CN110681074B - Tumor respiratory motion prediction method based on bidirectional GRU network - Google Patents
Tumor respiratory motion prediction method based on bidirectional GRU network Download PDFInfo
- Publication number
- CN110681074B CN110681074B CN201911035046.6A CN201911035046A CN110681074B CN 110681074 B CN110681074 B CN 110681074B CN 201911035046 A CN201911035046 A CN 201911035046A CN 110681074 B CN110681074 B CN 110681074B
- Authority
- CN
- China
- Prior art keywords
- network
- data
- prediction
- tumor
- respiratory motion
- 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
- 206010028980 Neoplasm Diseases 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 42
- 230000000241 respiratory effect Effects 0.000 title claims abstract description 39
- 230000002457 bidirectional effect Effects 0.000 title claims abstract description 19
- 238000009499 grossing Methods 0.000 claims abstract description 13
- 230000029058 respiratory gaseous exchange Effects 0.000 claims abstract description 12
- 238000013528 artificial neural network Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract 3
- 230000008569 process Effects 0.000 claims description 15
- 238000010606 normalization Methods 0.000 claims description 13
- 238000005070 sampling Methods 0.000 claims description 5
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 238000000338 in vitro Methods 0.000 claims description 3
- 239000003550 marker Substances 0.000 claims description 3
- 238000001959 radiotherapy Methods 0.000 description 11
- 230000008859 change Effects 0.000 description 4
- 238000001914 filtration Methods 0.000 description 4
- 230000003247 decreasing effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 208000037841 lung tumor Diseases 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000007639 printing Methods 0.000 description 2
- 230000001902 propagating effect Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 201000005202 lung cancer Diseases 0.000 description 1
- 208000020816 lung neoplasm Diseases 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000005180 public health Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- A61N5/1048—Monitoring, verifying, controlling systems and methods
- A61N5/1064—Monitoring, verifying, controlling systems and methods for adjusting radiation treatment in response to monitoring
- A61N5/1065—Beam adjustment
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Measuring devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb
- A61B5/113—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb occurring during breathing
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/48—Other medical applications
- A61B5/4836—Diagnosis combined with treatment in closed-loop systems or methods
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7271—Specific aspects of physiological measurement analysis
- A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/40—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Pathology (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Heart & Thoracic Surgery (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Artificial Intelligence (AREA)
- Theoretical Computer Science (AREA)
- Physiology (AREA)
- Psychiatry (AREA)
- Dentistry (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Signal Processing (AREA)
- Radiology & Medical Imaging (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Urology & Nephrology (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Radiation-Therapy Devices (AREA)
Abstract
本发明公开了一种基于双向GRU网络的肿瘤呼吸运动预测方法，包括以下步骤：获取肿瘤呼吸运动的历史数据，对所述历史数据进行平滑化处理和归一化处理，并对所述平滑化处理和归一化处理后的数据进行时相划分，获得预处理数据集；从所述预处理数据集中选定用于神经网络训练的训练集；构建双向GRU预测网络，将训练集带入预测网络以更新网络的参数，获得训练后的预测网络；将实时数据输入至训练后的预测网络，获得呼吸运动预测值的输出；定期获取实际的肿瘤呼吸运动数据，将实际的肿瘤呼吸运动数据与预测值进行比对，获得误差值；进行预测网络模型在线更新。其能够预测肿瘤呼吸，具有较高的预测精度和鲁棒性。
The invention discloses a tumor respiratory motion prediction method based on a bidirectional GRU network, comprising the following steps: obtaining historical data of tumor respiratory motion, performing smoothing and normalizing processing on the historical data, The processed and normalized data are divided into phases to obtain a preprocessing data set; a training set for neural network training is selected from the preprocessing data set; a bidirectional GRU prediction network is constructed, and the training set is brought into the prediction The network can update the parameters of the network to obtain the trained prediction network; input real-time data into the trained prediction network to obtain the output of the predicted value of respiratory motion; regularly obtain the actual tumor respiratory motion data, and compare the actual tumor respiratory motion data with the actual tumor respiratory motion data. The predicted values are compared to obtain error values; the prediction network model is updated online. It can predict tumor respiration with high prediction accuracy and robustness.
Description
Technical Field
The invention relates to the technical field of radiotherapy robot breath tracking, in particular to a tumor breath motion prediction method based on a bidirectional GRU network.
Background
The lung cancer is one of the main public health problems seriously threatening the global population health, and among the treatment means of lung tumors, radiotherapy is the treatment means with the highest use rate, and has the advantages of wide applicability, good treatment effect and the like. With the intellectualization and refinement of radiotherapy equipment, the concept of precise radiotherapy is proposed and applied to the treatment of lung tumors. The precise radiotherapy refers to a tumor treatment mode which combines radiotherapy medicine with computer network technology, physics and the like. Real-time image guidance and synchronous breath tracking are adopted in the treatment process, so that the treatment accuracy is ensured.
However, the change of the tumor position caused by the respiration of the human body brings great difficulty to the respiration tracking part of the precise radiotherapy, and the tumor movement completely destroys the effect of the dose distribution suitable for the shape of the three-dimensional static target area. Not only does the actual radiation dose to the target tumor be less than the planned dose reduce the efficiency of radiation therapy, but normal tissue surrounding the tumor enters the high dose region in the center of the field causing complications. Meanwhile, the breathing motion is a complex quasi-periodic motion without obvious regularity, and the breathing mode of the patient can change during the treatment process, such as frequency change, baseline drift and position amplitude offset. Therefore, it is necessary to introduce a prediction means of respiratory motion to compensate the tumor position of the radiotherapy device.
Although the traditional respiratory prediction means such as LMS, RLS, Kalman filtering and other methods are simple in algorithm construction, the traditional respiratory prediction means have the defects of low prediction precision, poor robustness and the like; and the network updating time of the RNN neural network prediction method based on LSTM and the like is too long, so that the online updating of the prediction model is difficult to realize on the basis of periodically acquiring tumor position data through X-rays.
Disclosure of Invention
The invention aims to provide a tumor respiration motion prediction method based on a bidirectional GRU network, which can predict tumor respiration and has higher prediction precision and robustness.
In order to solve the technical problem, the invention provides a tumor respiratory motion prediction method based on a bidirectional GRU network, which comprises the following steps:
acquiring historical data of tumor respiratory motion, performing smoothing and normalization processing on the historical data, and performing time-phase division on the data subjected to smoothing and normalization processing to obtain a preprocessed data set;
selecting a training set for neural network training from the preprocessed data set, and printing an expiration label and an inspiration label on the training set according to time phase division;
step three, constructing a bidirectional GRU prediction network, bringing the training set into the prediction network to update the parameters of the network, and obtaining the trained prediction network;
step four, collecting real-time data, inputting the real-time data into a trained prediction network, and obtaining the output of a respiratory motion prediction value;
acquiring actual tumor respiratory motion data regularly, and comparing the actual tumor respiratory motion data with a predicted value to obtain an error value; and setting an error threshold, and if the error value is higher than the error threshold, performing online updating on the actual tumor respiratory motion data by using the prediction network model.
Preferably, in the first step, the acquired historical data of tumor respiratory motion is in-vitro marker point motion data.
Preferably, in the first step, the "smoothing process" is a moving average method.
Preferably, in the first step, the Max-Min normalization method is adopted as the normalization processing.
Preferably, the "time-phase dividing the data after the smoothing processing and the normalization processing" specifically includes: the findpeak method is adopted to search a segmentation point corresponding to the inhalation and exhalation processes.
Preferably, in the second step, the training set is reconstructed into a three-dimensional tensor of (samples, times, input dim), where the dimensions of samples refer to samples of all input networks, the dimensions of times refer to sampling step lengths, and the input dim refers to input dimensions of data.
Preferably, the third step specifically includes:
s31, setting the number of GRU network layers, the number of units of each layer, training rounds, a training optimizer and the data length of each training by using a Keras library to obtain a prediction network;
and S32, inputting the training set data into the prediction network of S31, and reversely propagating the error back to the network by using a gradient descent method to correct the parameters of the prediction network.
Preferably, in the fifth step, the error threshold is a root mean square error.
Preferably, the root mean square error is 0.5 mm.
The invention has the beneficial effects that:
1. the method uses the bidirectional GRU network to predict the respiratory movement of the tumor, and has higher prediction precision and robustness compared with the traditional prediction methods such as RLS, LMS, Kalman method and the like.
2. Compared with other neural network prediction methods, the method has higher network updating speed, and can realize real-time prediction model updating of the neural network prediction model in radiotherapy; due to the continuity of the treatment process, online updating of the predictive model does not interrupt the treatment process.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a comparison graph of moving average filtering after data smoothing;
FIG. 3 shows the result of the division of the respiratory motion phase of the tumor;
FIG. 4 is a schematic diagram of a GRU network structure;
FIG. 5 is a graph comparing predicted results with actual values;
FIG. 6 is a flow chart of online updating of a predictive model.
Detailed Description
The present invention is further described below in conjunction with the following figures and specific examples so that those skilled in the art may better understand the present invention and practice it, but the examples are not intended to limit the present invention.
Referring to fig. 1, the invention discloses a tumor respiratory motion prediction method based on a bidirectional GRU network, comprising the following steps:
step one, historical data of tumor respiratory motion is obtained, smoothing and normalization processing are carried out on the historical data, and time phase division is carried out on the data after the smoothing and normalization processing, so that a preprocessed data set is obtained.
The acquired historical data of the tumor respiratory motion is motion data of the in-vitro marker points.
And the "smoothing process" employs a moving average method. The moving average filtering is based on statistical rules, and continuous sampling data is regarded as a queue with the length fixed as N, after a new measurement, the head data of the queue is removed, the rest N-1 data are sequentially moved forward, and the new sampling data is inserted to be used as the tail of the new queue; then, arithmetic operation is carried out on the queue, and the result is used as the result of the measurement. Assuming that the input is x and the output is y, the calculation formula of the moving average filter is as follows:
fig. 2 is a comparison of the smoothed data with the original data.
The 'normalization processing' adopts a Max-Min normalization method. The principle is to perform linear transformation on original data, minA and maxA are respectively set as the minimum value and the maximum value of an attribute A, and an original value x of A is mapped to a value y of an interval [0,1] through maximum-minimum standardization, so that the formula is as follows:
wherein, the time phase division adopts a findpeak method to search a division point corresponding to the inspiration and expiration processes. The principle is as follows: after filtering and smoothing, firstly traversing the array, and finding that the (N-N, N) is increased progressively and the (N, N + N) is decreased progressively; or (N-N, N) is decreasing and (N, N + N) is increasing, it can be considered as the cut-off point of breathing. The result of the breathing phase partition is shown in fig. 3.
And secondly, selecting a training set for neural network training from the preprocessed data set, and printing an expiration label and an inspiration label on the training set according to the time phase division.
Specifically, 10% of the data of the whole treatment process is selected as a training set to be input into the network. And reconstructing the data into a three-dimensional tensor in the shape of (samples, timestamps, input dim), wherein the dimensions of the samples refer to all samples of the input network, the dimensions of the timestamps refer to sampling step length, and the input dim refers to the input dimensions of the data. And marking a corresponding label for the training set according to the result of the time phase division in the previous step, wherein the label of the expiration process is 1, and the label of the inspiration process is 0.
And step three, constructing a bidirectional GRU prediction network, and bringing the training set into the prediction network to update the parameters of the network to obtain the trained prediction network.
S31, setting the number of GRU network layers, the number of units of each layer, training rounds, a training optimizer and the data length of each training by using a Keras library to obtain a prediction network;
and S32, inputting the training set data into the prediction network of S31, reversely propagating the error back to the network by using a gradient descent method, and correcting the parameters of the prediction network until the output error of the network is minimum.
And step four, acquiring real-time data, inputting the real-time data into the trained prediction network, and obtaining the output of the respiratory motion prediction value.
And selecting a section of motion data stream from the current time to the previous k times, reconstructing the motion data stream into a data structure like a training set, and inputting the data structure into the prediction network. The prediction network outputs a predicted value of the motion position at the next moment. The predicted result versus real data ratio is shown in fig. 5.
Acquiring actual tumor respiratory motion data through X-rays regularly, and comparing the actual tumor respiratory motion data with a predicted value to obtain an error value; and setting an error threshold, and if the error value is higher than the error threshold, performing online updating on the actual tumor respiratory motion data by using the prediction network model. Or the continuous prediction error tends to increase, and the acquired real tumor respiratory motion data is used for performing online updating of the prediction model. Due to the continuity of the treatment process, online updating of the prediction network model does not interrupt the treatment process. The error threshold is a root mean square error. The root mean square error is 0.5 mm. The tumor respiratory motion prediction and radiotherapy process does not stop while the model is updated. After the updating of the prediction model is finished, the new prediction model is used for prediction output, and the step of comparing the acquired real motion data is repeated, so that circulation is realized. FIG. 5 is a flow chart of online updating of a predictive model.
The above-mentioned embodiments are merely preferred embodiments for fully illustrating the present invention, and the scope of the present invention is not limited thereto. The equivalent substitution or change made by the technical personnel in the technical field on the basis of the invention is all within the protection scope of the invention. The protection scope of the invention is subject to the claims.
Claims (7)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911035046.6A CN110681074B (en) | 2019-10-29 | 2019-10-29 | Tumor respiratory motion prediction method based on bidirectional GRU network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911035046.6A CN110681074B (en) | 2019-10-29 | 2019-10-29 | Tumor respiratory motion prediction method based on bidirectional GRU network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110681074A CN110681074A (en) | 2020-01-14 |
| CN110681074B true CN110681074B (en) | 2021-06-15 |
Family
ID=69114364
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911035046.6A Active CN110681074B (en) | 2019-10-29 | 2019-10-29 | Tumor respiratory motion prediction method based on bidirectional GRU network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110681074B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115317806A (en) * | 2022-07-01 | 2022-11-11 | 南方医科大学 | Dynamic tumor respiratory motion prediction method, readable storage medium and computer equipment |
| CN116051603B (en) * | 2023-01-30 | 2025-07-04 | 北京理工大学 | An intelligent breathing prediction method for spinal surgery |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101512636A (en) * | 2006-09-11 | 2009-08-19 | 惠普开发有限公司 | Computational music-tempo estimation |
| WO2016156236A1 (en) * | 2015-03-31 | 2016-10-06 | Sony Corporation | Method and electronic device |
| CN106570335A (en) * | 2016-11-10 | 2017-04-19 | 苏州大学 | Colorless transformation based on correlation model between tumor and marker point in stereoscopic radiotherapy |
| CN106563210A (en) * | 2016-11-10 | 2017-04-19 | 苏州大学 | Unscented transformation (UT)-based radiotherapy robot respiratory movement estimation and prediction system and usage method thereof |
| CN108159576A (en) * | 2017-12-17 | 2018-06-15 | 哈尔滨理工大学 | Human body chest and abdomen surface region respiratory movement predicting method in a kind of radiotherapy |
| CN108875916A (en) * | 2018-06-27 | 2018-11-23 | 北京工业大学 | A kind of ad click rate prediction technique based on GRU neural network |
| CN108926338A (en) * | 2018-05-31 | 2018-12-04 | 中南民族大学 | Heart rate prediction technique and device based on deep learning |
| CN109727672A (en) * | 2018-12-28 | 2019-05-07 | 江苏瑞尔医疗科技有限公司 | Patient's thorax and abdomen malignant respiratory movement predicting tracing method |
| CN110176306A (en) * | 2019-05-17 | 2019-08-27 | 上海交通大学 | A kind of soft tissue drift target spot automatic positioning method based on the polynary LSTM network of dynamic |
| CN110245581A (en) * | 2019-05-25 | 2019-09-17 | 天津大学 | A Human Behavior Recognition Method Based on Deep Learning and Range-Doppler Sequence |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7260426B2 (en) * | 2002-11-12 | 2007-08-21 | Accuray Incorporated | Method and apparatus for tracking an internal target region without an implanted fiducial |
| WO2006002396A2 (en) * | 2004-06-24 | 2006-01-05 | Calypso Medical Technologies, Inc. | Radiation therapy of the lungs using leadless markers |
| CN105931262B (en) * | 2016-07-13 | 2017-02-22 | 中国科学院合肥物质科学研究院 | Tumour position prediction apparatus |
-
2019
- 2019-10-29 CN CN201911035046.6A patent/CN110681074B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101512636A (en) * | 2006-09-11 | 2009-08-19 | 惠普开发有限公司 | Computational music-tempo estimation |
| WO2016156236A1 (en) * | 2015-03-31 | 2016-10-06 | Sony Corporation | Method and electronic device |
| CN106570335A (en) * | 2016-11-10 | 2017-04-19 | 苏州大学 | Colorless transformation based on correlation model between tumor and marker point in stereoscopic radiotherapy |
| CN106563210A (en) * | 2016-11-10 | 2017-04-19 | 苏州大学 | Unscented transformation (UT)-based radiotherapy robot respiratory movement estimation and prediction system and usage method thereof |
| CN108159576A (en) * | 2017-12-17 | 2018-06-15 | 哈尔滨理工大学 | Human body chest and abdomen surface region respiratory movement predicting method in a kind of radiotherapy |
| CN108926338A (en) * | 2018-05-31 | 2018-12-04 | 中南民族大学 | Heart rate prediction technique and device based on deep learning |
| CN108875916A (en) * | 2018-06-27 | 2018-11-23 | 北京工业大学 | A kind of ad click rate prediction technique based on GRU neural network |
| CN109727672A (en) * | 2018-12-28 | 2019-05-07 | 江苏瑞尔医疗科技有限公司 | Patient's thorax and abdomen malignant respiratory movement predicting tracing method |
| CN110176306A (en) * | 2019-05-17 | 2019-08-27 | 上海交通大学 | A kind of soft tissue drift target spot automatic positioning method based on the polynary LSTM network of dynamic |
| CN110245581A (en) * | 2019-05-25 | 2019-09-17 | 天津大学 | A Human Behavior Recognition Method Based on Deep Learning and Range-Doppler Sequence |
Non-Patent Citations (2)
| Title |
|---|
| "基于双向门控循环单元的3D人体运动预测";桑海峰 等;《电子与信息学报》;20190915;第41卷(第9期);第2256-2263页 * |
| Julieta Martinez 等."On Human Motion Prediction Using Recurrent Neural Networks".《2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR)》.2017,第4674-4683页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110681074A (en) | 2020-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109727672B (en) | Prediction and tracking method for respiratory movement of patient thoracoabdominal tumor | |
| CN108883297B (en) | Treatment control using motion prediction | |
| Vedam et al. | Predicting respiratory motion for four‐dimensional radiotherapy | |
| CN106563210B (en) | Estimation and prediction method of tumor respiration motion of radiotherapy robot based on UT transform | |
| CN110681074B (en) | Tumor respiratory motion prediction method based on bidirectional GRU network | |
| Lombardo et al. | Offline and online LSTM networks for respiratory motion prediction in MR-guided radiotherapy | |
| CN108159576B (en) | A method for predicting respiratory motion in human thoracic and abdominal surface regions during radiotherapy | |
| WO2020199836A1 (en) | Method and system for dosimetry evaluation on organs at risk in esophageal radiotherapy plan | |
| Houssein et al. | Estimation of respiratory variables from thoracoabdominal breathing distance: a review of different techniques and calibration methods | |
| CN104268914B (en) | Reestablishing method of 4D-CT (Four Dimensional-Computed Tomography) different time phase sequence image | |
| CN113574565A (en) | A system for synthesizing real-time images using optical body surface motion signals | |
| Jafari et al. | In-vivo lung biomechanical modeling for effective tumor motion tracking in external beam radiation therapy | |
| CN115187608B (en) | Respiration characteristic extraction method based on body surface significance analysis | |
| CN103153398A (en) | Method for setting up a radiation planning and method for applying a spatially resolved radiation dose | |
| Borgmann et al. | Lung area estimation using functional tidal electrical impedance variation images and active contouring | |
| CN108010056B (en) | A blood vessel motion tracking method based on four-dimensional medical images | |
| CN113368403B (en) | Intelligent physiotherapy system capable of improving cardio-pulmonary function | |
| Yoganathan et al. | Predicting respiratory motion using a novel patient specific dual deep recurrent neural networks | |
| CN107808387B (en) | Target tracking method in medical image sequence | |
| CN109887573B (en) | Human body chest and abdomen surface motion prediction method facing radiotherapy | |
| Feng et al. | Clinical CFD applications 2 | |
| Xi et al. | Automatic delineation of hippocampus in CT images based on deep learning and dosimetry study in whole brain radiotherapy | |
| CN113841204A (en) | System and method for training a machine learning model and for providing estimated internal images of a patient | |
| Polak et al. | Estimation of lung properties using ANN-based inverse modeling of spirometric data | |
| Harry et al. | Cardiac dosimetric evaluation of deep inspiration breath-hold level variances using computed tomography scans generated from deformable image registration displacement vectors |
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 |