
- 235000012000 cholesterol Nutrition 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- 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
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention provides a carotid artery segmentation method, a carotid artery segmentation device and carotid artery segmentation equipment based on a magnetic resonance image, wherein the carotid artery segmentation method comprises the following steps: acquiring magnetic resonance three-dimensional image data of carotid artery of a target patient; dividing the magnetic resonance three-dimensional image data into a first data set and a second data set along the vascular axis; inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model, and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data; respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model, and segmenting a blood vessel region from each piece of two-dimensional image data; the blood vessel region obtained in each two-dimensional image data is corrected according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data, so that the carotid artery is fully automatically segmented, and the time consumption is short, the efficiency is high and the repeatability is high. Further, through the cooperation of three-dimensional segmentation and two-dimensional segmentation, the interference of the irrelevant area on the segmentation result is removed, and the accuracy of carotid artery segmentation is improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of image segmentation, in particular to a carotid artery segmentation method, device and equipment based on a magnetic resonance image.
Background
Carotid atherosclerosis is a common clinical condition that results from the accumulation of cholesterol, fat, calcium and other substances on the wall of the carotid artery. This build-up is commonly referred to as plaque, which can occlude the carotid artery, causing local changes in vessel size, and narrow vessels resulting in reduced blood supply. If the plaque breaks suddenly, a blood clot is generated, and the blood clot can cause stroke, so that the examination of carotid artery is of great importance for the prevention and diagnosis of diseases. The magnetic resonance imaging device is a highly integrated instrument integrating physical and chemical technologies, can provide a magnetic resonance image reflecting anatomical structure and pathological information of target tissues in a noninvasive manner, and has become one of the main means for checking carotid arteries in clinic.
After obtaining a magnetic resonance image of the carotid artery, the carotid artery region of interest needs to be segmented therefrom for clinical analysis, currently relying mainly on manual segmentation. Specifically, a three-dimensional image acquired by magnetic resonance is sliced along the axial direction of a blood vessel, the tube wall and the tube cavity of a carotid artery are manually identified in the two-dimensional slice image, and sketching is performed based on the harvesting operation. Because each patient has a plurality of two-dimensional slice images to be segmented, the whole segmentation process is very time-consuming, and human factors still play a dominant role in the segmentation process, so that the repeatability of the segmentation result is low. In view of the foregoing, there is a need for an efficient carotid artery segmentation method.
Disclosure of Invention
The embodiment of the invention provides a carotid artery segmentation method, device and equipment based on a magnetic resonance image, which are used for solving the problems of long time consumption, low efficiency, low repeatability and the like in the existing manual segmentation.
In a first aspect, an embodiment of the present invention provides a carotid artery segmentation method based on a magnetic resonance image, including:
acquiring magnetic resonance three-dimensional image data of carotid artery of a target patient;
dividing the magnetic resonance three-dimensional image data into a first data set and a second data set along the axial direction of the blood vessel, wherein the first data set comprises two-dimensional image data of the left carotid artery of a plurality of target patients, and the second data set comprises two-dimensional image data of the right carotid artery of the plurality of target patients;
inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model, and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data;
respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model, and segmenting a blood vessel region from each piece of two-dimensional image data;
and correcting the blood vessel region obtained in each piece of two-dimensional image data according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data to obtain the carotid artery region of interest of the target patient.
In one embodiment, the correction of the blood vessel region obtained in each two-dimensional image data based on the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data comprises:
the blood vessel region which falls into the bifurcation region of the carotid artery in each two-dimensional image data is reserved.
In one embodiment, prior to training the three-dimensional segmentation model and the two-dimensional segmentation model, the method further comprises:
and performing linear interpolation on the partially marked training samples to generate completely marked training samples.
In one embodiment, before correcting the blood vessel region obtained in each two-dimensional image data according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data, the method further comprises:
and carrying out morphological correction on the blood vessel region segmented in each piece of two-dimensional image data, wherein the morphological correction comprises corrosion, connected domain detection, maximum connected domain reservation and expansion operation.
In one embodiment, the method further comprises:
calculating the distance between the center of the lumen and the center of the tube wall;
if the distance is greater than the preset threshold, the morphological correction operation is iteratively executed until the distance is less than the preset threshold, or the iteration times reach the preset times.
In one embodiment, acquiring magnetic resonance three-dimensional image data of a carotid artery of a subject patient, comprises:
acquiring magnetic resonance three-dimensional image data of carotid artery of a target patient in real time through magnetic resonance imaging equipment;
or,
the method comprises the steps of acquiring magnetic resonance three-dimensional image data of the carotid artery of a target patient, which is stored in advance, from a storage device.
In one embodiment, the method further comprises: and visually displaying the carotid artery region of interest obtained by segmentation based on the magnetic resonance three-dimensional image data.
In a second aspect, an embodiment of the present invention provides a carotid artery segmentation device based on magnetic resonance images, including:
the acquisition module is used for acquiring magnetic resonance three-dimensional image data of the carotid artery of the target patient;
the dividing module is used for dividing the magnetic resonance three-dimensional image data into a first data set and a second data set along the blood vessel axial direction, wherein the first data set comprises two-dimensional image data of the left carotid artery of the plurality of target patients, and the second data set comprises two-dimensional image data of the right carotid artery of the plurality of target patients;
the three-dimensional segmentation module is used for inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data;
the two-dimensional segmentation module is used for respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model and segmenting a blood vessel region from each piece of two-dimensional image data;
and the correction module is used for correcting the blood vessel region obtained in each two-dimensional image data according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data to obtain the carotid artery region of interest of the target patient.
In a third aspect, an embodiment of the present invention provides an electronic device, including:
at least one processor and memory;
the memory stores computer-executable instructions;
at least one processor executes computer-executable instructions stored in a memory, causing the at least one processor to perform the magnetic resonance image-based carotid artery segmentation method as described in any one of the first aspects.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having stored therein computer-executable instructions for implementing the magnetic resonance image-based carotid artery segmentation method according to any one of the first aspects when executed by a processor.
The carotid artery segmentation method, the carotid artery segmentation device and the carotid artery segmentation equipment based on the magnetic resonance image provided by the embodiment of the invention acquire magnetic resonance three-dimensional image data of the carotid artery of a target patient; dividing the magnetic resonance three-dimensional image data into a first data set and a second data set along the axial direction of the blood vessel, wherein the first data set comprises two-dimensional image data of the left carotid artery of a plurality of target patients, and the second data set comprises two-dimensional image data of the right carotid artery of the plurality of target patients; inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model, and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data; respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model, and segmenting a blood vessel region from each piece of two-dimensional image data; the blood vessel region obtained in each two-dimensional image data is corrected according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data, and the carotid artery region of interest of the target patient is obtained, so that the carotid artery is fully automatically segmented, and the time consumption is short, the efficiency is high and the repeatability is high. Further, through the cooperation of three-dimensional segmentation and two-dimensional segmentation, the interference of the irrelevant area on the segmentation result is removed, and the accuracy of carotid artery segmentation is improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a flowchart of a carotid artery segmentation method based on magnetic resonance image according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a visual display interface according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a carotid artery segmentation device based on magnetic resonance image according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Specific embodiments of the present invention have been shown by way of the above drawings and will be described in more detail below. The drawings and the written description are not intended to limit the scope of the inventive concepts in any way, but rather to illustrate the inventive concepts to those skilled in the art by reference to the specific embodiments.
Detailed Description
The invention will be described in further detail below with reference to the drawings by means of specific embodiments. Wherein like elements in different embodiments are numbered alike in association. In the following embodiments, numerous specific details are set forth in order to provide a better understanding of the present application. However, one skilled in the art will readily recognize that some of the features may be omitted, or replaced by other elements, materials, or methods in different situations. In some instances, some operations associated with the present application have not been shown or described in the specification to avoid obscuring the core portions of the present application, and may not be necessary for a person skilled in the art to describe in detail the relevant operations based on the description herein and the general knowledge of one skilled in the art.
Furthermore, the described features, operations, or characteristics of the description may be combined in any suitable manner in various embodiments. Also, various steps or acts in the method descriptions may be interchanged or modified in a manner apparent to those of ordinary skill in the art. Thus, the various orders in the description and drawings are for clarity of description of only certain embodiments, and are not meant to be required orders unless otherwise indicated.
The numbering of the components itself, e.g. "first", "second", etc., is used herein merely to distinguish between the described objects and does not have any sequential or technical meaning. The terms "coupled" and "connected," as used herein, are intended to encompass both direct and indirect coupling (coupling), unless otherwise indicated.
Carotid bifurcation area is a multiple site of atherosclerosis, which has the following unique challenges compared to other traditional medical segmentation tasks: first, the lumen and vessel contours are generally concentric, smooth, and closed circles in shape, i.e., all points on the contour are closely spaced from the vessel centroid. However, the atherosclerotic vessels have a variety of lumen shapes, including oval, small eccentric circles, crescent shapes, and even are not visible in the magnetic resonance image (i.e. occluded), the lumen shape being primarily dependent on the plaque distribution. Second, the lumen and wall should have a sufficiently smooth shape without sharp areas, given the smoothness and tubular nature of the carotid artery. Thus, in carotid vessel wall segmentation tasks, the lumen and wall shapes may be very different, mainly due to atherosclerosis, i.e. the presence of plaque. Thirdly, when interpretation is performed, the concerned part is concentrated at the position of 4cm above and below the carotid bifurcation, and how to accurately divide the concerned region of interest and reduce the interference of blood vessels at other parts is also important to carotid division.
Conventional segmentation algorithms such as thresholding are only applicable to images where the target and background occupy different gray level ranges, and are not applicable to high resolution carotid magnetic resonance images. When the segmentation is performed based on methods such as edge detection and wavelet transformation, the subsequent processing or other related algorithms are combined to complete the segmentation task. The existing deep learning neural network segmentation algorithm can take the characteristic information of different subareas and multiple scales into consideration, and can automatically and rapidly segment at pixel level, so that a huge result is obtained in the field of image segmentation, but the current deep learning model does not further optimize carotid artery segmentation tasks, does not take special properties of vessel walls and clinically focused parts into consideration, has poor segmentation effect, and is easy to generate wrong segmentation on tissues similar to blood vessels.
In order to solve at least one problem in the prior art, the application provides a carotid artery segmentation method based on the combination of three-dimensional segmentation and two-dimensional segmentation, which can remove interference of an irrelevant region segmentation result and accurately segment a clinically focused region of interest. The following is a detailed description of specific examples.
Fig. 1 is a flowchart of a carotid artery segmentation method based on a magnetic resonance image according to an embodiment of the present invention. As shown in fig. 1, the carotid artery segmentation method based on the magnetic resonance image provided in this embodiment may include:
s101, acquiring magnetic resonance three-dimensional image data of carotid artery of a target patient.
In this embodiment, the acquiring of magnetic resonance three-dimensional image data of the carotid artery of the target patient may be acquiring, in real time, magnetic resonance three-dimensional image data of the carotid artery of the target patient by using a magnetic resonance imaging device; the magnetic resonance three-dimensional image data of the carotid artery of the target patient stored in advance can also be acquired from the storage device.
S102, dividing magnetic resonance three-dimensional image data into a first data set and a second data set along the axial direction of a blood vessel, wherein the first data set comprises two-dimensional image data of left carotid arteries of a plurality of target patients, and the second data set comprises two-dimensional image data of right carotid arteries of the plurality of target patients.
In order to improve the accuracy of the segmentation, the left and right carotid arteries are segmented in this embodiment. Carotid artery data scanned by a magnetic resonance imaging device is usually derived from the machine in digital imaging and communications in medicine (Digital Imaging and Communications in Medicine, DICOM) format, and in this embodiment, the magnetic resonance three-dimensional image data may be divided in half into a first data set and a second data set according to the left-right direction defined in DICOM meta information. The first data set corresponds to the left carotid artery of the target patient, and two-dimensional image data of the left carotid artery of a plurality of target patients can be generated by performing two-dimensional slicing along the axial direction; the second data set corresponds to the right carotid artery of the target patient, and two-dimensional slicing along the axial direction can generate two-dimensional image data of the right carotid artery of the plurality of target patients.
S103, inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model, and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data.
The model used for three-dimensional segmentation in this embodiment may be a nnU-net model, and by performing key characterization analysis on the data set used for training, various super-parameters are adjusted without any manual intervention, and integration of the optimal model is performed through 5-time cross validation, so as to finally realize automatic training of the optimal three-dimensional segmentation model. In particular, a three-dimensional cascade of nnU-net may be employed for the segmentation of the three-dimensional vessel wall.
The important part of clinical interest is a region located about 4cm above and below the carotid bifurcation, i.e., a carotid bifurcation region. Therefore, in order to remove the interference of other irrelevant areas on carotid artery segmentation, the three-dimensional segmentation model may be trained by using a three-dimensional training sample marked with a carotid bifurcation area vessel wall mask in this embodiment.
S104, respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model, and segmenting a blood vessel region from each piece of two-dimensional image data.
The model used for two-dimensional segmentation in this embodiment may be a nnU-net model, and by performing key characterization analysis on the data set used for training, various super-parameters are adjusted without any manual intervention, and integration of the optimal model is performed through 5-time cross validation, so as to finally realize automatic training of the optimal two-dimensional segmentation model. In particular, a two-dimensional cascade of nnU-net may be employed for the segmentation of the two-dimensional vessel wall.
In this embodiment, the left carotid artery and the right carotid artery are processed separately, so that only one vessel segmentation result is generated in each two-dimensional image. In this embodiment, a two-dimensional training sample labeled with a vascular wall mask may be used to train the two-dimensional segmentation model. The two-dimensional training sample marked with the vascular wall mask can be generated by carrying out two-dimensional slicing on the three-dimensional training sample marked with the vascular wall mask of the carotid bifurcation area along the axial direction.
S105, correcting the blood vessel region obtained in each two-dimensional image data according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data to obtain the carotid artery region of interest of the target patient.
In an alternative embodiment, the correction of the blood vessel region obtained in each two-dimensional image data according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data specifically may include: the blood vessel region which falls into the bifurcation region of the carotid artery in each two-dimensional image data is reserved. The intersection of the carotid bifurcation area obtained in the magnetic resonance three-dimensional image data and the vascular area obtained in each two-dimensional image data is taken as a final segmentation result. That is, the 2D segmentation result in the 3D segmentation result range is reserved, because only the blood vessel region falling into the carotid bifurcation region is the region to be focused in clinic, and the blood vessel region obtained in each two-dimensional image data is corrected according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data, so that the segmentation accuracy can be further improved.
According to the carotid artery segmentation method based on the magnetic resonance image, magnetic resonance three-dimensional image data of a carotid artery of a target patient are obtained; dividing the magnetic resonance three-dimensional image data into a first data set and a second data set along the axial direction of the blood vessel, wherein the first data set comprises two-dimensional image data of the left carotid artery of a plurality of target patients, and the second data set comprises two-dimensional image data of the right carotid artery of the plurality of target patients; inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model, and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data; respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model, and segmenting a blood vessel region from each piece of two-dimensional image data; the blood vessel region obtained in each two-dimensional image data is corrected according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data, and the carotid artery region of interest of the target patient is obtained, so that the carotid artery is fully automatically segmented, and the time consumption is short, the efficiency is high and the repeatability is high. Further, through the cooperation of three-dimensional segmentation and two-dimensional segmentation, the interference of the irrelevant area on the segmentation result is removed, and the accuracy of carotid artery segmentation is improved.
It should be noted that, training of the three-dimensional segmentation model and the two-dimensional segmentation model requires a large number of labeled training samples, however, manual labeling is not only long, but also extremely costly, so that the actually obtained training samples usually only delineate the vessel wall contour coordinates in a limited axial 2D slice. If the model is trained by directly adopting the training samples marked by the parts, the accuracy of segmentation is reduced; if the training samples are completely marked manually, the time cost and the economic cost are both intolerable. In order to achieve both cost and segmentation accuracy, in the carotid artery segmentation method based on the magnetic resonance image provided in the embodiment, before training a three-dimensional segmentation model and a two-dimensional segmentation model, linear interpolation is performed on a partially labeled training sample to generate a completely labeled training sample. Considering that the manually drawn vessel wall contour coordinates are only delineated in a limited axial 2D slice, it is necessary to generate by linear interpolation on slices without manual labeling, supplementing a three-dimensional complete but rough vessel wall mask. Carotid artery data scanned on a magnetic resonance machine is derived from the machine in a DICOM format for 3D segmentation network training, images are segmented in half according to the left-right direction defined in DICOM meta-information, and input images are roughly divided into left and right carotid arteries for 2D segmentation network training. After the trained two-dimensional segmentation model and the trained three-dimensional segmentation model are obtained, in practical application, only DICOM-format images of the carotid artery of a patient are acquired and exported, and the trained two-dimensional segmentation model and the trained three-dimensional segmentation model are respectively input to perform 2D and 3D segmentation.
In order to further improve the accuracy of segmentation, in the carotid artery segmentation method based on the magnetic resonance image provided by the embodiment, before the blood vessel region obtained in each two-dimensional image data is corrected according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data, the blood vessel region segmented in each two-dimensional image data is also morphologically corrected to remove discontinuous blood vessels, so that adverse effects of tissues similar to blood vessels on segmentation are avoided. Specifically, morphological modification includes corrosion, connected domain detection, retention of the largest connected domain, and expansion operation, with the largest area retained. For the two-dimensional segmentation model, the left carotid blood vessel and the right carotid blood vessel are processed separately, so that each two-dimensional image only has a segmentation result of one blood vessel, and the largest connected domain component is reserved during morphological correction, thereby correcting discontinuous and wrong segmentation results of the blood vessels. This morphological optimization procedure helps correct erroneous bifurcation results at carotid bifurcation.
And centroid analysis is performed on the segmentation result, namely the distance between the center of the lumen and the center of the blood vessel wall is smaller than a certain threshold value so as to keep the lumen in the blood vessel wall. On the basis of the above embodiment, the carotid artery segmentation method based on the magnetic resonance image provided in this embodiment may further include: calculating the distance between the center of the lumen and the center of the tube wall; if the distance is greater than the preset threshold, the morphological correction operation is iteratively executed until the distance is less than the preset threshold, or the iteration times reach the preset times.
It can be understood that, because the data set for 3D segmentation network training is generated through discontinuous 2D slices, the generated data tag has a certain error with the real tag, and then the segmentation effect of the 3D network is slightly poorer than that of the 2D network, but the positioning effect of the 3D segmentation network before and after carotid bifurcation of clinical interest is better due to the space information of the 3D segmentation network; the two-dimensional segmentation results are finer, but because the network input is a slice of each layer, inaccurate segmentation is also performed at sites that are not of clinical interest, and these results should be ignored to improve the final segmentation effect. Therefore, when the segmentation results are combined, the segmentation after 2D morphology correction is taken as the final segmentation, when the sudden decrease of the blood vessel area in the 3D segmentation result is detected, namely the blood vessel area is considered to be away from the key region of interest, the segmentation results of other distal sections are removed, and the segmentation results of the clinical region of interest are reserved. And finally, outputting the 2D segmentation result corrected according to the 3D result as a final result.
On the basis of any one of the embodiments, in order to enable a user to conveniently and intuitively view the segmentation result of the carotid artery, the carotid artery segmentation method based on the magnetic resonance image provided in the embodiment may further include: and visually displaying the carotid artery region of interest obtained by segmentation based on the magnetic resonance three-dimensional image data. Referring to fig. 2, fig. 2 is a schematic diagram of a visual display interface according to an embodiment of the invention. As shown in fig. 2, the white part is the vessel wall, and the part wrapped by the white part is the vessel cavity. From the left hand side of fig. 2, it can be seen that the clinically interesting site is located about 4cm above and below the carotid bifurcation. In the concrete display, the left and right carotid arteries may be displayed as shown in the right two figures in fig. 2.
In summary, the carotid artery segmentation method based on the magnetic resonance image provided by the application can accurately locate the carotid artery bifurcation area of clinical interest by combining the three-dimensional segmentation network and the two-dimensional segmentation network, and remove the interference of the segmentation result of the irrelevant area. The two-dimensional segmentation model is beneficial to improving the accuracy of the segmentation result, the three-dimensional segmentation model is beneficial to providing the positioning of the clinically concerned region, and the two models can achieve better segmentation effect in combination. And the two-dimensional segmentation result adopts morphological correction operations such as maximum connected domain, corrosion, expansion and the like, which are helpful for removing the wrong segmentation result, so as to achieve better segmentation effect. The method can solve the problem that the existing carotid artery wall lumen segmentation marking time is long, and compared with the existing manual marking method, the automatic marking method of linear interpolation saves more manpower and time; compared with an algorithm of manual segmentation, the method has the advantages that the repeatability is higher, and the segmentation result is more stable; compared with other segmentation methods based on a single neural network, the segmentation method based on the two-dimensional and three-dimensional collaborative segmentation neural network respectively carries out two-dimensional segmentation on the left carotid artery and the right carotid artery, fully considers the carotid artery characteristics focused clinically, improves the segmentation accuracy, and has faster processing speed and more stable performance compared with the traditional algorithm.
Fig. 3 is a schematic structural diagram of a carotid artery segmentation device based on magnetic resonance image according to an embodiment of the present invention. As shown in fig. 3, the carotid artery segmentation device 30 based on the magnetic resonance image provided in the present embodiment may include: an acquisition module 301, a division module 302, a three-dimensional segmentation module 303, a two-dimensional segmentation module 304, and a correction module 305.
An acquisition module 301, configured to acquire magnetic resonance three-dimensional image data of a carotid artery of a target patient;
a dividing module 302, configured to divide the magnetic resonance three-dimensional image data into a first data set and a second data set along the vascular axis, where the first data set includes two-dimensional image data of left carotid arteries of a plurality of target patients, and the second data set includes two-dimensional image data of right carotid arteries of the plurality of target patients;
the three-dimensional segmentation module 303 is configured to input magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model, and segment a carotid bifurcation area from the magnetic resonance three-dimensional image data;
the two-dimensional segmentation module 304 is configured to input the first dataset and the second dataset into a pre-trained two-dimensional segmentation model, and segment a blood vessel region from each piece of two-dimensional image data;
the correction module 305 is configured to correct a blood vessel region obtained in each two-dimensional image data according to a carotid bifurcation region obtained in the magnetic resonance three-dimensional image data, so as to obtain a carotid artery region of interest of the target patient.
The device of this embodiment may be used to implement the technical solution of the method embodiment shown in fig. 1, and its implementation principle and technical effects are similar, and are not described here again.
An embodiment of the present invention further provides an electronic device, and referring to fig. 4, the embodiment of the present invention is illustrated by taking fig. 4 as an example only, and the present invention is not limited thereto. Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. As shown in fig. 4, the electronic device 40 provided in this embodiment may include: memory 401, processor 402, and bus 403. Wherein the bus 403 is used to implement the connections between the elements.
The memory 401 stores a computer program, which when executed by the processor 402 can implement the technical solution of any of the above-mentioned method embodiments.
Wherein the memory 401 and the processor 402 are electrically connected directly or indirectly to enable transmission or interaction of data. For example, the elements may be electrically coupled to each other via one or more communication buses or signal lines, such as via bus 403. The memory 401 stores therein a computer program for implementing a carotid artery segmentation method based on magnetic resonance images, including at least one software functional module that can be stored in the memory 401 in the form of software or firmware, and the processor 402 executes various functional applications and data processing by running the software program and the module stored in the memory 401.
The Memory 401 may be, but is not limited to, random access Memory (Random Access Memory, RAM), read Only Memory (ROM), programmable Read Only Memory (Programmable Read-Only Memory, PROM), erasable Read Only Memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable Read Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), etc. The memory 401 is used for storing a program, and the processor 402 executes the program after receiving an execution instruction. Further, the software programs and modules within the memory 401 described above may also include an operating system, which may include various software components and/or drivers for managing system tasks (e.g., memory management, storage device control, power management, etc.), and may communicate with various hardware or software components to provide an operating environment for other software components.
The processor 402 may be an integrated circuit chip with signal processing capabilities. The processor 402 may be a general-purpose processor, including a central processing unit (Central Processing Unit, abbreviated as CPU), a network processor (Network Processor, abbreviated as NP), and the like. The disclosed methods, steps, and logic blocks in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. It will be appreciated that the configuration of fig. 4 is merely illustrative and may include more or fewer components than shown in fig. 4 or have a different configuration than shown in fig. 4. The components shown in fig. 4 may be implemented in hardware and/or software.
The embodiment of the invention also provides a computer readable storage medium, on which a computer program is stored, the computer program being executed by a processor to implement the technical solution of any of the method embodiments described above.
The various embodiments in this disclosure are described in a progressive manner, and identical and similar parts of the various embodiments are all referred to each other, and each embodiment is mainly described as different from other embodiments.
The scope of the present disclosure is not limited to the above-described embodiments, and it is apparent that various modifications and variations can be made to the present disclosure by those skilled in the art without departing from the scope and spirit of the disclosure. Such modifications and variations are intended to be included herein within the scope of the following claims and their equivalents.
Claims (10)
1. A carotid artery segmentation method based on magnetic resonance images, comprising:
acquiring magnetic resonance three-dimensional image data of carotid artery of a target patient;
dividing the magnetic resonance three-dimensional image data into a first data set and a second data set along the axial direction of a blood vessel, wherein the first data set comprises a plurality of two-dimensional image data of the left carotid artery of the target patient, and the second data set comprises a plurality of two-dimensional image data of the right carotid artery of the target patient;
inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model, and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data;
respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model, and segmenting a blood vessel region from each piece of two-dimensional image data;
and correcting the blood vessel region obtained in each two-dimensional image data according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data to obtain the carotid artery region of interest of the target patient.
2. The method of claim 1, wherein the modifying the vessel region from each two-dimensional image data from the carotid bifurcation region from the magnetic resonance three-dimensional image data comprises:
and reserving a blood vessel region falling into the carotid bifurcation region in each piece of two-dimensional image data.
3. The method of claim 1, wherein prior to training the three-dimensional segmentation model and the two-dimensional segmentation model, the method further comprises:
and performing linear interpolation on the partially marked training samples to generate completely marked training samples.
4. The method of claim 1, wherein before the correcting the vessel region from each two-dimensional image data from the carotid bifurcation region from the magnetic resonance three-dimensional image data, the method further comprises:
and carrying out morphological correction on the blood vessel region segmented in each piece of two-dimensional image data, wherein the morphological correction comprises corrosion, connected domain detection, maximum connected domain reservation and expansion operation.
5. The method according to claim 4, wherein the method further comprises:
calculating the distance between the center of the lumen and the center of the tube wall;
and if the distance is greater than a preset threshold, iteratively executing the morphological correction operation until the distance is less than the preset threshold, or the iteration times reach the preset times.
6. The method of claim 1, wherein the acquiring magnetic resonance three-dimensional image data of the carotid artery of the subject patient comprises:
acquiring magnetic resonance three-dimensional image data of carotid artery of a target patient in real time through magnetic resonance imaging equipment;
or,
the method comprises the steps of acquiring magnetic resonance three-dimensional image data of the carotid artery of a target patient, which is stored in advance, from a storage device.
7. The method according to any one of claims 1-6, further comprising:
and visually displaying the carotid artery region of interest obtained by segmentation based on the magnetic resonance three-dimensional image data.
8. A carotid artery segmentation device based on magnetic resonance images, comprising:
the acquisition module is used for acquiring magnetic resonance three-dimensional image data of the carotid artery of the target patient;
the dividing module is used for dividing the magnetic resonance three-dimensional image data into a first data set and a second data set along the blood vessel axial direction, wherein the first data set comprises a plurality of two-dimensional image data of the left carotid artery of the target patient, and the second data set comprises a plurality of two-dimensional image data of the right carotid artery of the target patient;
the three-dimensional segmentation module is used for inputting the magnetic resonance three-dimensional image data into a pre-trained three-dimensional segmentation model and segmenting a carotid bifurcation area from the magnetic resonance three-dimensional image data;
the two-dimensional segmentation module is used for respectively inputting the first data set and the second data set into a pre-trained two-dimensional segmentation model and segmenting a blood vessel region from each piece of two-dimensional image data;
and the correction module is used for correcting the blood vessel region obtained in each piece of two-dimensional image data according to the carotid bifurcation region obtained in the magnetic resonance three-dimensional image data to obtain the carotid artery region of interest of the target patient.
9. An electronic device, comprising: at least one processor and memory;
the memory stores computer-executable instructions;
the at least one processor executing computer-executable instructions stored in the memory causes the at least one processor to perform the magnetic resonance image-based carotid artery segmentation method as defined in any one of claims 1-7.
10. A computer readable storage medium having stored therein computer executable instructions for implementing the magnetic resonance image based carotid artery segmentation method as claimed in any one of claims 1-7 when executed by a processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310308272.7A CN116485810A (en) | 2023-03-27 | 2023-03-27 | Carotid artery segmentation method, device and equipment based on magnetic resonance image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310308272.7A CN116485810A (en) | 2023-03-27 | 2023-03-27 | Carotid artery segmentation method, device and equipment based on magnetic resonance image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116485810A true CN116485810A (en) | 2023-07-25 |
Family
ID=87225990
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310308272.7A Pending CN116485810A (en) | 2023-03-27 | 2023-03-27 | Carotid artery segmentation method, device and equipment based on magnetic resonance image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116485810A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116934741A (en) * | 2023-09-11 | 2023-10-24 | 首都医科大学附属北京天坛医院 | Method and device for acquiring composition and quantitative parameters of one-stop type blood vessel wall |
| CN117084654A (en) * | 2023-10-19 | 2023-11-21 | 深圳市鹰瞳智能技术有限公司 | Method and system for evaluating ocular artery collateral blood flow dynamics based on facial thermal image |
| CN117115187A (en) * | 2023-10-24 | 2023-11-24 | 北京联影智能影像技术研究院 | Carotid artery wall segmentation method, carotid artery wall segmentation device, carotid artery wall segmentation computer device, and carotid artery wall segmentation storage medium |
-
2023
- 2023-03-27 CN CN202310308272.7A patent/CN116485810A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116934741A (en) * | 2023-09-11 | 2023-10-24 | 首都医科大学附属北京天坛医院 | Method and device for acquiring composition and quantitative parameters of one-stop type blood vessel wall |
| CN116934741B (en) * | 2023-09-11 | 2023-12-26 | 首都医科大学附属北京天坛医院 | Method and device for acquiring composition and quantitative parameters of one-stop type blood vessel wall |
| CN117084654A (en) * | 2023-10-19 | 2023-11-21 | 深圳市鹰瞳智能技术有限公司 | Method and system for evaluating ocular artery collateral blood flow dynamics based on facial thermal image |
| CN117084654B (en) * | 2023-10-19 | 2024-01-12 | 深圳市鹰瞳智能技术有限公司 | Method and system for evaluating ocular artery collateral blood flow dynamics based on facial thermal image |
| CN117115187A (en) * | 2023-10-24 | 2023-11-24 | 北京联影智能影像技术研究院 | Carotid artery wall segmentation method, carotid artery wall segmentation device, carotid artery wall segmentation computer device, and carotid artery wall segmentation storage medium |
| CN117115187B (en) * | 2023-10-24 | 2024-02-09 | 北京联影智能影像技术研究院 | Carotid artery wall segmentation method, carotid artery wall segmentation device, carotid artery wall segmentation computer device, and carotid artery wall segmentation storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116485810A (en) | Carotid artery segmentation method, device and equipment based on magnetic resonance image | |
| CN108520519B (en) | Image processing method and device and computer readable storage medium | |
| US7773791B2 (en) | Analyzing lesions in a medical digital image | |
| US10497123B1 (en) | Isolation of aneurysm and parent vessel in volumetric image data | |
| CN101558321B (en) | Fibre tracking on the basis of macroscopic information | |
| CN107895364B (en) | A kind of three-dimensional reconstruction system for the preoperative planning of virtual operation | |
| US20120323547A1 (en) | Method for intracranial aneurysm analysis and endovascular intervention planning | |
| US20050110791A1 (en) | Systems and methods for segmenting and displaying tubular vessels in volumetric imaging data | |
| CN106709920B (en) | Blood vessel extraction method and device | |
| EP2901934B1 (en) | Method and device for generating virtual endoscope image, and program | |
| US9691157B2 (en) | Visualization of anatomical labels | |
| Debarba et al. | Efficient liver surgery planning in 3D based on functional segment classification and volumetric information | |
| CN107481254A (en) | Processing method, device, medium and the electronic equipment of medical image | |
| EP2901935B1 (en) | Method and device for generating virtual endoscope image, and program | |
| US8934690B2 (en) | Method for processing vascular structure images | |
| CN117115150B (en) | Method, computing device and medium for determining branch vessels | |
| CN110706236A (en) | Three-dimensional reconstruction method and device of blood vessel image | |
| CN111105873A (en) | Auxiliary diagnosis and measurement method and system in endoscopic retrograde cholangiopancreatography | |
| CN104146766A (en) | Scanning device, scanning method and medical image device | |
| CN111627023B (en) | Method and device for generating coronary artery projection image and computer readable medium | |
| CN112489051A (en) | Liver cutting method and system based on blood vessels and lesion areas | |
| Zahel et al. | Rapid assessment of liver volumetry by a novel automated segmentation algorithm | |
| CN115861347A (en) | Method and device for extracting focus region, electronic equipment and readable storage medium | |
| CN114931435A (en) | Three-dimensional model processing method and device and electronic equipment | |
| Boneš et al. | Automatic segmentation and alignment of uterine shapes from 3D ultrasound data |
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 |