CN108159576B - Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy - Google Patents

Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy Download PDF

Info

Publication number
CN108159576B
CN108159576B CN201711359660.9A CN201711359660A CN108159576B CN 108159576 B CN108159576 B CN 108159576B CN 201711359660 A CN201711359660 A CN 201711359660A CN 108159576 B CN108159576 B CN 108159576B
Authority
CN
China
Prior art keywords
characteristic
region
respiratory motion
prediction
points
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.)
Expired - Fee Related
Application number
CN201711359660.9A
Other languages
Chinese (zh)
Other versions
CN108159576A (en
Inventor
于晓洋
史领
韩玉翠
赵烟桥
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin University of Science and Technology
Original Assignee
Harbin University of Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Harbin University of Science and Technology filed Critical Harbin University of Science and Technology
Priority to CN201711359660.9A priority Critical patent/CN108159576B/en
Publication of CN108159576A publication Critical patent/CN108159576A/en
Application granted granted Critical
Publication of CN108159576B publication Critical patent/CN108159576B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N5/00—Radiation therapy
    • A61N5/10—X-ray therapy; Gamma-ray therapy; Particle-irradiation therapy
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N5/00—Radiation therapy
    • A61N5/10—X-ray therapy; Gamma-ray therapy; Particle-irradiation therapy
    • A61N5/1048—Monitoring, verifying, controlling systems and methods
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N5/00—Radiation therapy
    • A61N5/10—X-ray therapy; Gamma-ray therapy; Particle-irradiation therapy
    • A61N5/1048—Monitoring, verifying, controlling systems and methods
    • A61N5/1049—Monitoring, verifying, controlling systems and methods for verifying the position of the patient with respect to the radiation beam
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G06T7/248—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments involving reference images or patches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G06T7/74—Determining position or orientation of objects or cameras using feature-based methods involving reference images or patches
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N5/00—Radiation therapy
    • A61N5/10—X-ray therapy; Gamma-ray therapy; Particle-irradiation therapy
    • A61N5/1048—Monitoring, verifying, controlling systems and methods
    • A61N5/1049—Monitoring, verifying, controlling systems and methods for verifying the position of the patient with respect to the radiation beam
    • A61N2005/1059—Monitoring, verifying, controlling systems and methods for verifying the position of the patient with respect to the radiation beam using cameras imaging the patient
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N5/00—Radiation therapy
    • A61N5/10—X-ray therapy; Gamma-ray therapy; Particle-irradiation therapy
    • A61N2005/1092—Details
    • 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
    • G06T2207/10016—Video; Image sequence
    • 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
    • G06T2207/10024—Color image
    • 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
    • G06T2207/30096—Tumor; Lesion

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Veterinary Medicine (AREA)
  • Radiology & Medical Imaging (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Pathology (AREA)
  • Public Health (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Radiation-Therapy Devices (AREA)

Abstract

The invention relates to a method for predicting respiratory motion of a chest and abdomen surface area of a human body in radiotherapy, belonging to the field of tumor treatment and respiratory motion prediction; the prediction method with the respiratory motion prediction function adopts the double cameras to collect respiratory motion signal data, preprocesses the collected data, and uses the obtained data as the input of the Gaussian process regression prediction algorithm. Then, a prediction is carried out on the regression problem based on the Gaussian process model. And constructing a proper regression model by using the data, training and predicting, selecting a kernel function in the training process, solving the hyperparameter in the kernel function, and verifying the feasibility of the algorithm through an off-line simulation experiment. And finally, selecting a quasiperiodic kernel function, solving the hyperparameter by using a conjugate gradient method, and performing model prediction on the respiratory motion data by using a Gaussian process regression method. The method for predicting the respiratory motion of the chest and abdomen surface area of the human body in the radiotherapy can reduce the influence of the respiratory motion on the radiotherapy precision.

Description

Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy
Technical Field
A method for predicting respiratory motion of a chest and abdomen surface area of a human body in radiotherapy belongs to the technical field of respiratory motion prediction.
Background
Respiratory motion can cause tumor and normal tissue to move at a certain frequency and amplitude, which causes a difference between the actual absorbed dose in radiotherapy and the calculation result of the planning system and affects the treatment effect, even causes radiotherapy damage to the human body. Respiratory motion is the main reason for the position movement and volume change of the thoracoabdominal tumor during the radiation therapy, the phenomena of target area missing illumination and normal tissue illuminated increase are aggravated, and the difference between the actual absorbed dose and the planned dose is caused by the deformation and movement of the tumor caused by the respiratory motion, so the influence on the dose effect is very obvious clinically.
The current conventional methods for treating respiratory motion in radiotherapy include: clinical techniques such as exercise inclusion method, compression type shallow breathing method, breath holding method and respiratory gating method have already produced positive effects on thoracoabdominal tumor radiotherapy, but the expected effects are not yet achieved due to poor tolerance of patients, low treatment efficiency, large damage to normal tissues and the like.
There are two types of algorithms for the prediction of respiratory motion implemented. The first category, prediction models based on mathematical and physical methods; the second category is predictive models based on modern scientific techniques and methods. The former includes linear estimation algorithm and arima (automated Integrated Moving operation Model) in the time series Model, which is the most widely used one of time series Model, subsequence matching Model, Kalman Filtering Model (Kalman Filtering Model), Parametric regression Model (Parametric regression Model), Exponential Smoothing Model (explicit Smoothing Model) and various combined prediction models formed by these models. The latter Model includes nonparametric regression Model, KARIMA algorithm, adaptive weight Model, spectrum analysis, state empty I sentence reconstruction Model, Wavelet Network, multidimensional fractal based method, Wavelet decomposition and reconstruction based method, and various composite prediction models related to Neural Network.
However, compared with these models, the gaussian process has the advantage that firstly, the priori knowledge in the process can be expressed in the form of a priori probability, and the performance of the model is improved. Secondly, an output prediction with precision parameters can be made for unknown input items, wherein the precision parameters mainly refer to estimated variance model parameters and are expressed by the characteristics of obvious parameter reduction, relatively easy parameter optimization, easier parameter convergence and the like. And through the continuous exploration of scholars at home and abroad in nearly 10 years, the Gaussian process is approved in practice, and the application degree in supervised learning is obviously improved.
Disclosure of Invention
In order to solve the problem, the utility model provides a human chest abdomen surface area respiratory motion prediction method in radiotherapy, this respiratory motion predictive analysis based on the gaussian process through constructing correlation model and prediction model and combining both and finally constructing a more accurate prediction method, and then for the control with predict tumour respiratory motion during radiotherapy provide better basis, solved the problem that arouses by respiratory motion in the radiotherapy process.
The purpose of the invention is realized as follows:
a method for predicting respiratory motion of a chest and abdomen surface area of a human body in radiotherapy comprises the following steps:
step a, a color pattern formed by compounding three RGB cosine curves with different frequencies is projected to the surface of the chest and abdomen of a human body by adopting a light pattern projector, a 3CCD color camera is respectively arranged at two sides of the light pattern projector to collect scene images and send the scene images to a computer for post-processing, and three-dimensional coordinates of characteristic mark points and area boundary lines are obtained by adopting two cameras according to a binocular vision principle;
b, aiming at two video sequences obtained by the left camera and the right camera, extracting the same region, the boundary line and the characteristic mark points of the region from the corresponding image pairs, respectively matching the boundary line and the characteristic mark points of the region, then obtaining three-dimensional coordinates of points on the boundary line and the characteristic mark points by adopting the left camera and the right camera according to a binocular vision principle, obtaining three-dimensional coordinates of surface points in the region by adopting a camera and projector combination according to a stripe analysis and phase expansion method, and calculating 7 region characteristic quantities of the three-dimensional coordinates of the characteristic points, the horizontal projection perimeter and the geometric center of the boundary line of the region, the average value and the perimeter of the coordinates of each point of the boundary line of the region, the average value of the coordinates;
step c, determining an interested area, a boundary thereof and a characteristic mark point according to the position of the specific prediction characteristic quantity; taking the predicted characteristic quantity training observation value as a reference, carrying out correlation analysis and significance analysis on the predicted characteristic quantity training observation value and all other regional characteristic quantity training observation values, and optimizing a regional characteristic quantity set which participates in modeling and prediction and recording as Y;
step d, selecting the quasiperiodic kernel function as follows
Figure GDA0002217107960000021
Wherein r ═||x-x'||2Representing the Euclidean distance, θ, between two data pointsS、θL、θpIs a hyper-parameter;
step e, to ensure KcFor an effective positive definite covariance function, Cholesky decomposition is used and the elements of the lower triangular matrix are parameterized to obtain Kc＝L(θc)L(θc)TWherein L (θ)c) Is a lower triangular matrix with the size of m multiplied by m;
f, minimizing the negative logarithmic edge probability with respect to the hyper-parameter, namely-log (y | theta), and then solving the optimal value of the negative logarithmic edge probability by adopting a conjugate gradient method;
step g of giving the hyper-parameter θ by maximizing the cross-correlation between the predicted region feature quantity and the other region feature quantitiesSInitial value, giving the hyper-parameter theta according to the training data of the characteristic quantities of other regionspAssigning an initial value, repeating the experiment for multiple times, and initializing other hyper-parameters randomly;
step h, in the measuring and predicting stage, firstly measuring other region characteristic quantities except the predicted region characteristic quantity according to the sampling frequency, and then aiming at the predicted time x*And predicting the predicted value and the error confidence interval of the predicted value, wherein the t is the prediction time.
The method for predicting respiratory motion of chest and abdomen surface area of human body in radiotherapy comprisesc) Is composed of a non-zero element of thetacTo specify the correlation hyperparameter thetacIs given by
Figure GDA0002217107960000031
In the method for predicting the respiratory motion of the chest and abdomen surface area of the human body in the radiotherapy, the sufficient necessary condition for forming the kernel function is that a matrix formed between a test concentrated point and a point must be a semi-positive definite matrix.
Has the advantages that: the accurate prediction of the tumor depends on a perfect real-time tracking prediction system to a great extent, the system can help a doctor to track the motion state of the tumor in real time and predict the position of the tumor at the next moment in time, the readjustment of a radiation beam is facilitated, the damage to healthy tissues is avoided, and the system is an important method for improving the curative effect of tumor treatment. The high-precision prediction algorithm plays a crucial role in readjusting the radiation beam, and can ensure that the radiation beam has minimum damage to normal tissues in the process of completely killing tumor cancer cells, which has great clinical significance.
Through comparison and analysis at home and abroad, the mathematical Gaussian process and a large-scale neural network, a Bayesian model and the like are found to be equivalent to those known by people. However, compared with these models, the gaussian process regression prediction model adopted in the invention has the advantages that firstly, the priori knowledge in the process can be expressed in the form of the prior probability, and the performance of the model is improved. Secondly, an output prediction with precision parameters can be made for unknown input items, wherein the precision parameters mainly refer to estimated variance model parameters and are expressed by the characteristics of obvious parameter reduction, relatively easy parameter optimization, easier parameter convergence and the like. And through the continuous exploration of scholars at home and abroad in nearly 10 years, the Gaussian process is approved in practice, and the application degree in supervised learning is obviously improved.
Drawings
FIG. 1 is a Gaussian process regression prediction.
Detailed Description
The following describes embodiments of the present invention in further detail with reference to the accompanying drawings.
Detailed description of the preferred embodiment
The method for predicting the respiratory motion of the chest and abdomen surface area of the human body in the radiotherapy comprises the following steps:
step a, a color pattern formed by compounding three RGB cosine curves with different frequencies is projected to the surface of the chest and abdomen of a human body by adopting a light pattern projector, a 3CCD color camera is respectively arranged at two sides of the light pattern projector to collect scene images and send the scene images to a computer for post-processing, and three-dimensional coordinates of characteristic mark points and area boundary lines are obtained by adopting two cameras according to a binocular vision principle;
b, aiming at two video sequences obtained by the left camera and the right camera, extracting the same region, the boundary line and the characteristic mark points of the region from the corresponding image pairs, respectively matching the boundary line and the characteristic mark points of the region, then obtaining three-dimensional coordinates of points on the boundary line and the characteristic mark points by adopting the left camera and the right camera according to a binocular vision principle, obtaining three-dimensional coordinates of surface points in the region by adopting a camera and projector combination according to a stripe analysis and phase expansion method, and calculating 7 region characteristic quantities of the three-dimensional coordinates of the characteristic points, the horizontal projection perimeter and the geometric center of the boundary line of the region, the average value and the perimeter of the coordinates of each point of the boundary line of the region, the average value of the coordinates;
step c, determining an interested area, a boundary thereof and a characteristic mark point according to the position of the specific prediction characteristic quantity; taking the predicted characteristic quantity training observation value as a reference, carrying out correlation analysis and significance analysis on the predicted characteristic quantity training observation value and all other regional characteristic quantity training observation values, and optimizing a regional characteristic quantity set which participates in modeling and prediction and recording as Y;
step d, selecting the quasiperiodic kernel function as follows
Figure GDA0002217107960000041
Wherein r | | | x-x' | non-luminous2Representing the Euclidean distance, θ, between two data pointsS、θL、θpIs a hyper-parameter; the sufficient necessary condition for forming the kernel function is that a matrix formed between the test concentrated points and the points must be a semi-positive definite matrix;
step e, to ensure KcFor an effective positive definite covariance function, Cholesky decomposition is used and the elements of the lower triangular matrix are parameterized to obtain Kc＝L(θc)L(θc)TWherein L (θ)c) Is a lower triangular matrix with the size of m multiplied by m; l (theta)c) Is composed of a non-zero element of thetacTo specify the correlation hyperparameter thetacIs given by
F, minimizing the negative logarithmic edge probability with respect to the hyper-parameter, namely-log (y | theta), and then solving the optimal value of the negative logarithmic edge probability by adopting a conjugate gradient method;
step g of giving the hyper-parameter θ by maximizing the cross-correlation between the predicted region feature quantity and the other region feature quantitiesSInitial value, giving the hyper-parameter theta according to the training data of the characteristic quantities of other regionspAssigning an initial value, repeating the experiment for multiple times, and initializing other hyper-parameters randomly;
step h, in the measuring and predicting stage, firstly measuring other region characteristic quantities except the predicted region characteristic quantity according to the sampling frequency, and then aiming at the predicted time x*And predicting the predicted value and the error confidence interval of the predicted value, wherein the t is the prediction time.
Finally, the regression prediction result of the gaussian process shown in the figure I is obtained, wherein the dark line is an original data curve, the light line is a prediction result curve, and the gray area represents the confidence interval of the prediction result curve.

Claims (2)

1. A method for predicting respiratory motion of a chest and abdomen surface area of a human body in radiotherapy is characterized by comprising the following steps:
step a, a color pattern formed by compounding three RGB cosine curves with different frequencies is projected to the surface of the chest and abdomen of a human body by adopting a light pattern projector, a 3CCD color camera is respectively arranged at two sides of the light pattern projector to collect scene images and send the scene images to a computer for post-processing, and three-dimensional coordinates of characteristic mark points and area boundary lines are obtained by adopting two cameras according to a binocular vision principle;
b, aiming at two video sequences obtained by the left camera and the right camera, extracting the same region, the boundary line and the characteristic mark points of the region from the corresponding image pairs, respectively matching the boundary line and the characteristic mark points of the region, then obtaining three-dimensional coordinates of points on the boundary line and the characteristic mark points by adopting the left camera and the right camera according to a binocular vision principle, obtaining three-dimensional coordinates of surface points in the region by adopting a camera and projector combination according to a stripe analysis and phase expansion method, and calculating 7 region characteristic quantities of the three-dimensional coordinates of the characteristic points, the horizontal projection perimeter and the geometric center of the boundary line of the region, the average value and the perimeter of the coordinates of each point of the boundary line of the region, the average value of the coordinates;
step c, determining an interested area, a boundary thereof and a characteristic mark point according to the position of the specific prediction characteristic quantity; taking the predicted characteristic quantity training observation value as a reference, carrying out correlation analysis and significance analysis on the predicted characteristic quantity training observation value and all other regional characteristic quantity training observation values, and optimizing a regional characteristic quantity set which participates in modeling and prediction and recording as Y;
step d, selecting the quasiperiodic kernel function as follows
Figure FDA0002217107950000011
Wherein r | | | x-x' | non-luminous2Representing the Euclidean distance, θ, between two data pointsS、θL、θpIs a hyper-parameter;
step e, to ensure KcFor an effective positive definite covariance function, Cholesky decomposition is used and the elements of the lower triangular matrix are parameterized to obtain Kc＝L(θc)L(θc)TWherein L (θ)c) Is a lower triangular matrix with the size of m multiplied by m; said L (theta)c) Is composed of a non-zero element of thetacTo specify the correlation hyperparameter thetacIs given by
Figure FDA0002217107950000012
F, minimizing the negative logarithmic edge probability with respect to the hyper-parameter, namely-log (y | theta), and then solving the optimal value of the negative logarithmic edge probability by adopting a conjugate gradient method;
step g of giving the mobile hyper-parameter theta by maximizing the cross-correlation between the predicted area feature quantity and the other area feature quantitiesSAssigning initial values to the hyper-parameter theta according to the training data of the characteristic quantities of other regionspAssigning an initial value, repeating the experiment for multiple times, and initializing other hyper-parameters randomly;
step h, in the measuring and predicting stage, firstly measuring other region characteristic quantities of which the predicted region characteristic quantities are unexpected according to the sampling frequency, and then measuring the other region characteristic quantitiesFor predicted time x*And predicting the predicted value and the error confidence interval of the predicted value, wherein the t is the prediction time.
2. The method for predicting respiratory motion of the thoracoabdominal surface area of a human body during radiotherapy according to claim 1, wherein: a sufficient requirement for constructing the kernel function is that the matrix constructed between the test concentration points and the points must be a semi-positive definite matrix.
CN201711359660.9A 2017-12-17 2017-12-17 Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy Expired - Fee Related CN108159576B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711359660.9A CN108159576B (en) 2017-12-17 2017-12-17 Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711359660.9A CN108159576B (en) 2017-12-17 2017-12-17 Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy

Publications (2)

Publication Number Publication Date
CN108159576A CN108159576A (en) 2018-06-15
CN108159576B true CN108159576B (en) 2020-01-07

Family

ID=62522113

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711359660.9A Expired - Fee Related CN108159576B (en) 2017-12-17 2017-12-17 Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy

Country Status (1)

Country Link
CN (1) CN108159576B (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109741831B (en) * 2019-01-09 2022-12-06 哈尔滨理工大学 Human chest and abdomen surface area respiratory signal period prediction method based on variance calculation
CN109741828B (en) * 2019-01-09 2022-11-25 哈尔滨理工大学 Human chest and abdomen surface area respiratory signal period prediction method in radiotherapy
CN110681074B (en) * 2019-10-29 2021-06-15 苏州大学 Tumor respiratory motion prediction method based on bidirectional GRU network
CN113674393B (en) * 2021-07-12 2023-09-26 中国科学院深圳先进技术研究院 Method for constructing respiratory motion model and method for predicting unmarked respiratory motion
CN114177545B (en) * 2022-01-17 2023-11-07 中国科学院合肥物质科学研究院 Contactless respiratory rhythm monitoring device and method for radiotherapy
CN115187608B (en) * 2022-09-14 2023-02-03 苏州大学 Respiration characteristic extraction method based on body surface significance analysis

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103761745A (en) * 2013-07-31 2014-04-30 深圳大学 Estimation method and system for lung motion model
CN104984481A (en) * 2015-05-20 2015-10-21 清华大学 Method and system for carrying out on-line compensation on tumor motion generated by breathing

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103761745A (en) * 2013-07-31 2014-04-30 深圳大学 Estimation method and system for lung motion model
CN104984481A (en) * 2015-05-20 2015-10-21 清华大学 Method and system for carrying out on-line compensation on tumor motion generated by breathing

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
《基于体外信号的呼吸运动跟踪模型的研究》;欧阳斌;《中国优秀硕士学位论文全文数据库 信息科技辑》;20130415;全文 *
《放疗中肿瘤运动基于实时跟踪呼吸预测的算法》;万伟权;《中国优秀硕士学位论文全文数据库 医药卫生科技辑》;20160315;全文 *

Also Published As

Publication number Publication date
CN108159576A (en) 2018-06-15

Similar Documents

Publication Publication Date Title
CN108159576B (en) Human body chest and abdomen surface area respiratory motion prediction method in radiotherapy
CN107403446B (en) Method and system for image registration using intelligent artificial agents
Sminchisescu et al. Generative modeling for continuous non-linearly embedded visual inference
Khan et al. RMS-UNet: Residual multi-scale UNet for liver and lesion segmentation
CN110378881B (en) Tumor positioning system based on deep learning
Rios et al. Population model of bladder motion and deformation based on dominant eigenmodes and mixed-effects models in prostate cancer radiotherapy
Ye et al. Medical image diagnosis of prostate tumor based on PSP-Net+ VGG16 deep learning network
Nakao et al. Image-to-graph convolutional network for deformable shape reconstruction from a single projection image
CN114792326A (en) Surgical navigation point cloud segmentation and registration method based on structured light
Weimin et al. Enhancing Liver Segmentation: A Deep Learning Approach with EAS Feature Extraction and Multi-Scale Fusion
Luo et al. An optimized two-stage cascaded deep neural network for adrenal segmentation on CT images
Jiao et al. TransDose: Transformer-based radiotherapy dose prediction from CT images guided by super-pixel-level GCN classification
Che et al. AMNet: Adaptive multi-level network for deformable registration of 3D brain MR images
Afshar et al. Lung tumor area recognition in CT images based on Gustafson-Kessel clustering
Gamal et al. Brain tumor segmentation using 3D U-Net with hyperparameter optimization
Mambo et al. A review on medical image registration techniques
Romaguera et al. Conditional-based Transformer network with learnable queries for 4D deformation forecasting and tracking
Chen et al. Joint segmentation and discontinuity-preserving deformable registration: Application to cardiac cine-mr images
Fu et al. Divergence-free fitting-based incompressible deformation quantification of liver
CN107808387A (en) A kind of method for tracking target in medical image sequences
Sadikine et al. Semi-overcomplete convolutional auto-encoder embedding as shape priors for deep vessel segmentation
Nambakhsh et al. Right ventricle segmentation with probability product kernel constraints
Suresh et al. Object tracking based 3D modelling and quantification of abnormal contours in brain MRI DICOM study
Zhang et al. Automatic implant shape design for minimally invasive repair of pectus excavatum using deep learning and shape registration
CN111388882A (en) Method for dynamically adjusting radiation angle and dose according to movement of tumor

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
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20200107

Termination date: 20201217