CN112070731B - Method for guiding registration of human body model atlas and case CT image by artificial intelligence - Google Patents
Method for guiding registration of human body model atlas and case CT image by artificial intelligence Download PDFInfo
- Publication number
- CN112070731B CN112070731B CN202010879567.6A CN202010879567A CN112070731B CN 112070731 B CN112070731 B CN 112070731B CN 202010879567 A CN202010879567 A CN 202010879567A CN 112070731 B CN112070731 B CN 112070731B
- Authority
- CN
- China
- Prior art keywords
- data
- case
- registration
- model
- organ
- 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
Images
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/04—Indexing scheme for image data processing or generation, in general involving 3D image data
-
- 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
- G06T2207/30204—Marker
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (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)
Abstract
A method for guiding registration of a human body model atlas and a case CT image by applying artificial intelligence relates to the field of medical image registration and segmentation, and comprises the following steps: s1: establishing a parameter system capable of expressing the diversity of CT data in a parameterization manner; s2: making representative data sets M { M0 and M1 … Mn } of the human body whole body model atlas; s3: adaptively finding model data which is most matched with the case CT in the data set M { M0, M1 … Mn } of the human whole body model atlas as candidate registration model data Mj; s4: guiding the position information of the candidate registration model data Mj to be updated to obtain initialized registration model data Mj 2; s5: the initialized registration model data Mj2 is registered to the case CT. The registration is carried out by establishing a parameter system and guiding the mode of selecting the model and initializing the registration by artificial intelligence, so that the problem that the model form cannot meet the diversity of real data is avoided, a model closer to the real case data is provided, and a better registration initialization state is obtained to improve the precision of the registration and the segmentation.
Description
Technical Field
The invention relates to the field of medical image registration and segmentation, in particular to a method for guiding registration of a human body model atlas and a case CT image by applying artificial intelligence.
Background
Medical image registration refers to a process of seeking one or a series of spatial transformation for one piece of medical image data to make the medical image data spatially consistent with a region of interest of another piece of target data, and with the continuous development of medical imaging equipment, the registration and segmentation of organs or regions of interest of the data are required to perform comprehensive analysis by utilizing more and more abundant medical data. For human medical data, on one hand, a large number of tissues or organs have fuzzy boundaries, complex gradients and difficulty in direct segmentation, and on the other hand, the sizes, shapes and corresponding structures of organs of the organs or the tissues in the human body are relatively stable. The prior knowledge of the relative relation between the human body structure and the tissue is utilized for segmentation, and the segmentation can be used as an effective method for improving the conditions of unclear organ boundaries and complex gradient.
For human tissues and organs, an Atlas model set based segmentation method is often used to perform the organ segmentation task using a priori knowledge of the human structure. The Atlas model set is created by manually labeling the designated organs or tissues in the data, and the selection of the model in the segmentation process usually has two types: and selecting the model closest to the target by sequentially registering the plurality of models to the result of the case CT or obtaining a registration model by using a model set to perform cluster analysis, and finally segmenting the case CT through registration and model marking information. In practical application, the Atlas model based segmentation method has some defects: firstly, when the registration model is selected, the optimal mode selected by multiple times of registration is used for calculating, so that the segmentation duration is greatly increased, and the mode of selecting the registration model by using the clustering method influences the registration precision; secondly, because the morphology of human organs is relatively stable and the interrelationship between organs and tissues may generate obvious differences with different patients, the Atlas segmentation method usually segments a specific organ or tissue with relatively stable morphology, such as tissues of cervix, spleen, brain, etc., and for a larger range of objects, such as abdomen, thorax, and half-body data, it cannot process effectively due to lack of integral prior knowledge judgment on the anatomical structure and imaging posture of human body. 7-10, for the larger range of case data segmentation requirements in the clinic, Atlas segmentation based methods will face the mismatch problem caused by the large difference between the selected registration model and the target CT anatomy or the poor registration initialization state.
Disclosure of Invention
Aiming at the defects, the invention aims to provide a method for guiding registration of a human body model atlas and a case CT image by applying artificial intelligence, solves the problem that the model form cannot meet the diversity of real data, provides a model closer to the real case data, and obtains a better registration initialization state so as to improve the registration and segmentation precision.
In order to achieve the purpose, the invention adopts the following technical scheme:
the method for guiding registration of the human body model atlas and the case CT image by applying artificial intelligence comprises the following steps:
s1: collecting medical data statistical information of crowds in different age groups and different regions, and establishing a parameter system capable of expressing CT data diversity in a parameterization mode;
s2: according to the relevance, mutual limitation and crowd representativeness of each parameter in the parameter system on actual data, a group of representative human body whole body model atlas data sets M { M0 and M1 … Mn };
s3: analyzing the case CT through artificial intelligence, respectively calculating corresponding parameter values of the case CT according to the parameter types in the parameter system, and adaptively finding model data which is most matched with the case CT from the data set M { M0 and M1 … Mn } of the human body whole body model atlas as candidate registration model data Mj according to the calculated parameter values;
s4: analyzing the case CT through artificial intelligence, calculating the size of an organ, a CT imaging range and position information corresponding to the organ contained in the case CT, and guiding the position information of each corresponding organ in the candidate registration model data Mj to update by combining the position information of the corresponding organ in the candidate registration model data Mj to obtain initialized registration model data Mj 2;
s5: and registering the initialized registered model data Mj2 to the case CT by applying non-rigid registration to complete image registration and organ segmentation between the human body model and the case CT.
Further, in step S1, the parameter system includes an anatomical structure parameter, a data imaging posture parameter, and a data specific information parameter.
Further, in the step S2, each item Mk (k 0,1, 2, …, n) in the human whole-body model atlas data set M { M0, M1 … Mn } represents model data created using the same anatomical structure parameter, the data imaging posture parameter, and the data-specific parameter as respective CT data in the demographic information, respectively, wherein each item Mk (k 0,1, 2, …, n) includes two three-dimensional model images corresponding to each other and having respective pixel values of a CT value and an organ label index.
Further, the step S3 includes the following steps:
analyzing the case CT through artificial intelligence, and calculating the anatomical structure parameters, the data imaging posture parameters and the data specific information parameters in the case CT;
and adaptively selecting model data which is most matched with the case CT from the data set M { M0 and M1 … Mn } of the whole-body model atlas of the human body as candidate registration model data Mj according to the anatomical structure parameters, the data imaging posture parameters and the data specific information parameters which are calculated in the case CT.
Further, the step S4 includes the following steps:
analyzing the case CT through artificial intelligence, judging whether the case CT is whole body data or partial data, and calculating organs contained in the case CT, the size of a CT imaging range and position information corresponding to the organs;
calculating the relative displacement relation between the candidate registration model data Mj and each organ in the case CT according to the position information of the corresponding organ in the candidate registration model data Mj;
and updating the position of the corresponding organ in the candidate registration model data Mj to obtain initialized registration model data Mj2, so that each organ in the case CT is respectively and preliminarily aligned with the corresponding organ in the initialized registration model data Mj 2.
Further, the step S5 includes the following steps:
extracting a three-dimensional model image with a pixel value of a CT value in the initialized registration model Mj2, and registering the three-dimensional model image with the pixel value of the CT value in case CT by adopting a non-rigid registration method, so that the three-dimensional model image with the pixel value of the CT value is consistent with organs in case CT in space, and a registered conversion parameter T is obtained;
and extracting a three-dimensional model image with the pixel value in the initialized registration model Mj2 as an organ label index, and performing spatial transformation according to the transformation parameter T to obtain label data consistent with the spatial position of the case CT.
Further, the method also comprises the step of optimizing the result of the organ segmentation by applying a dynamic contour calculation method.
Has the advantages that:
in the registration method of the human body model and case CT images guided by the artificial intelligence algorithm, a parameter system capable of expressing the diversity of CT data in a parameterization mode is established according to the statistical information of the medical data of people, a representative data set M of the whole body model Atlas of the human body is established by using the priori knowledge of the relative relation of the human body structure, and a model for marking real data is not needed, so that the problem that the model form cannot meet the diversity of the real data is solved.
According to the method, candidate registration model data Mj which is most matched with the case CT is selected from a human body whole body model atlas data set M through artificial intelligence, and the position information of the candidate registration model data Mj is guided to be updated to obtain initialized registration model data Mj2, so that each organ in the case CT is respectively and initially aligned with the corresponding organ in the initialized registration model data Mj 2. Compared with the mode of selecting a fixed model or selecting a clustering model by other methods, the method can provide a model which is closer to real case data, and obtain a better registration initialization state so as to improve the precision of registration and segmentation.
The method applies non-rigid registration to register the initial shape registration model data Mj2 to the case CT, completes image registration and organ segmentation between the human model data and the case CT, realizes the task of performing full data direct registration and segmentation on the local or integral case CT, and obtains the integral segmentation result of the case CT without respectively segmenting each organ and then combining, thereby improving the efficiency of image registration and organ segmentation.
Drawings
FIG. 1 is a schematic flow diagram of one embodiment of the present invention;
FIG. 2 is a three-dimensional model image with CT values as pixel values in one item of model data according to an embodiment of the present invention;
FIG. 3 is a three-dimensional model image with pixel values indexed by organ labels in one item of model data according to an embodiment of the present invention;
FIG. 4 is a schematic illustration of liver orthostatic for case CT in one embodiment of the present invention;
fig. 5 is an orthographic view of the liver of the case CT of fig. 4 after registration to the initialized registration model data Mj 2;
fig. 6 is a pictorial representation of the liver of the case CT of fig. 4 in true position after segmentation from liver label index data, wherein the gray regions are shown in actual operation as red;
FIG. 7 is a schematic illustration of a prior art orthostatic CT of a female local case;
fig. 8 is an orthographic view of the case CT registration of fig. 7 to a female model;
fig. 9 is an orthotopic schematic of the case CT registration of fig. 7 to a male model;
fig. 10 is an orthostatic schematic of the case CT registration of fig. 7 to a female model with wrong arm orientation.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention.
Referring to fig. 1-10, a method for guiding registration of a manikin atlas and a case CT image by applying artificial intelligence comprises the following steps:
s1: collecting medical data statistical information of crowds in different age groups and different regions, and establishing a parameter system capable of expressing CT data diversity in a parameterization mode;
s2: according to the relevance, mutual limitation and crowd representativeness of each parameter in the parameter system on actual data, a group of representative human body whole body model atlas data sets M { M0 and M1 … Mn };
s3: analyzing the case CT through artificial intelligence, respectively calculating corresponding parameter values of the case CT according to the parameter types in the parameter system, and adaptively finding model data which is most matched with the case CT from the data set M { M0 and M1 … Mn } of the human body whole body model atlas as candidate registration model data Mj according to the calculated parameter values;
s4: analyzing the case CT through artificial intelligence, calculating the size of an organ, a CT imaging range and position information corresponding to the organ contained in the case CT, and guiding the position information of each corresponding organ in the candidate registration model data Mj to update by combining the position information of the corresponding organ in the candidate registration model data Mj to obtain initialized registration model data Mj 2;
s5: and registering the initialized registered model data Mj2 to the case CT by applying non-rigid registration to complete image registration and organ segmentation between the human body model and the case CT.
In the registration method of the human body model and case CT images guided by the artificial intelligence algorithm, a parameter system capable of expressing the diversity of CT data in a parameterization mode is established according to the statistical information of the medical data of people, a representative data set M of the whole body model Atlas of the human body is established by using the priori knowledge of the relative relation of the human body structure, and a model for marking real data is not needed, so that the problem that the model form cannot meet the diversity of the real data is solved.
According to the method, candidate registration model data Mj which is most matched with the case CT is selected from a human body whole body model atlas data set M through artificial intelligence, and the position information of the candidate registration model data Mj is guided to be updated to obtain initialized registration model data Mj2, so that each organ in the case CT is respectively and initially aligned with the corresponding organ in the initialized registration model data Mj 2. Compared with the mode of selecting a fixed model or selecting a clustering model by other methods, the method can provide a model which is closer to real case data, and obtain a better registration initialization state so as to improve the precision of registration and segmentation.
The method applies non-rigid registration to register the initial shape registration model data Mj2 to the case CT, completes image registration and organ segmentation between the human model data and the case CT, realizes the task of performing full data direct registration and segmentation on the local or integral case CT, and obtains the integral segmentation result of the case CT without respectively segmenting each organ and then combining, thereby improving the efficiency of image registration and organ segmentation.
Specifically, in step S1, the parameter system includes an anatomical structure parameter, a data imaging posture parameter, and a data specific information parameter.
Wherein the anatomical structure parameters are used for expressing information which can cause the difference of the human anatomical structure, such as the sex, the age group, the height, the weight, the body type (olive type, lean type, fat type, strong type, etc.), and the like, and the absolute position of organs, the relative position between organs, the volume and the shape of organs, and the like;
the data imaging posture parameter is used for expressing the difference information of the patient posture in the CT imaging process, such as the respiratory state of a human body, the positioning and orientation of the human body during imaging, the posture of the human body during imaging, the size of the imaging range and the like;
the data specific information parameter is used for expressing the information of special conditions existing in case CT, such as the absence of organs and the position and range of corresponding organs, whether large-range or obvious lesion and the position and range thereof appear in the data and the like.
The information is used for establishing a set of parameter system based on human anatomy structure parameters, imaging attitude parameters and specific parameters, so that the parameter system can carry out parametric representation on the diversity of case CT data, and for any case CT, a series of parameters can represent the characteristics of the basic anatomy structure, the imaging attitude, the specific information and the like.
Specifically, in step S2, each item Mk (k 0,1, 2, …, n) in the human body whole-body model atlas data set M { M0, M1 … Mn } represents a set of model data, wherein each item Mk (k 0,1, 2, …, n) includes two three-dimensional model images, and the anatomical structure parameter, the data imaging posture parameter and the data specific information parameter in the two three-dimensional model images correspond to the same parameter, respectively, and a pixel value of one three-dimensional model image is a CT value and a pixel value of the other three-dimensional model image is an organ label index. In some embodiments of the present invention, each item Mk (k is 0,1, 2, …, n) in the human whole-body model atlas data set M { M0, M1 … Mn } includes mk.ct, mk.tag, and mk.par, respectively, where mk.ct refers to CT data, and is a three-dimensional model image with CT values as shown in fig. 2. Tag is organ label index data, and is a three-dimensional model image having a pixel value as an organ label index as shown in fig. 3, and the pixel values of the organ in the three-dimensional model image correspond to the organ label indexes. The anatomical parameters, data imaging pose parameters and data specific parameters in the model data are stored at mk. The three-dimensional model image with the pixel value being the CT value is used for realizing image registration with case CT, and the three-dimensional model image with the pixel value being the organ label index is used for marking each pixel in the case CT so as to complete organ segmentation.
Specifically, the step S3 includes the following steps:
analyzing the case CT through artificial intelligence, and calculating the anatomical structure parameters, the data imaging posture parameters and the data specific information parameters in the case CT;
and adaptively selecting model data which is most matched with the case CT from the data set M { M0 and M1 … Mn } of the whole-body model atlas of the human body as candidate registration model data Mj according to the anatomical structure parameters, the data imaging posture parameters and the data specific information parameters which are calculated in the case CT.
In some embodiments of the invention, a case CT is input into an artificial intelligence network, an anatomical structure parameter, a data imaging posture parameter and a data specific information parameter in the case CT are calculated by the artificial intelligence network, so that parameter values corresponding to the case CT and a parameter system are calculated by the artificial intelligence network, then, the human body whole body model data which is most matched with the case CT is selected from a human body whole body model atlas data set M in a self-adaptive manner as candidate registration model data Mj according to the anatomical structure parameter, the data imaging posture parameter and the data specific information parameter in the calculated case CT, so that model data selected by the method can meet real data diversity, and the matching degree between the model data and the case CT is improved.
Specifically, the step S4 includes the following steps:
analyzing the case CT through artificial intelligence, judging whether the case CT is whole body data or partial data, and calculating organs contained in the case CT, the size of a CT imaging range and position information corresponding to the organs;
calculating the relative displacement relation between the candidate registration model data Mj and each organ in the case CT according to the position information of the corresponding organ in the candidate registration model data Mj;
and guiding the position information of the candidate registration model data Mj to be updated based on the alignment of the candidate registration model data Mj and the organs in the case CT to obtain initialized registration model data Mj2, so that each organ in the case CT is respectively and preliminarily aligned with the corresponding organ in the initialized registration model data Mj 2. The method realizes the registration initialization of the model data by artificial intelligence, obtains a better registration initialization state to improve the precision of registration and segmentation, is favorable for initializing the registration model data Mj2 to be more suitable for the whole data and the large-range data, and is not limited by the range of target data and the number of organs. And guiding the position information update of the candidate registration model data Mj, wherein the position information update of the candidate registration model data Mj is to update the position information of the candidate registration model data Mj according to the relative displacement relation between the candidate registration model data Mj and each organ in the case CT, so that each organ in the candidate registration model data Mj is basically aligned with the corresponding organ in the case CT in space.
Specifically, the step S5 includes the following steps:
using the three-dimensional model image with the pixel value of the CT value in the initialized registration model Mj2 to register the three-dimensional model image with the pixel value of the CT value in case CT by using a non-rigid registration method, so that the three-dimensional model image with the pixel value of the CT value is consistent with the organ in case CT in space, and obtaining a registered conversion parameter T;
and performing spatial transformation according to the transformation parameter T by using the three-dimensional model image with the pixel value in the initialized registration model Mj2 as the index of the organ label to obtain label data consistent with the spatial position of the case CT.
In some embodiments of the present invention, as shown in fig. 4 and 5, the non-rigid registration method is adopted to register the three-dimensional model image with the pixel value being the CT value to the case CT, so that the three-dimensional model image with the pixel value being the CT value is spatially consistent with the organ in the case CT, thereby completing the image registration between the human model data and the case CT, and obtaining the conversion parameter T of the registration. Then, as shown in fig. 4 and 6, the three-dimensional model image with the pixel value as the organ label index is spatially converted according to the conversion parameter T to obtain label data that is consistent with the spatial position of the case CT, so as to clarify the organ to which each pixel in the case CT belongs, and realize organ segmentation. Non-rigid registration applied includes, but is not limited to, affine registration and bspline registration, among others.
Further, in some embodiments of the present invention, the result of organ segmentation is optimized by applying a dynamic contour calculation method. Specifically, after organ segmentation is completed, the organ segmentation result is optimized by using a dynamic contour calculation method, so that the accuracy of the organ segmentation result is improved, and a more accurate organ segmentation result is obtained. Wherein the dynamic contour calculation method may be a level set method.
The technical principle of the present invention is described above in connection with specific embodiments. The description is made for the purpose of illustrating the principles of the invention and should not be construed in any way as limiting the scope of the invention. Based on the explanations herein, those skilled in the art will be able to conceive of other embodiments of the present invention without inventive effort, which would fall within the scope of the present invention.
Claims (7)
1. The method for guiding registration of the human body model atlas and the case CT image by applying artificial intelligence is characterized by comprising the following steps of:
s1: collecting medical data statistical information of crowds in different age groups and different regions, and establishing a parameter system capable of expressing CT data diversity in a parameterization mode;
s2: according to the relevance, mutual limitation and crowd representativeness of each parameter in the parameter system on actual data, a group of representative human body whole body model atlas data sets M { M0 and M1 … Mn };
s3: analyzing the case CT through artificial intelligence, respectively calculating corresponding parameter values of the case CT according to the parameter types in the parameter system, and adaptively finding model data which is most matched with the case CT from the data set M { M0 and M1 … Mn } of the human body whole body model atlas as candidate registration model data Mj according to the calculated parameter values;
s4: analyzing the case CT through artificial intelligence, calculating the size of an organ, a CT imaging range and position information corresponding to the organ contained in the case CT, guiding the position information of the candidate registration model data Mj to be updated based on the alignment of the candidate registration model data Mj and the organ in the case CT, and obtaining initialized registration model data Mj 2;
s5: and registering the initialized registered model data Mj2 to the case CT by applying non-rigid registration to complete image registration and organ segmentation between the human body model and the case CT.
2. The method for guiding registration of an atlas of mannequins with case CT images using artificial intelligence as claimed in claim 1, wherein in step S1 the parameter system includes anatomical parameters, data imaging pose parameters and data specific information parameters.
3. The method for guiding registration of a atlas of human body model with CT image of case as claimed in claim 2, wherein in step S2, each Mk (k 0,1, 2, …, n) in the data set M { M0, M1 … Mn } of the atlas of human body model respectively represents a set of model data, wherein each Mk (k 0,1, 2, …, n) contains two three-dimensional model images, and the anatomical structure parameter, the data imaging pose parameter and the data specific information parameter in the two three-dimensional model images respectively correspond to the same, wherein the pixel value of one three-dimensional model image is CT value and the pixel value of the other three-dimensional model image is organ label index.
4. The method for guiding registration of a manikin atlas and case CT images using artificial intelligence as claimed in claim 3, wherein said step S3 includes the following procedures:
analyzing the case CT through artificial intelligence, and calculating the anatomical structure parameters, the data imaging posture parameters and the data specific information parameters in the case CT;
and adaptively selecting model data which is most matched with the case CT from the data set M { M0 and M1 … Mn } of the whole-body model atlas of the human body as candidate registration model data Mj according to the anatomical structure parameters, the data imaging posture parameters and the data specific information parameters which are calculated in the case CT.
5. The method for guiding registration of a manikin atlas and case CT images using artificial intelligence as claimed in claim 4, wherein said step S4 includes the following procedures:
analyzing the case CT through artificial intelligence, judging whether the case CT is whole body data or partial data, and calculating organs contained in the case CT, the size of a CT imaging range and position information corresponding to the organs;
calculating the relative displacement relation between the candidate registration model data Mj and each organ in the case CT according to the position information of the corresponding organ in the candidate registration model data Mj;
and guiding the position information of the candidate registration model data Mj to be updated based on the alignment of the candidate registration model data Mj and the organs in the case CT to obtain initialized registration model data Mj2, so that each organ in the case CT is respectively and preliminarily aligned with the corresponding organ in the initialized registration model data Mj 2.
6. The method for guiding registration of an atlas of mannequins with case CT images using artificial intelligence as claimed in claim 5, wherein the step S5 includes the following procedures:
using the three-dimensional model image with the pixel value of the CT value in the initialized registration model Mj2, and adopting a non-rigid registration method to register the three-dimensional model image with the pixel value of the CT value in case CT, so that the three-dimensional model image with the pixel value of the CT value is consistent with the organs in case CT in space, and obtaining a registered conversion parameter T;
and performing spatial transformation according to the transformation parameter T by using the three-dimensional model image with the pixel value in the initialized registration model Mj2 as the index of the organ label to obtain label data consistent with the spatial position of the case CT.
7. The method for guiding registration of an atlas of mannequin images with case CT images using artificial intelligence as recited in claim 1, further comprising optimizing the results of organ segmentation using dynamic contour computation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010879567.6A CN112070731B (en) | 2020-08-27 | 2020-08-27 | Method for guiding registration of human body model atlas and case CT image by artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010879567.6A CN112070731B (en) | 2020-08-27 | 2020-08-27 | Method for guiding registration of human body model atlas and case CT image by artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112070731A CN112070731A (en) | 2020-12-11 |
| CN112070731B true CN112070731B (en) | 2021-05-11 |
Family
ID=73659990
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010879567.6A Active CN112070731B (en) | 2020-08-27 | 2020-08-27 | Method for guiding registration of human body model atlas and case CT image by artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112070731B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104603836A (en) * | 2012-08-06 | 2015-05-06 | 范德比尔特大学 | Enhanced method for correcting data for deformations during image guided procedures |
| CN105118068A (en) * | 2015-09-29 | 2015-12-02 | 常熟理工学院 | Medical image automatic annotation method under small sample condition |
| CN107103207A (en) * | 2017-04-05 | 2017-08-29 | 浙江大学 | Based on the multigroup accurate medical knowledge search system and implementation method for learning variation features of case |
| EP3441939A1 (en) * | 2007-01-17 | 2019-02-13 | Mediguide Ltd. | A method and system for registering a 3d pre-acquired image coordinates system with a medical positioning system coordinate system and with a 2d image coordinate system |
| CN110189815A (en) * | 2019-05-29 | 2019-08-30 | 成都光魔科技有限公司 | A kind of case solid contrast teaching system of medical education |
| CN110335665A (en) * | 2019-04-01 | 2019-10-15 | 佛山原子医疗设备有限公司 | It is a kind of applied to medical image auxiliary diagnosis analysis to scheme to search drawing method and system |
| CN110335358A (en) * | 2019-06-18 | 2019-10-15 | 大连理工大学 | The personalized deformation method of deformable Digital Human anatomical model |
| CN110689521A (en) * | 2019-08-15 | 2020-01-14 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Automatic identification method and system for human body part to which medical image belongs |
| CN111414393A (en) * | 2020-03-26 | 2020-07-14 | 湖南科创信息技术股份有限公司 | Semantic similar case retrieval method and equipment based on medical knowledge graph |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8090166B2 (en) * | 2006-09-21 | 2012-01-03 | Surgix Ltd. | Medical image analysis |
| CN104866727A (en) * | 2015-06-02 | 2015-08-26 | 陈宽 | Deep learning-based method for analyzing medical data and intelligent analyzer thereof |
| CN106204520B (en) * | 2015-07-17 | 2019-05-24 | 北京大学第一医院 | A kind of method and device of magnetic resonance image selective positioning |
| US11202652B2 (en) * | 2017-08-11 | 2021-12-21 | Canon U.S.A., Inc. | Registration and motion compensation for patient-mounted needle guide |
| CN107742538A (en) * | 2017-10-10 | 2018-02-27 | 首都医科大学附属北京朝阳医院 | Lesion analogy method and device |
| US11382603B2 (en) * | 2018-05-07 | 2022-07-12 | Siemens Healthcare Gmbh | System and methods for performing biomechanically driven image registration using ultrasound elastography |
| CN110097944B (en) * | 2019-04-11 | 2021-05-07 | 山东大学齐鲁医院 | Display regulation and control method and system for human organ model |
| CN111583385B (en) * | 2020-04-15 | 2024-02-13 | 滨州医学院 | Personalized deformation method and system for deformable digital human anatomy model |
-
2020
- 2020-08-27 CN CN202010879567.6A patent/CN112070731B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3441939A1 (en) * | 2007-01-17 | 2019-02-13 | Mediguide Ltd. | A method and system for registering a 3d pre-acquired image coordinates system with a medical positioning system coordinate system and with a 2d image coordinate system |
| CN104603836A (en) * | 2012-08-06 | 2015-05-06 | 范德比尔特大学 | Enhanced method for correcting data for deformations during image guided procedures |
| CN105118068A (en) * | 2015-09-29 | 2015-12-02 | 常熟理工学院 | Medical image automatic annotation method under small sample condition |
| CN107103207A (en) * | 2017-04-05 | 2017-08-29 | 浙江大学 | Based on the multigroup accurate medical knowledge search system and implementation method for learning variation features of case |
| CN110335665A (en) * | 2019-04-01 | 2019-10-15 | 佛山原子医疗设备有限公司 | It is a kind of applied to medical image auxiliary diagnosis analysis to scheme to search drawing method and system |
| CN110189815A (en) * | 2019-05-29 | 2019-08-30 | 成都光魔科技有限公司 | A kind of case solid contrast teaching system of medical education |
| CN110335358A (en) * | 2019-06-18 | 2019-10-15 | 大连理工大学 | The personalized deformation method of deformable Digital Human anatomical model |
| CN110689521A (en) * | 2019-08-15 | 2020-01-14 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Automatic identification method and system for human body part to which medical image belongs |
| CN111414393A (en) * | 2020-03-26 | 2020-07-14 | 湖南科创信息技术股份有限公司 | Semantic similar case retrieval method and equipment based on medical knowledge graph |
Non-Patent Citations (3)
| Title |
|---|
| Viktor G＇al 等.Organ Detection in Medical Images with Discriminately Trained Deformable Part Model.《IEEE 9th International Conference on Computational Cybernetics》.2013, * |
| 基于图序列模型的病例近似搜索和疾病辅助诊断算法研究;高楚楚;《中国优秀硕士学位论文全文数据库 医药卫生科技辑》;20200715(第07期);E054-17 * |
| 基于机器学习的疾病自动诊断研究;刘巧利;《中国优秀硕士学位论文全文数据库 医药卫生科技辑》;20190915(第09期);E054-31 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112070731A (en) | 2020-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111091589B (en) | Ultrasonic and nuclear magnetic image registration method and device based on multi-scale supervised learning | |
| CN109166133B (en) | Soft tissue organ image segmentation method based on key point detection and deep learning | |
| McInerney et al. | Deformable models in medical image analysis | |
| CN108416802B (en) | Multimode medical image non-rigid registration method and system based on deep learning | |
| van Rikxoort et al. | Automatic segmentation of the liver in computed tomography scans with voxel classification and atlas matching | |
| CN110838140A (en) | Ultrasound and nuclear magnetic image registration fusion method and device based on hybrid supervised learning | |
| CN114972366B (en) | Full-automatic segmentation method and system for cerebral cortex surface based on graph network | |
| CN113570627B (en) | Training method of deep learning segmentation network and medical image segmentation method | |
| CN115457020B (en) | 2D medical image registration method fusing residual image information | |
| Luboz et al. | A segmentation and reconstruction technique for 3D vascular structures | |
| Jin et al. | Object recognition in medical images via anatomy-guided deep learning | |
| US20200357162A1 (en) | Modeling method, apparatus, device and storage medium of dynamic cardiovascular system | |
| CN110570430B (en) | Orbital bone tissue segmentation method based on volume registration | |
| CN117274599A (en) | Brain magnetic resonance segmentation method and system based on combined double-task self-encoder | |
| CN112164447B (en) | Image processing method, device, equipment and storage medium | |
| CN112070731B (en) | Method for guiding registration of human body model atlas and case CT image by artificial intelligence | |
| CN109741439A (en) | A kind of three-dimensional rebuilding method of two dimension MRI fetus image | |
| CN108961256A (en) | Image partition method, operation navigation device, electronic equipment and storage medium | |
| Gooya et al. | An EM algorithm for brain tumor image registration: A tumor growth modeling based approach | |
| Urschler et al. | Assessing breathing motion by shape matching of lung and diaphragm surfaces | |
| CN112598669B (en) | Lung lobe segmentation method based on digital human technology | |
| Ellingsen et al. | Robust deformable image registration using prior shape information for atlas to patient registration | |
| Lötjönen et al. | Four-chamber 3-D statistical shape model from cardiac short-axis and long-axis MR images | |
| Liu et al. | Organ labeling using anatomical model-driven global optimization | |
| Lelieveldt et al. | Anatomical modeling with fuzzy implicit surface templates: Application to automated localization of the heart and lungs in thoracic MR volumes |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |