CN117422647B - Heart CT image geometric calibration method and system based on machine learning - Google Patents
Heart CT image geometric calibration method and system based on machine learning Download PDFInfo
- Publication number
- CN117422647B CN117422647B CN202311217461.XA CN202311217461A CN117422647B CN 117422647 B CN117422647 B CN 117422647B CN 202311217461 A CN202311217461 A CN 202311217461A CN 117422647 B CN117422647 B CN 117422647B
- Authority
- CN
- China
- Prior art keywords
- heart
- image
- model
- dimensional
- calibrated
- 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
- 238000010801 machine learning Methods 0.000 title claims abstract description 62
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000005094 computer simulation Methods 0.000 claims abstract description 11
- 238000002474 experimental method Methods 0.000 claims abstract description 11
- 238000000338 in vitro Methods 0.000 claims abstract description 11
- 238000012549 training Methods 0.000 claims abstract description 9
- 230000000747 cardiac effect Effects 0.000 claims description 23
- 238000012937 correction Methods 0.000 claims description 20
- 238000013135 deep learning Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 5
- 238000013136 deep learning model Methods 0.000 claims description 3
- 238000002591 computed tomography Methods 0.000 abstract description 139
- 230000003068 static effect Effects 0.000 abstract description 14
- 238000012795 verification Methods 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 7
- 238000010009 beating Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 230000010247 heart contraction Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 239000013013 elastic material Substances 0.000 description 2
- 210000000056 organ Anatomy 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000010146 3D printing Methods 0.000 description 1
- 210000003484 anatomy Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30048—Heart; Cardiac
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a geometric calibration method and a geometric calibration system for a heart CT image based on machine learning, which relate to the field of Computer Tomography (CT), and comprise the following steps: building a heart in-vitro dynamic simulation experiment platform; inputting a plurality of different scanning parameters and heart rate parameters into an experimental platform, and obtaining a heart CT dynamic scanning image as an image database to be calibrated; scanning or shooting an experimental platform by using a static scanning or high-speed camera to obtain a standard image database consistent with the dynamic scanning image period of the heart CT; constructing a heart CT image geometric calibration model based on machine learning, and sending an image database to be calibrated and a standard image database into the heart CT geometric calibration model for training and verification; and acquiring a trained geometric calibration model of the heart CT image, and inputting an individualized heart CT dynamic scanning image of the patient to obtain a calibrated real CT image. The invention can realize the rapid and accurate calibration of the CT image of the dynamic scanning of the heart.
Description
Technical Field
The invention relates to the field of Computer Tomography (CT), in particular to a geometric calibration method and system for a cardiac CT image based on machine learning.
Background
With the continued development of science and technology, computed tomography (computed tomography, CT) is increasingly being used for imaging of moving organs or tissues of the human body. The heart is one of the organs with the highest motion frequency of the human body, and the heart continuously beats at the frequency of 60 to 100 times per minute, so that the CT scan image obtained by reconstruction at each angle has larger motion artifact.
In the prior art, in order to inhibit motion artifact, a heart gating technology is generally used in clinic to select CT scan data obtained by scanning at a relatively stable moment of heart motion for reconstruction, so as to obtain a heart image.
However, due to limitations of the number of rows of CT detectors, the rotational speed, and the high heart rate of patients, scanning of the entire heart cannot be completed in the same cardiac cycle, and thus the obtained CT image may be affected by the heart beat, so that the reconstructed heart shape may not conform to the actual heart shape, and thus the observation of the patient-specific anatomical structure by the clinician is affected to diagnose the disease, and the accuracy of the image calibration method is further affected due to the non-conforming to the actual heart shape.
Disclosure of Invention
The embodiment of the invention provides a geometric calibration method and a geometric calibration system for a cardiac CT image based on machine learning, which can solve the problem that the cardiac dynamic scanning CT image cannot be accurately calibrated in the prior art.
The embodiment of the invention provides a geometric calibration method of a heart CT image based on machine learning, which comprises the following steps: acquiring heart CT dynamic scanning images under different scanning parameters and heart rates on an in-vitro dynamic simulation experiment platform of the heart, and taking the heart CT dynamic scanning images as an image database to be calibrated;
acquiring a heart CT image obtained by static scanning of a CT machine consistent with the period of a heart CT dynamic scanning image or a two-dimensional image obtained by multi-azimuth shooting of a heart model by a high-speed camera, and taking the two-dimensional image as a standard image database;
When a heart CT image obtained by static scanning of a CT machine is used as a standard image database, training a first heart CT image geometric calibration model based on machine learning through the image database to be calibrated; inputting an original CT image of a heart to be calibrated into a first geometric calibration model of the heart CT image based on machine learning to obtain a calibrated heart CT image;
Wherein the first machine learning based cardiac CT image geometric calibration model comprises: a multi-modal self-encoder and decoder interconnected;
when a two-dimensional image which is shot in multiple directions by the high-speed camera is used as a standard image database, training a second geometric calibration model of the heart CT image based on machine learning through the image database to be calibrated; inputting the heart dynamic scanning CT image to be calibrated into a second geometric calibration model of the heart CT image based on machine learning, outputting a three-dimensional correction vector field, and superposing the three-dimensional correction vector field on the heart CT dynamic scanning image to be calibrated to obtain a calibrated heart CT image;
the construction of the second machine learning-based cardiac CT image geometric calibration model comprises the following steps:
Automatically reconstructing a heart three-dimensional geometric model to be corrected according to an image database to be calibrated, and automatically reconstructing a standard heart three-dimensional geometric model according to a standard image database to construct a three-dimensional automatic reconstruction model based on deep learning;
Performing spatial registration on the heart three-dimensional geometric model to be calibrated and the standard heart three-dimensional geometric model to construct a geometric registration model based on depth learning;
And inputting the heart three-dimensional geometric model to be calibrated and the registered standard heart three-dimensional geometric model into an unsupervised machine learning model formed by a multi-mode self-encoder and a decoder to obtain a three-dimensional correction vector field, and constructing a three-dimensional correction model based on machine learning.
Further, the three-dimensional automatic reconstruction model based on the deep learning specifically comprises: 3D UNet, VNet, GAN deep learning model.
Further, the mode of obtaining the image database to be calibrated is constructed by heart CT dynamic scanning images obtained under the assistance of a plurality of different parameters by an electrocardiographic gating technology.
Further, the parameters include: CT number of rows, scan speed, and heart rate.
Furthermore, the heart CT image obtained by static scanning of the CT machine or the two-dimensional image obtained by multi-azimuth shooting of the heart model by the high-speed camera can accurately reflect the form of the heart under the real beating.
The embodiment of the invention provides a heart CT image geometric calibration system based on machine learning, which comprises the following steps: the first data unit is used for acquiring heart CT dynamic scanning images under different scanning parameters and heart rates on an in-vitro dynamic simulation experiment platform of the heart and taking the heart CT dynamic scanning images as an image database to be calibrated;
The second data unit is used for acquiring a heart CT image obtained by static scanning of the CT machine consistent with the dynamic scanning image period of the heart CT or a two-dimensional image obtained by multi-azimuth shooting of the heart model by the high-speed camera, and taking the two-dimensional image as a standard image database;
The first model unit trains a first geometric calibration model of the heart CT image based on machine learning through an image database to be calibrated when the heart CT image obtained by static scanning of the CT machine is used as a standard image database; inputting an original CT image of a heart to be calibrated into a first geometric calibration model of the heart CT image based on machine learning to obtain a calibrated heart CT image; wherein the first machine learning based cardiac CT image geometric calibration model comprises: a multi-modal self-encoder and decoder interconnected;
The second model unit is used for training a second geometric calibration model of the heart CT image based on machine learning through the image database to be calibrated when a two-dimensional image which is shot in multiple directions by the high-speed camera is used as a standard image database; inputting the heart dynamic scanning CT image to be calibrated into a second geometric calibration model of the heart CT image based on machine learning, outputting a three-dimensional correction vector field, and superposing the three-dimensional correction vector field on the heart CT dynamic scanning image to be calibrated to obtain a calibrated heart CT image; the construction of the second machine learning-based cardiac CT image geometric calibration model comprises the following steps: automatically reconstructing a heart three-dimensional geometric model to be corrected according to an image database to be calibrated, and automatically reconstructing a standard heart three-dimensional geometric model according to a standard image database to construct a three-dimensional automatic reconstruction model based on deep learning; performing spatial registration on the heart three-dimensional geometric model to be calibrated and the standard heart three-dimensional geometric model to construct a geometric registration model based on depth learning; and inputting the heart three-dimensional geometric model to be calibrated and the registered standard heart three-dimensional geometric model into an unsupervised machine learning model formed by a multi-mode self-encoder and a decoder to obtain a three-dimensional correction vector field, and constructing a three-dimensional correction model based on machine learning.
The embodiment of the invention provides a geometric calibration method and a geometric calibration system for a heart CT image based on machine learning, which have the following beneficial effects compared with the prior art:
Aiming at the problem of artifacts caused by heart beating, which is commonly existed in heart CT images, the invention provides a geometric calibration method of heart CT images based on machine learning, which is to train a first geometric calibration model and a second geometric calibration model of heart CT images based on machine learning through an image database to be calibrated, so as to obtain the calibrated heart CT images. Therefore, the rapid and accurate geometric calibration of the cardiac dynamic scanning CT image is realized, the calibrated cardiac CT image can truly reflect the geometric form of the heart, the calibrated real image can be obtained by inputting the individualized cardiac CT dynamic scanning image of the patient, and a more reliable and accurate real image without motion artifacts of the beating heart is provided for a clinician to carry out disease diagnosis and treatment decision.
The built in-vitro dynamic simulation experiment platform for the heart can truly simulate the heart beating of a human body, and the HU value of the heart model is close to the HU value of the real heart of the human body by adopting an elastic material 3D printed heart model with similar density to the heart of the human body.
And acquiring heart CT dynamic scanning images under parameters including CT ranking, scanning speed and heart rate by using an electrocardiographic gating technology, so that the constructed image database to be calibrated can basically cover heart CT clinical examination. The heart CT image obtained by static scanning of the CT machine or the two-dimensional image obtained by multi-azimuth shooting of the heart model by the high-speed camera can accurately reflect the form of the heart under the real beating.
Drawings
FIG. 1 is a schematic diagram of a method and a system for geometric calibration of cardiac CT images based on machine learning according to an embodiment of the present invention;
FIG. 2 is a diagram showing a method and a system for geometric calibration of cardiac CT images based on machine learning (a standard image database is obtained by static scanning);
fig. 3 is a diagram of a geometric calibration method and a system calibration model for cardiac CT images based on machine learning (a standard image database is obtained by a high-speed camera) according to an embodiment of the present invention.
Detailed Description
In order that the above objects, features and advantages of the invention will be readily understood, a more particular description of the invention will be rendered by reference to the appended drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The present invention may be embodied in many other forms than described herein and similarly modified by those skilled in the art without departing from the spirit of the invention, whereby the invention is not limited to the specific embodiments disclosed below.
In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings are merely for convenience in describing the present invention and simplifying the description, and do not indicate or imply that the device or element being referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention.
Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In the description of the present invention, the meaning of "plurality" means at least two, for example, two, three, etc., unless specifically defined otherwise.
In the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," "secured," and the like are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; either directly or indirectly, through intermediaries, or both, may be in communication with each other or in interaction with each other, unless expressly defined otherwise. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.
In the present invention, unless expressly stated or limited otherwise, a first feature "up" or "down" a second feature may be the first and second features in direct contact, or the first and second features in indirect contact via an intervening medium. Moreover, a first feature being "above," "over" and "on" a second feature may be a first feature being directly above or obliquely above the second feature, or simply indicating that the first feature is level higher than the second feature. The first feature being "under", "below" and "beneath" the second feature may be the first feature being directly under or obliquely below the second feature, or simply indicating that the first feature is less level than the second feature.
It will be understood that when an element is referred to as being "fixed" or "disposed" on another element, it can be directly on the other element or intervening elements may also be present. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present. The terms "vertical," "horizontal," "upper," "lower," "left," "right," and the like are used herein for illustrative purposes only and are not meant to be the only embodiment.
Referring to fig. 1 to 3, an embodiment of the present invention provides a method and a system for geometric calibration of a cardiac CT image based on machine learning, the method and the system for geometric calibration of a cardiac CT image based on machine learning include the following steps: and (5) constructing a heart in-vitro dynamic simulation experiment platform. And acquiring a plurality of heart CT dynamic scanning images with different scanning parameters and heart rate by using an electrocardiographic gating technology as an image database to be calibrated. A standard image database consistent with the cardiac CT dynamic scan image session is acquired using a static scan or high speed camera. And constructing a geometric calibration model of the heart CT image based on machine learning, training and verifying the geometric calibration model of the heart CT by using an image database to be calibrated and related parameters as samples and using a standard image database consistent with the dynamic scanning image period of the heart CT as a label. The improvement is that a trained heart CT image geometric calibration model is obtained, and a calibrated real CT image can be obtained by inputting an individualized heart CT dynamic scanning image of a patient.
Further, the in-vitro dynamic simulation experiment platform for the heart comprises a CT machine and a heart simulation device, and a controller for controlling the movement of the heart simulation device. The improvement is that the heart model is formed by 3D printing of an elastic material with the density similar to that of a human heart, so that the HU value of each part in the heart model and the HU value of a human real heart are close to each other.
Further, the image database to be calibrated is constructed by heart CT dynamic scanning images obtained under the assistance of a plurality of different scanning parameters and heart rate electrocardio gating technologies. The improvement is that parameters include, but are not limited to, CT number of rows and scan speed.
Further, heart CT dynamic scanning images under different scanning parameters and heart rates are obtained on a heart in-vitro dynamic simulation experiment platform and used as an image database to be calibrated; acquiring a heart CT image obtained by static scanning of a CT machine consistent with the period of a heart CT dynamic scanning image or a two-dimensional image obtained by multi-azimuth shooting of a heart model by a high-speed camera, and taking the two-dimensional image as a standard image database;
Further, when a heart CT image obtained by static scanning of the CT machine is used as a standard image database, training a first heart CT image geometric calibration model based on machine learning through the image database to be calibrated; inputting an original CT image of a heart to be calibrated into a first geometric calibration model of the heart CT image based on machine learning to obtain a calibrated heart CT image; wherein the first machine learning based cardiac CT image geometric calibration model comprises: a multi-modal self-encoder and decoder interconnected;
further, when a two-dimensional image which is shot in multiple directions by the high-speed camera is used as a standard image database, training a second geometric calibration model of the heart CT image based on machine learning through the image database to be calibrated; inputting the heart dynamic scanning CT image to be calibrated into a second geometric calibration model of the heart CT image based on machine learning, outputting a three-dimensional correction vector field, and superposing the three-dimensional correction vector field on the heart CT dynamic scanning image to be calibrated to obtain a calibrated heart CT image; the construction of the second machine learning-based cardiac CT image geometric calibration model comprises the following steps:
Automatically reconstructing a heart three-dimensional geometric model to be corrected according to an image database to be calibrated, and automatically reconstructing a standard heart three-dimensional geometric model according to a standard image database to construct a three-dimensional automatic reconstruction model based on deep learning;
Performing spatial registration on the heart three-dimensional geometric model to be calibrated and the standard heart three-dimensional geometric model to construct a geometric registration model based on depth learning;
And inputting the heart three-dimensional geometric model to be calibrated and the registered standard heart three-dimensional geometric model into an unsupervised machine learning model formed by a multi-mode self-encoder and a decoder to obtain a three-dimensional correction vector field, and constructing a three-dimensional correction model based on machine learning.
Further, after the first and second geometric calibration models of the heart CT image based on machine learning are constructed, the image database to be calibrated is used as a sample, the standard image database is used as a label, the sample and the label are input into the geometric calibration model of the heart CT image based on machine learning for training, and finally, the output result obtained from the trained geometric calibration model of the heart CT image is expected to be the standard image conforming to the label.
Further, a trained geometric calibration model of the heart CT image is obtained, and an individualized heart CT dynamic scanning image of the patient is input to obtain a calibrated real CT image.
Furthermore, the invention provides a geometric calibration method of the heart CT image based on machine learning aiming at the problem of artifacts caused by heart beating commonly existing in the heart CT image, which can quickly and accurately eliminate the artifacts, so that the calibrated heart CT image can truly reflect the geometric form of the heart, and more reliable and accurate patient-specific anatomical information is provided for a clinician to perform disease diagnosis and treatment decision.
Furthermore, the image database to be calibrated constructed by the invention is composed of heart CT dynamic scanning images obtained under the assistance of a plurality of different parameters by an electrocardiographic gating technology, and can basically cover heart CT clinical examination. The constructed standard image database is a heart CT image obtained by static scanning of a CT machine under the condition that the heart model is kept from any deformation at the dynamic scanning starting moment, or a two-dimensional image obtained by multi-azimuth shooting of the heart model by a high-speed camera at the dynamic scanning starting moment, so that the form of the heart under real beating can be accurately reflected. The constructed geometric calibration model of the heart CT image based on machine learning can quickly realize quick geometric calibration of the heart dynamic scanning CT image, and a calibrated real image can be obtained by inputting the patient individuation heart CT dynamic scanning image.
One specific example is as follows:
firstly, a heart in-vitro dynamic simulation experiment platform is built.
Secondly, an image database to be calibrated is constructed in a mode of acquiring a plurality of heart CT dynamic scanning images under different scanning parameters, heart rate and other parameters as the image database to be calibrated.
Then, a standard image database is constructed by acquiring the standard image database consistent with the dynamic scanning image period of the heart CT by using a static scanning or high-speed camera.
And then, constructing a geometric calibration model of the heart CT image based on machine learning, utilizing an image database to be calibrated and related parameters as samples, and utilizing a corresponding standard image database consistent with the dynamic scanning image period of the heart CT as a label to train and verify the geometric calibration model of the heart CT.
Finally, the obtained calibrated heart CT image can truly reflect the geometric form of the heart, so that the whole method can obtain a calibrated real image by inputting an individualized heart CT dynamic scanning image of a patient.
The invention aims to provide a geometric calibration method and a geometric calibration system for a heart CT image based on machine learning, which are characterized in that a heart in-vitro dynamic simulation experiment platform is built to obtain a heart CT dynamic scanning image data set to be calibrated, and a standard image data set obtained by a static scanning or high-speed camera is utilized to realize rapid and accurate calibration of the heart CT dynamic scanning image by utilizing a geometric calibration model for the heart CT image based on machine learning.
The above examples illustrate only a few embodiments of the invention, which are described in detail and are not to be construed as limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of protection of the present invention is to be determined by the appended claims.
Claims (4)
1. The geometric calibration method of the heart CT image based on machine learning is characterized by comprising the following steps of:
Acquiring heart CT dynamic scanning images under different scanning parameters and heart rates on an in-vitro dynamic simulation experiment platform of the heart, and taking the heart CT dynamic scanning images as an image database to be calibrated;
acquiring a two-dimensional image of the heart model shot in multiple directions by a high-speed camera, and taking the two-dimensional image as a standard image database;
When a two-dimensional image which is shot in multiple directions by the high-speed camera is used as a standard image database, training a geometric calibration model of the heart CT image based on machine learning through the image database to be calibrated; inputting a heart CT image to be calibrated into a geometric calibration model of the heart CT image based on machine learning, outputting a three-dimensional correction vector field, and superposing the three-dimensional correction vector field on the heart CT image to be calibrated to obtain a calibrated heart CT image;
The construction of the geometric calibration model of the heart CT image based on machine learning comprises the following steps:
Automatically reconstructing a heart three-dimensional geometric model to be corrected according to an image database to be calibrated, and automatically reconstructing a standard heart three-dimensional geometric model according to a standard image database to construct a three-dimensional automatic reconstruction model based on deep learning, wherein the three-dimensional automatic reconstruction model based on the deep learning specifically comprises: 3D UNet, VNet and GAN deep learning models;
Performing spatial registration on the heart three-dimensional geometric model to be calibrated and the standard heart three-dimensional geometric model to construct a geometric registration model based on depth learning;
And inputting the heart three-dimensional geometric model to be calibrated and the registered standard heart three-dimensional geometric model into an unsupervised machine learning model formed by a multi-mode self-encoder and a decoder to obtain a three-dimensional correction vector field, and constructing a three-dimensional correction model based on machine learning.
2. The method for geometric calibration of cardiac CT images based on machine learning as recited in claim 1 wherein said database of images to be calibrated is constructed from cardiac CT dynamic scan images obtained under assistance of a plurality of different parameters by means of an electrocardiographic gating technique.
3. A machine learning based cardiac CT image geometry calibration method as recited in claim 2 in which said parameters include: CT number of rows, scan speed, and heart rate.
4. A machine learning based cardiac CT image geometry calibration system, comprising:
The first data unit is used for acquiring heart CT dynamic scanning images under different scanning parameters and heart rates on an in-vitro dynamic simulation experiment platform of the heart and taking the heart CT dynamic scanning images as an image database to be calibrated;
the second data unit is used for acquiring a two-dimensional image of the heart model shot in multiple directions by the high-speed camera, and taking the two-dimensional image as a standard image database;
The model unit trains a geometric calibration model of the heart CT image based on machine learning through the image database to be calibrated when a two-dimensional image which is shot in multiple directions by the high-speed camera is used as a standard image database; inputting a heart CT image to be calibrated into a geometric calibration model of the heart CT image based on machine learning, outputting a three-dimensional correction vector field, and superposing the three-dimensional correction vector field on the heart CT image to be calibrated to obtain a calibrated heart CT image; the construction of the geometric calibration model of the heart CT image based on machine learning comprises the following steps: automatically reconstructing a heart three-dimensional geometric model to be corrected according to an image database to be calibrated, and automatically reconstructing a standard heart three-dimensional geometric model according to a standard image database to construct a three-dimensional automatic reconstruction model based on deep learning, wherein the three-dimensional automatic reconstruction model based on the deep learning specifically comprises: 3D UNet, VNet and GAN deep learning models; performing spatial registration on the heart three-dimensional geometric model to be calibrated and the standard heart three-dimensional geometric model to construct a geometric registration model based on depth learning; and inputting the heart three-dimensional geometric model to be calibrated and the registered standard heart three-dimensional geometric model into an unsupervised machine learning model formed by a multi-mode self-encoder and a decoder to obtain a three-dimensional correction vector field, and constructing a three-dimensional correction model based on machine learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311217461.XA CN117422647B (en) | 2023-09-20 | 2023-09-20 | Heart CT image geometric calibration method and system based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311217461.XA CN117422647B (en) | 2023-09-20 | 2023-09-20 | Heart CT image geometric calibration method and system based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117422647A CN117422647A (en) | 2024-01-19 |
| CN117422647B true CN117422647B (en) | 2024-07-16 |
Family
ID=89531594
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311217461.XA Active CN117422647B (en) | 2023-09-20 | 2023-09-20 | Heart CT image geometric calibration method and system based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117422647B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111462020A (en) * | 2020-04-24 | 2020-07-28 | 上海联影医疗科技有限公司 | Method, system, storage medium and device for correcting motion artifact of heart image |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8938111B2 (en) * | 2010-01-13 | 2015-01-20 | Fei Company | Computed tomography imaging process and system |
| DE102016207367A1 (en) * | 2016-04-29 | 2017-11-02 | Siemens Healthcare Gmbh | Specify scan parameters of a CT image capture using an external image capture |
| US11517197B2 (en) * | 2017-10-06 | 2022-12-06 | Canon Medical Systems Corporation | Apparatus and method for medical image reconstruction using deep learning for computed tomography (CT) image noise and artifacts reduction |
| CN110458779B (en) * | 2019-08-08 | 2021-11-16 | 江苏赛诺格兰医疗科技有限公司 | Method for acquiring correction information for attenuation correction of PET images of respiration or heart |
| JP7455542B2 (en) * | 2019-09-27 | 2024-03-26 | キヤノン株式会社 | Image processing method, program, image processing device, learned model manufacturing method, and image processing system |
| US20210177295A1 (en) * | 2019-12-11 | 2021-06-17 | GE Precision Healthcare LLC | Systems and methods for generating diagnostic scan parameters from calibration images |
| US20210304457A1 (en) * | 2020-03-31 | 2021-09-30 | The Regents Of The University Of California | Using neural networks to estimate motion vectors for motion corrected pet image reconstruction |
| CN113256753B (en) * | 2021-06-30 | 2021-10-29 | 之江实验室 | PET image region-of-interest enhancement reconstruction method based on multitask learning constraint |
| CN114820857A (en) * | 2022-05-06 | 2022-07-29 | 山东理工大学 | Bimodal feature set construction method for machine learning |
| CN116188614A (en) * | 2023-02-23 | 2023-05-30 | 北京师范大学 | Magnetic sensitivity artifact correction method and device based on deep learning |
| CN116740218B (en) * | 2023-08-11 | 2023-10-27 | 南京安科医疗科技有限公司 | Heart CT imaging image quality optimization method, device and medium |
-
2023
- 2023-09-20 CN CN202311217461.XA patent/CN117422647B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111462020A (en) * | 2020-04-24 | 2020-07-28 | 上海联影医疗科技有限公司 | Method, system, storage medium and device for correcting motion artifact of heart image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117422647A (en) | 2024-01-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7271000B2 (en) | Patch guide method and program | |
| JP7263324B2 (en) | Method and program for generating 3D brain map | |
| CN105025803B (en) | Segmentation from multiple 3-D views to blob | |
| US5889524A (en) | Reconstruction of three-dimensional objects using labeled piecewise smooth subdivision surfaces | |
| CN105451663B (en) | Guidance is fed back to the ultrasound acquisition of target view | |
| JP2020021488A (en) | Left atrium shape reconstruction from sparse location measurement using neural networks | |
| US9747689B2 (en) | Image processing system, X-ray diagnostic apparatus, and image processing method | |
| EP1832233A1 (en) | Cardiac valve data measuring method and device | |
| JP2008534082A (en) | Systems, methods and media for acquiring and generating standard ultrasound images independent of operators of fetal, neonatal and adult organs | |
| JP7463395B2 (en) | DEVICE AND METHOD FOR ANALYZING PHOTOACOUS DATA, PHOTOACOUS SYSTEM AND COMPUTER PROGRAM - Patent application | |
| US8655040B2 (en) | Integrated image registration and motion estimation for medical imaging applications | |
| JP4584232B2 (en) | Brain model | |
| Liao et al. | Analysis of surface geometry of the human stomach using real‐time 3‐D ultrasonography in vivo | |
| CN109770943A (en) | A kind of ultrasonic automatic optimization method positioned using computer vision | |
| Remme et al. | Extraction and quantification of left ventricular deformation modes | |
| KR20190125592A (en) | 3D Blood Vessel Construction Method using medical images | |
| CN114943688A (en) | Method for extracting interest region in mammary gland image based on palpation and ultrasonic data | |
| CN113057677B (en) | Heart image modeling method, system and equipment for fusing ultrasonic image and CT image | |
| JP4879872B2 (en) | Image processing apparatus, image processing program, storage medium, and ultrasonic diagnostic apparatus | |
| US9521980B2 (en) | Method for registering medical images, apparatus performing the method, and computer readable media including the method | |
| CN117422647B (en) | Heart CT image geometric calibration method and system based on machine learning | |
| CN115300809B (en) | Image processing method and device, computer equipment and storage medium | |
| French et al. | 3D perfusion mapping in post-infarct mice using myocardial contrast echocardiography | |
| Biswas et al. | Clinical utility of three-dimensional echocardiography for the evaluation of ventricular function | |
| Shanmuganathan et al. | Image registration for multi-view three-dimensional echocardiography sequences |
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 |