CN107358607A - Tumour radiotherapy visual monitoring and visual servo intelligent control method - Google Patents
Tumour radiotherapy visual monitoring and visual servo intelligent control method Download PDFInfo
- Publication number
- CN107358607A CN107358607A CN201710688862.1A CN201710688862A CN107358607A CN 107358607 A CN107358607 A CN 107358607A CN 201710688862 A CN201710688862 A CN 201710688862A CN 107358607 A CN107358607 A CN 107358607A
- Authority
- CN
- China
- Prior art keywords
- radiotherapy
- visual
- pose
- patient
- therapeutical bed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- 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
- A61N5/103—Treatment planning systems
-
- 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
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
-
- 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
- G06T2207/10072—Tomographic images
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Pathology (AREA)
- Animal Behavior & Ethology (AREA)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Radiation-Therapy Devices (AREA)
Abstract
The present invention is tumour radiotherapy visual monitoring and visual servo intelligent control method:1)The Multimodal medical image of the collections such as stage airborne medical image system and more visual optics images are performed according to optical imaging system, radiotherapy planning in radiotherapy planning step simulations positioning medical image system, radiotherapy position phase treatment room, handled by intellectual analysis, to the tumor radiotherapy target area in radiotherapy overall process and jeopardize organ, auxiliary treatment personnel and object are monitored, identify, positioned in the radiation therapeutical bed for carrying patient and therapeutic equipment and therapeutic room;2)According to tumor radiotherapy overall process different phase gather optical imagery and(Or)Medical image and clinical radiation therapy procedure quality ensure requirement, pass through multistage more visual servo intelligence control systems(Algorithm)Intelligent, automation, high accuracy, high robust Visual servoing control are carried out to optical imaging system in the radiation therapeutical bed in tumor radiotherapy overall process, therapeutic equipment, airborne medical image system, therapeutic room.
Description
Technical field
The present invention relates to computer vision, robot, intelligent control, artificial intelligence, medical image, image procossing, image
The technology such as analysis, machine learning, pattern-recognition and tumour radiotherapy physics, radiation biological, radiotherapy, biomedical engineering is led
Domain;More particularly to the dynamic vision monitoring image of tumour radiotherapy overall process and medical oncology image intellectual analysis processing side
Method;Radiation therapeutical bed(Carry patient), therapeutic equipment(Accelerator, multi-leaf optical grating), airborne imaging system, therapeutic room's visual monitoring
The intellectuality of imaging system, automation, high accuracy, the visual servo intelligent control method of high robust;Specifically, it shows tumour
Radiotherapy visual monitoring and visual servo intelligent control method.
Background technology
Tumour radiotherapy is one of current three big oncotherapy technologies.Current clinically malignant tumour accurate radiotherapy process
Quality assurance rely primarily on clinical radiation therapy doctor, radiotherapy physics teacher and treat the timely and effective manipulation of technician and realize, radiotherapy
Operating efficiency is low, and Patients During Radiotherapy reliability is not high, and radiotherapy quality is difficult to ensure card.
Modern accurate radiotherapy procedure quality ensures to depend on:1)Radiotherapy planning step simulations position imaging system(Including meter
Calculation machine Tomography (CT), Magnetic resonance imaging (MRI), positron emission computerized tomography imaging (PET))Collection is swollen
Knurl target of prophylactic radiotherapy(GTV)With jeopardize organ(OAR)Medical image;2)Optical imaging system in the phase treatment room of radiotherapy position(Bag
Include visible ray, infrared light, structure light, laser radar)The visible ray monitoring image and airborne imaging system of collection(Pencil-beam calculates
Machine Tomography (CBCT))The medical image of collection;3)Radiotherapy planning performs stage airborne imaging system(Ultrasound or MRI
Imaging system)The medical image of collection;And the image gathered according to above different phase, different image-forming principles, to tumor radiotherapy
Overall process carries out dynamic vision monitoring, tracking.Further, intelligence is carried out to the visual optics image and medical image that monitor
Change, the image analysis processing of automation, and image analysing computer result quantitative accordingly, to radiation therapeutical bed, therapeutic equipment(Add
Fast device, multi-leaf optical grating), airborne imaging system, therapeutic room visual monitoring imaging system carry out effective intelligent, automation, high
The visual monitoring servo intelligent control of precision, high robust.
Therefore, it is necessary to provide effective tumour radiotherapy visual monitoring and visual servo intelligent control method.
The content of the invention
It is an object of the invention to provide tumour radiotherapy visual monitoring and visual servo intelligent control method, to solve mesh
The problem of preceding clinically manual control Patients During Radiotherapy efficiency is low, reliability is not high, improve the efficiency and quality of radiotherapy.
The present invention realizes above-mentioned purpose by following implementer's case, comprises the following steps:
Tumour radiotherapy visual monitoring and visual servo intelligent control method, comprise the following steps:
1)Tumour radiotherapy overall process visual monitoring
1-1)By radiotherapy planning step simulations be positioned to as system acquisition patient CT, MRI, PET/CT and(Or)PET/MRI
Medical image, high accuracy determine tumor radiotherapy target area(GTV)With jeopardize organ(OAR)Pose;
1-2)By multiple color depth imaging systems in the phase treatment room of radiotherapy position, laser infrared radar imaging system, multigroup
The visible ray monitoring image that near infrared light binocular vision imaging system, structure light imaging system etc. gather in real time, in real time dynamic with
Track and high accuracy determine radiation therapeutical bed(Carry patient)With the position of auxiliary treatment personnel and object in therapeutic equipment and therapeutic room
Appearance;
1-3)Pass through radiotherapy position phase treatment equipment(Accelerator)CBCT the or CT medical images of airborne imaging system acquisition,
High accuracy determines the target of prophylactic radiotherapy before every interval procedure(GTV)With jeopardize organ(OAR)Current pose;
1-4)The ultrasound or MRI medical image data gathered in real time by radiotherapy planning execution stage airborne imaging system, it is real
When dynamic tracking and high accuracy determine radiotherapy planning implementation procedure in target of prophylactic radiotherapy(GTV)With jeopardize organ(OAR)It is current
Pose
2)Tumour radiotherapy overall process visual servo intelligent control
2-1)According to being put in radiotherapy position stage optical Imaging: Monitoring system real time dynamic tracing and the therapeutic room determined in high precision
Penetrate therapeutic bed(Carry patient), therapeutic equipment, the spatial pose of auxiliary treatment personnel and object, pass through visual servo intelligent control
System(Algorithm), guide and control radiation therapeutical bed(Carry patient)Predetermined radiotherapy district is moved to, and makes the centre bits such as treatment
Put in the certain limits of center such as the radiotherapy planning for reaching the determination of radiotherapy planning stage;
2-2)Pass through automatic comparison therapeutic equipment(Accelerator)Every fractionated radiotherapy pendulum position stage that airborne imaging system image determines
The radiotherapy planning that target of prophylactic radiotherapy determines with jeopardizing current pose and the radiotherapy planning step simulations of organ to be positioned to as system image
Tumor radiotherapy target area and the pose for jeopardizing organ, automatic guiding and control, adjustment radiation therapeutical bed(Carry patient)Pose, make to work as
The centers such as the radiotherapy planning that the centers such as preceding treatment and radiotherapy planning stage determine are controlled in tolerance scope
It is interior, and make Pose Control when target of prophylactic radiotherapy pose and simulator locating plan in the range of clinical radiation therapy tolerance, then,
The radiotherapy planning determined by the radiotherapy planning stage carries out the treatment of this gradation;Otherwise, radiotherapy planning is corrected, and is put by revised
The plan for the treatment of carries out the treatment of this gradation, it is ensured that the quality of therapeutic process is effectively ensured;
2-3)Stage airborne imaging system image real time dynamic tracing and the treatment determined in high precision are performed according to radiotherapy planning
Target area and the current pose for jeopardizing organ in journey, guiding in real time and control radiation therapeutical bed(Carry patient)With the position of therapeutic equipment
Appearance, radiotherapy planning implementation procedure is set to be carried out by the radiotherapy planning determined before treatment;Otherwise, radiotherapy planning is corrected in real time, and by repairing
Radiotherapy planning after just carries out this interval procedure, it is ensured that the quality of therapeutic process is effectively ensured.
Further, the present invention is realized by radiation treatment robot system.Radiation treatment robot system is by tumour radiotherapy
Treatment simulator locating medical image system, radiotherapy position optical imaging system, the multiple degrees of freedom radiation therapeutical bed for carrying patient, control
Treat equipment(Accelerator and multi-leaf optical grating), the composition such as airborne medical image system.
Further, visual monitoring and visual servo intelligence control system(Algorithm)By multistage more visual optics imagings, prison
Measurement equipment(System)Composition, is specifically included:
(3-1)In radiotherapy room centered on the centers such as radiotherapy planning, to 10m*10m*10m(Wherein m:For long measure:Rice)
Interesting target in spatial dimension carries out vision positioning, and a kind of embodiment is as follows:
A certain fixing point establishes global space coordinates as origin using in room.It is big using being fixedly mounted in quality room
Field laser radar and(Or)The various visual angles color depth imaging of multiple fixed installations(RGBD)Equipment(System)Carry out radiotherapy room
Static three-dimensional rebuild, reconstructed results show as three-dimensional point cloud;According to this reconstructed results, indoor static target is determined(Interior Space
Between, radiation therapeutical bed, therapeutic equipment, airborne medical imaging devices etc.)Coarse positioning model and the position in global space coordinate system
Put, the interesting target three dimensional local information of lower accuracy is provided for first order radiotherapy position visual monitoring and visual servo;
(3-2)In radiotherapy room centered on the centers such as radiotherapy planning, the interesting target in 5m*5m*5m spatial dimensions is entered
Row degree of precision vision positioning, a kind of embodiment are as follows:
Using the near-infrared multiple binocular stereovision equipment that multigroup radiotherapy apparatus is airborne(System), gather the radiation in radiation treatment region
Therapeutic bed and entire patient's realtime graphic.It is based on(3-1)Coarse positioning model, determine radiation therapeutical bed static three-dimensional model, build
Vertical radiation therapeutical bed local coordinate system;Using technique of binocular stereoscopic vision, moving target interested is monitored in real time(Radiotherapy
Bed, patient), entire patient's three-dimensional space model and the position under radiation therapeutical bed local coordinate system and posture are determined, is second
Level radiotherapy position visual monitoring and visual servo provide the interesting target three-dimensional space position information of degree of precision;
(3-3)In radiotherapy room centered on the centers such as radiotherapy planning, to interested in 0.5m*0.5m*0.5m spatial dimensions
Target carries out high-precision vision positioning, and a kind of embodiment is as follows:
It is based on(3-2)Entire patient's three-dimensional space model of determination, establish patient's body coordinate system.Using active structural light three-dimensional
Scanning device(System), to patient's local body surface marker progress Real-time modeling set, and and(3-2)The entire patient of middle determination
Threedimensional model carry out real-time matching, determine real time position of patient's local volume surface marker position in patient's body coordinate system and
Posture, high-precision interesting target three-dimensional space position letter is provided for third level radiotherapy position visual monitoring and visual servo
Breath.
Further, it is as follows using the multistage pendulum position visual monitoring of tumour radiotherapy and visual servo, a kind of embodiment:
4-1)The foundation and demarcation of each coordinate system
This method is related to global coordinate system, radiation therapeutical bed local coordinate system, three kinds of space coordinates of patient's body coordinate system, is
Accurate three-dimensional position is obtained, it is necessary to be demarcated to coordinate systems at different levels, obtain the transformational relation between coordinate system.It is at different levels
The foundation rule of coordinate system is as follows:
(4-1-1)Global coordinate system
Using a certain fixing point in radiotherapy room as origin, Descartes's three-dimensional system of coordinate is established with right-handed coordinate system.It is by radiotherapy room
It is interior
Radiation therapeutical bed, therapeutic equipment, airborne medical imaging devices and optical monitoring imaging device of patient etc. is carried uniformly to arrive
World coordinate system together
(4-1-2)Radiation therapeutical bed local coordinate system
Using a certain fixing point in radiation therapeutical bed as origin, the various kinds of equipment on radiation therapeutical bed is often corrected to this during task
Coordinate system, the equipment that optical imagery monitoring is used in therapeutic room are also demarcated under the coordinate system;
(4-1-3)Patient's body coordinate system
Using patient tumors radiotherapy planning isocenter point as origin, three-dimensional system of coordinate is established, X-axis represents left and right directions, and Y-axis represents head
Pin direction, Z axis represent abdomen back of the body direction, and horizontal plane is rotated in a clockwise direction vector value as just.
Wherein, the demarcation between global coordinate system and radiation therapeutical bed local coordinate system can enter rower using cooperation mark
It is fixed, some cooperation marks are set on space and radiation therapeutical bed indoors respectively, gathers 3 pairs using total powerstation and the above is not conllinear
The coordinate in two spaces coordinate system of index point, calculate rigid body transformational relation of the radiation therapeutical bed in global coordinate system.
Matched by threedimensional model, calculate the transformation relation between patient's body coordinate and radiation therapeutical bed local coordinate, together
When can have a local displacement due to patient, therefore the conversion needs to carry out real time correction, is set by Near-infrared Double item stereo vision
The standby position and attitude for obtaining patient in real time is simultaneously transformed under radiation therapeutical bed coordinate system.
Further, using following multistage radiotherapy position visual monitoring and visual servo intelligence control system(Algorithm):
4-2)First order radiotherapy position
The real-time pose of first order radiotherapy position by the scanning of the big field laser that is arranged in therapeutic room and(Or)Various visual angles are color
Color depth is imaged(RGBD)Equipment(System)Three-dimensional modeling determination is carried out to indoor static interesting target.Wherein, radiotherapy
The position of bed is obtained by demarcation, and the position of other static objects can be obtained by the matching of threedimensional model.Calculate patient's
Real-time pose simultaneously feeds back to Visual servoing control system, and radiation therapeutical bed is moved in the central areas such as Patients Treated by Radiotherapy plan;
4-3)Second level radiotherapy position
The real-time pose of second level radiotherapy position is determined by near-infrared multiple binocular stereovision imaging system.By to patient and
Radiation therapeutical bed region carries out real-time three-dimensional reconstruction, and is matched with static radiation therapeutic bed model, calculates the reality of patient
Shi Weizi, and Visual servoing control system is fed back to, radiation therapeutical bed is moved to the near central regions such as Patients Treated by Radiotherapy plan;
4-4)Third level radiotherapy position
Third level radiotherapy position pose passes through active binocular structure light scan equipment(System)It is determined that.Pass through real time scan patient
The central areas such as radiotherapy planning, and matched with entire patient's threedimensional model, calculate reality of the patient on radiation therapeutical bed
Shi Weizi, and result is fed back into Visual servoing control system, radiation therapeutical bed is moved to the central areas such as radiotherapy planning.
Further, step 1)With step 2)Using multistage more Vision imaging systems come the radiation therapeutical bed to carrying patient
Carry out dynamic vision monitoring and visual servo intelligently pendulum level controlling system(Algorithm), a kind of embodiment is as follows:
5-1)First order radiotherapy position
Various visual angles color depth imaging device is installed on the wall of therapeutic room's surrounding, for 10m*10m*10m therapeutic rooms physics
Thick pendulum position in spatial dimension.In a 10m*10m*10m therapeutic room, more set various visual angles color depth imagings can be installed and set
It is standby, to the radiation therapeutical bed in the therapeutic room(Carry patient), therapeutic equipment, imaging device, auxiliary treatment personnel and object enter
Row visual information gathers.Using digital picture or video intelligent analyzing and processing algorithm, calculate determine radiation therapeutical bed in real time(Carry
Patient), therapeutic equipment, imaging device, the current pose of auxiliary treatment personnel and object.Using visual servo intelligent control algorithm
To radiation therapeutical bed(Carry patient)It is controlled, control radiation therapeutical bed is moved to the position of radiotherapy planning determination by optimal path
Appearance, the positioning precision of patient reaches 0.5 meter when making radiotherapy position;
5-2)Second level radiotherapy position
The field range of the Near-infrared Double item stereo vision equipment used is 5m*5m*5m.Colored deep by various visual angles various visual angles
Spend imaging device monitoring with behind servo intelligent control one-level pendulum position, by covering near infrared binocular stereo visual system, controlling this more
Treat indoor radiation therapeutical bed(Carry patient), therapeutic equipment, imaging device, auxiliary treatment personnel and object carry out visual information
Collection.Using digital picture or video intelligent analyzing and processing algorithm, calculate determine radiation therapeutical bed in real time(Carry patient), treatment
Equipment, imaging device, the current pose of auxiliary treatment personnel and object.Using visual servo intelligent control algorithm to radiotherapy
Bed(Carry patient)It is controlled, control radiation therapeutical bed is moved to the pose of radiotherapy planning determination by optimal path, makes radiotherapy
The positioning precision of patient reaches 0.5 millimeter when putting position;
5-3)Third level radiotherapy position
The field range of the active structural light three-dimensional scanning system-system used can be carried out more smart for 0.5m*0.5m*0.5m
True pendulum position.By active structural light three-dimensional scanning system, to the radiation therapeutical bed in the therapeutic room(Carry patient), treatment
Equipment, imaging device, auxiliary treatment personnel and object carry out visual information collection.At digital picture or video intelligent analysis
Adjustment method, calculate determine radiation therapeutical bed in real time(Carry patient), therapeutic equipment, imaging device, auxiliary treatment personnel and object
Current pose.Using visual servo intelligent control algorithm to radiation therapeutical bed(Carry patient)It is controlled, control radiation is controlled
The pose that bed is moved to radiotherapy planning determination by optimal path is treated, the positioning precision of patient reaches 0.1 millimeter when making radiotherapy position;
5-4)Fourth stage radiotherapy position
After by third level radiotherapy position, before starting treatment, shadow is carried by airborne imaging Visual servoing control system control machine
CBCT or CT scan before being treated as system to patient, and using medical image intellectual analysis Processing Algorithm to corresponding
CBCT or CT images are analyzed and processed, and are determined to treat isocentric current location in real time, will be treated isocentric current location
Contrasted with centers such as the radiotherapy plannings that is determined before treatment, if error not in permissible range, using visual servo
Intelligent control algorithm is to radiation therapeutical bed and therapeutic equipment(Linear accelerator, multi-leaf optical grating)Motion control is carried out, is ensured current
The centers such as the radiotherapy planning determined before the centers such as treatment and treatment are controlled in tolerance scope(2 millimeters)It is interior, then put
Penetrate treatment.
Further, all the time can not be by the centers such as current treatment and radiotherapy planning rank if in the radiotherapy position stage
The control errors of the position of center line such as the radiotherapy planning that section determines then correct radiotherapy planning, again within clinical allowed band
Establish new radiotherapy planning.It is measured with airborne imagery system again afterwards it is real-time wait center contrast, error is fed back to and regarded
Feel servo intelligent control system, and by visual servo intelligent control algorithm, to radiation therapeutical bed and therapeutic equipment(Straight line adds
Fast device, multi-leaf optical grating)Carry out visual servo motion control, it is ensured that per fractionated radiotherapy plan execution phase the center such as treatment and
The centers such as the radiotherapy planning that the radiotherapy planning stage determines are controlled in the range of clinical tolerance.
Further, the radiotherapy planning stage, can be according only to a certain tumor radiotherapy simulator locating imaging image data(Such as
CT)Determine tumor radiotherapy target area(GTV)With jeopardize organ(OAR)Pose, two kinds can also be used in combination, or two or more
Tumor radiotherapy simulator locating image determines tumor radiotherapy target area(GTV)With jeopardize organ(OAR)Pose, such as combination tumor CT/
MRI;Or CT/CBCT;Or PET/CT;Or PET/MRI;Or PET/CT/MRI.
In the radiotherapy position stage, a kind of monitoring optical imaging system of dynamic vision in real time can be used only(Such as binocular vision system
System)Real-time dynamic monitoring, tracking and high accuracy positioning radiation therapeutical bed(Carry patient)It is auxiliary in pose and therapeutic equipment, therapeutic room
The pose for the treatment of personnel and object is helped, two kinds, or two or more imaging systems can also be used in combination(Such as more binocular vision systems
System, structured-light system, laser radar system) real time dynamic tracing and high accuracy positioning radiation therapeutical bed(Carry patient)Pose and
The pose of auxiliary treatment personnel and object in therapeutic equipment, therapeutic room.
Per fractionated radiotherapy during pendulum position, airborne CBCT systems, or CT system can be used to determine in high precision per interval procedure
Pre-neoplastic target of prophylactic radiotherapy(GTV)With jeopardize organ(OAR)Current pose;Ultrasound or MRI systems can be used in therapeutic process
Real time dynamic tracing and the high accuracy of uniting determine tumor radiotherapy target area in therapeutic process(GTV)With jeopardize organ(OAR)Pose
Current pose;The different times of therapeutic process, use different visual monitoring image intelligent analysis and processing methods and visual servo
Intelligent control method carries out guiding in real time and control radiation therapeutical bed(Carry patient)And therapeutic equipment, make therapeutic process by radiotherapy
Plan is carried out, it is ensured that the quality of Patients During Radiotherapy is effectively controlled.
Further, intelligently controlled pendulum position with visual servo by the visual monitoring of tumour radiotherapy radiotherapy position process, and
After completing third level radiotherapy position, start airborne image system(CBCT)Intelligent control, and the patient on radiation therapeutical bed is entered
Row CBCT is scanned, to determine the centers such as the current treatment of patient and tumor radiotherapy target area pose;Specifically comprise the following steps:
6-1)Using near-infrared multiple binocular stereovision system, airborne image system is monitored(CBCT), radiation therapeutical bed(Carry and suffer from
Person), therapeutic equipment(Linear accelerator, multi-leaf optical grating), auxiliary treatment personnel and object current pose, and by airborne imagery system
System(CBCT)Current pose and airborne imagery system(CBCT)Expected pose contrasted, caused error propagation is to airborne
CBCT system Visual servoing control devices.The Visual servoing control device exports different FREQUENCY CONTROLs according to different control information to be believed
Number, control airborne image system(CBCT)Motor driver.Different frequency control signal controlled motor drivers, produce not
Same voltage acts on airborne imagery system(CBCT)Servomotor on so that the driving servomotor of airborne imagery system
Rotating speed is changed.Torque caused by the gear-box amplification servomotor being connected with servomotor, and act on airborne imagery system
System(CBCT)On, make its motion, until airborne imagery system(CBCT)Current pose and the error of expected pose allowing model
In enclosing.Then, send DC power supply switch to servomotor brake by Visual servoing control device to instruct, and pass through servo electricity
Torque Control servomotor stop motion on machine, lock airborne imagery system(CBCT)Pose;
6-2) in locking airborne imagery system(CBCT)After pose, by the expected pose of CBCT Image Plates and the shadow that in real time monitors
As plate real-time pose is contrasted, caused error is transmitted to CBCT Image Plate servo controllers, servo controller output CBCT
Image version motor driver frequency control signal, controlled motor driver export different voltage, act on CBCT Image Plates and watch
Take on motor, so as to controlled motor rotating speed, after gear-box amplifies, produce certain moment loading on Image Plate, make it
Move to imaging pose;
6-3)In locking airborne imagery system(CBCT)After pose, by the expected pose of CBCT X-ray machine and the X-ray machine of monitoring
Real-time pose is contrasted, and caused error is transmitted to servo controller, servo controller output X-ray machine motor driver frequency
Control signal, controlled motor driver exports different voltage, acted on X-ray machine servomotor, so that controlled motor rotating speed,
After the amplification of gear-box, certain moment loading is produced on X-ray machine, causes it to move to imaging pose.
Further, the fourth stage radiotherapy position stage, simulator locating CT shadows when airborne CBCT images and radiotherapy planning are passed through
High registration accuracy as between, determine target of prophylactic radiotherapy and jeopardize the Deformation Field of organ.Further, will by the Deformation Field of determination
The target of prophylactic radiotherapy and jeopardize organ information, radiotherapy planning information auto-changing that simulator locating CT is delineated during based on radiotherapy planning, with
Keep consistent with current CBCT images.And by SERVO CONTROL, radiation therapeutical bed, linear accelerator and multi-leaf optical grating are adjusted in real time
Pose, ensure that simulation is fixed during radiotherapy planning before jeopardizing the current pose of organ and target of prophylactic radiotherapy during per fractionated radiotherapy and treating
Jeopardize organ and target area position and attitude error in allowed band in the CT images of position, when being treated with reducing the center such as target of prophylactic radiotherapy with
The error of the centers such as the radiotherapy planning determined before treatment.Comprise the following steps that:
7-1)Import fractionated radiotherapy process CBCT images and original scheme simulator locating CT images;
7-2)Fractionated radiotherapy process CBCT images carry out high registration accuracy with image in original scheme CT;
7-3)According to registering Deformation Field, automatic conversion target of prophylactic radiotherapy and jeopardize organ and sketch outline information;
7-4)Before parameter, fractionated radiotherapy in original scheme CT CBCT images and this sketch outline recalculate dosage by several times
Distribution and dose volume histogram, constrained with reference to original prescription dosage and radiotherapy clinical dosage, judge whether to need to change radiotherapy
Plan.Plan if desired for modification, then carry out 7-5) and 7-6);Plan need not be such as changed, then using former radiotherapy planning as currently
Fractionated radiotherapy plan, and implement current fractionated radiotherapy;
7-5)The quick on-line amending of patient's radiotherapy planning is carried out with reference to clinical requirement, automatically generates current fractionated radiotherapy plan;
7-6)According to amended radiotherapy planning, radiation therapeutical bed, linear accelerator and multi-leaf optical grating are adjusted by SERVO CONTROL
Implement current Patients During Radiotherapy after pose again.
Further, a kind of image high registration accuracy algorithm is as follows:
8-1)The CBCT images of current fractionated radiotherapy and plan CT images are first subjected to Rigid Registration, that is, translates and rotates;It
Carry out non-rigid registration again afterwards.Present example is matched somebody with somebody using affine with the two-way consistent, high precision elastic based on Demons algorithms
Quasi- algorithm;
8-2)Image registration algorithm is multiple dimensioned, multiresolution, two-way consistent iteration registration process, by constantly to deformation field
Optimize until algorithmic statement, realize tumor imaging high precision elastic registration.The algorithm is converted using topography's information
Image, in order to ensure that the conversion is continuous in global scope, so as to keep the topological structure of image, after iteration each time, use
Skew obtained by Gaussian filter is smooth, improves computational efficiency.Further, to improve algorithm the convergence speed, by driving force
Bidirectional force from two images gradient is changed into from the unidirectional force that floating image pixel spreads to reference picture pixel, i.e., on image
The power of arbitrfary point not only drives floating image to be spread to reference picture, while also drives reference picture to be spread to floating image.
Further, to ensure that registration result is two-way consistent, present example uses two-way consistent registration Algorithm.It is double
Differomorphism demons, reference picture are based on to consistent registration AlgorithmI 1And floating imageI 0Deformation Field there is symmetrical characteristic,
Similarity measure criterion is ensured in differomorphism transformation space.The deformable registration makes reference picture and floating image shape simultaneously
Become, select anyway registration similarity measure and Optimal Parameters, the symmetric derivative homeomorphism ensure byI 1ArriveI 0Conversion with byI 0ArriveI 1Conversion there is uniformity.The two-way uniformity registration ensures the symmetry of two images transformed distances estimation, does not limit
How still image and reference picture selected.By directly adding reversible limitation during optimization, this method ensures sub- body
The reversible deformation of the registration accuracy and discrete domain of element.Meanwhile symmetrical deformation makes registration error minimum using invertibity.
Further, using Diagonal Recurrent Neural Network(DRNN)Visual servoing control algorithm is to radiation therapeutical bed, treatment
Equipment etc. carries out multistage visual servo intelligent control, and a kind of embodiment is as follows:
9-1)Diagonal Recurrent Neural Network structure
Diagonal Recurrent Neural Network(DRNN)With three layers:Input layer, hidden layer and output layer.Diagonal Recurrent Neural Network implies
Between the neuron of layer or network is output between input there is feedback link, and its feedback element also has time delayses special
Property, network has short-term memory capability.A kind of embodiment of the present invention is the Patients During Radiotherapy to complexity using DRNN networks
It is modeled, recognizes and control;
The expression formula of DRNN network states equation and output equation is:
(1)
(2)
(3)
In formula,It is the of networkIndividual input vector;It is in hidden layerThe summation of the input of individual neuron
Vector;It is hidden layerThe state vector of individual neuron;It is the output vector of network;It is Sigmoid functions.
9-2)Diagonal Recurrent Neural Network identification structure
When being recognized using Diagonal Recurrent Neural Network to system,kFor the iterative steps of network,u(k) it is the defeated of identification objects
Incoming vector,y d (k) be identification objects output vector,y(k) be DRNN networks output vector.The object that needs are recognized
The difference of the input of input and output as DRNN networks, the output of identification objects and the output of DRNN networks is believed as DRNN adjustment
Number.And the parameter of identification network is adjusted according to the size of mean square error;
9-3)Diagonal Recurrent Neural Network learning algorithm
Present example carries out network using dynamic error back-propagation algorithm (Dynamic Back Propagation, DBP)
Training study;
9-4)The training flow of diagonal recurrent neural network
9-4-1):Initialization network parameter;
9-4-2):Forward-propagating, according to the expression formula of DRNN network states equation and output equation calculate the output of network to
Amount, the input vector of hidden layerAnd state vector;
9-4-3):The Identification Errors of calculating network;
9-4-4):Backpropagation, the connection weight between each layer of network is adjusted;
9-4-5):JudgeWhether performance requirement is reached;
9-4-6):If reaching requirement, weights are preserved, terminate training;If not up to, go to 9-4-2);
9-5)Radiation therapeutical bed Visual servoing control system based on DRNN
Desired image characteristic point is compared with the image characteristic point obtained by optical imagery detecting system, by error propagation
Give Visual servoing control device, the speed and Jacobean matrix of the radiation therapeutical bed mechanical arm tail end of output in three-dimensional Cartesian space
Battle array inverse matrix is multiplied, and obtains the speed in each joint of radiation therapeutical bed mechanical arm, and passes to each joint control of radiation therapeutical bed
Device, Diagonal Recurrent Neural Network and its corresponding, being inaccurately partly modified and distinguishing to radiation therapeutical bed kinetic model
Know, and pass the result to radiation therapeutical bed kinetic model, so as to control the pose in each joint of radiation therapeutical bed mechanical arm.
Compared with prior art, it is low, reliable effectively to solve at present clinically manual control Patients During Radiotherapy efficiency by the present invention
Property it is not high the problem of, improve radiotherapy implementation procedure operating efficiency and work quality, it is ensured that the quality of therapeutic process obtains
It is effectively ensured.
Brief description of the drawings
Fig. 1 is by tumour radiotherapy simulator locating medical image system, pendulum position optical imaging system, carries the more of patient
Free degree radiation therapeutical bed, therapeutic equipment(Accelerator and multi-leaf optical grating), the composition such as airborne medical image system radiotherapy
Robot system forms block diagram;Fig. 2 is tumour radiotherapy radiotherapy position multistage vision guide and monitoring optical imaging system totality structure
Into block diagram;Fig. 3 be tumour radiotherapy radiotherapy visual monitoring and Visual servoing control system in radiation therapeutical bed local coordinate system with
Patient's body coordinate system block diagram;Fig. 4 is that the tumour radiotherapy first order puts position visual monitoring and visual servo intelligence pendulum position control is former
Manage block diagram;Fig. 5 is that position visual monitoring and visual servo intelligence pendulum position control principle block diagram are put in the tumour radiotherapy second level;Fig. 6
Position visual monitoring, which is put, for the tumour radiotherapy third level intelligently puts position control principle block diagram with visual servo;Fig. 7 is airborne CBCT
System visual servo motion control theory diagram;Fig. 8 is airborne CBCT Image Plates servo control theory diagram;Fig. 9 is machine
Carry CBCT X-ray machine servo control theory diagram;Figure 10 is that the tumour radiotherapy radiotherapy fourth stage puts position visual monitoring and vision
Servo intelligent puts position control principle block diagram;Figure 11 is Diagonal Recurrent Neural Network(DRNN)The structure chart of identification system;Figure 12 is
The training figure of Diagonal Recurrent Neural Network DRNN connection weights;Figure 13 be based on registration target of prophylactic radiotherapy and crisis organ delineate with
On-line planning adjusts flow;Figure 14 is the high precision elastic registration flow based on demons algorithms;Figure 15 is two-way consistent registration
Algorithm schematic diagram;Figure 16 is radiotherapy visual monitoring and visual servo intelligent control theory diagram.
Embodiment
The present embodiment shows tumour radiotherapy visual monitoring and visual servo intelligent control method, comprises the following steps:
1)Tumour radiotherapy overall process visual monitoring
1-1)By radiotherapy planning step simulations be positioned to as system acquisition patient CT, MRI, PET/CT and(Or)PET/MRI
Medical image, high accuracy determine tumor radiotherapy target area(GTV)With jeopardize organ(OAR)Pose;
1-2)By multiple color depth imaging systems in the phase treatment room of radiotherapy position, laser infrared radar imaging system, multigroup
The visible ray monitoring image that near infrared light binocular vision imaging system, structure light imaging system etc. gather in real time, in real time dynamic with
Track and high accuracy determine radiation therapeutical bed(Carry patient)Pose and therapeutic equipment, therapeutic room in auxiliary treatment personnel and object
Pose;
1-3)Pass through radiotherapy position phase treatment equipment(Accelerator)CBCT the or CT medical images of airborne imaging system acquisition,
High accuracy determines the target of prophylactic radiotherapy before every interval procedure(GTV)With jeopardize organ(OAR)Current pose;
1-4)The ultrasound or MRI medical image data gathered in real time by radiotherapy planning execution stage airborne imaging system, it is real
When dynamic tracking and high accuracy determine radiotherapy planning implementation procedure in target of prophylactic radiotherapy(GTV)With jeopardize organ(OAR)It is current
Pose;
2)Tumour radiotherapy overall process visual servo intelligent control:
2-1)According to being put in radiotherapy position stage optical Imaging: Monitoring system real time dynamic tracing and the therapeutic room determined in high precision
Penetrate therapeutic bed(Carry patient), therapeutic equipment, the spatial pose of auxiliary treatment personnel and object, pass through visual servo intelligent control
System(Algorithm), guide and control radiation therapeutical bed(Carry patient)Predetermined radiotherapy district is moved to, and makes the centre bits such as treatment
Put in the certain limits of center such as the radiotherapy planning for reaching the determination of radiotherapy planning stage;
2-2)Pass through automatic comparison therapeutic equipment(Accelerator)Every fractionated radiotherapy pendulum position stage that airborne imaging system image determines
The radiotherapy planning that target of prophylactic radiotherapy determines with jeopardizing current pose and the radiotherapy planning step simulations of organ to be positioned to as system image
Tumor radiotherapy target area and the pose for jeopardizing organ, automatic guiding and control, adjustment radiation therapeutical bed(Carry patient)Pose, make to work as
The centers such as the radiotherapy planning that the centers such as preceding treatment and radiotherapy planning stage determine are controlled in tolerance scope
It is interior, and make Pose Control when target of prophylactic radiotherapy pose and simulator locating plan in the range of clinical radiation therapy tolerance, then,
The radiotherapy planning determined by the radiotherapy planning stage carries out the treatment of this gradation;Otherwise, radiotherapy planning is corrected, and is put by revised
The plan for the treatment of carries out the treatment of this gradation, it is ensured that the quality of therapeutic process is effectively ensured;
2-3)Stage airborne imaging system image real time dynamic tracing and the treatment determined in high precision are performed according to radiotherapy planning
Target area and the current pose for jeopardizing organ in journey, guiding in real time and control radiation therapeutical bed(Carry patient)With the position of therapeutic equipment
Appearance, radiotherapy planning implementation procedure is set to be carried out by the radiotherapy planning determined before treatment;Otherwise, radiotherapy planning is corrected in real time, and by repairing
Radiotherapy planning after just carries out this interval procedure, it is ensured that the quality of therapeutic process is effectively ensured.
Further, step 1)Realized by radiation treatment robot system, it is specific as shown in Figure 1:
The bed surface of radiation therapeutical bed can on rectangular coordinate system in space X, Y, translated in Z-direction, and around space right-angle
The rotation of X, Y, Z-direction on coordinate system.Multiple degrees of freedom radiation therapeutical bed disclosure satisfy that target area on patient's body different directions
Radiotherapy, the effective quality for ensureing radiation therapy process;
The Patients Treated by Radiotherapy simulator locating that tumour radiotherapy simulator locating medical image system is used for the radiotherapy planning stage is imaged, and is
Patients Treated by Radiotherapy plan provides target of prophylactic radiotherapy and jeopardizes organ OAR simulator locating Multimodal medical images, and passes through multi-modal medical science
The processing of image intellectual analysis determines(Delineate)Patients Treated by Radiotherapy target area and OAR.Pendulum position optical imaging system is used for the radiotherapy position stage
Real time imagery, multiple degrees of freedom radiation therapeutical bed, the therapeutic equipment of patient are carried in monitoring treatment room(Accelerator and grating), it is airborne
The pose of medical image system, auxiliary treatment personnel and object, and handled by multi-modal video image intelligent analysis and determine it
Current state, radiotherapy position visual servo intelligence control system is fed back to, so as to control radiation therapeutical bed and therapeutic equipment(Accelerate
Device and grating)Motion, realize multistage vision guide high accuracy radiotherapy bit function.Carry the multiple degrees of freedom radiotherapy of patient
Bed can carry out multifreedom motion under multistage more visual monitoring guiding and radiotherapy position visual servo intelligent control, it is ensured that
The radiation that patient implements tumour in the range of the errors of centration such as the target of prophylactic radiotherapy that clinical radiation therapy allows and OAR poses, radiotherapy is controlled
Treat;Therapeutic equipment(Accelerator and grating)Can be in multistage more visual monitoring guiding and radiotherapy position visual servo intelligent control
Under, the center such as treatment that can determine radiotherapy planning before the centers such as Current therapeutic and treatment is controlled in tolerance scope
It is interior;Airborne medical image system can monitor the centers such as the current treatment of patient in real time, feed back to radiation therapy process vision and watch
Take intelligence control system so that the center control such as treatment that radiotherapy planning determines before the center such as patient's Current therapeutic and treatment is being held
Perhaps in error range.
Tumour per must all carry out high-precision pendulum position before fractionated radiation, during ensuring every fractionated radiotherapy patient and its
When being positioned by the pose of the tumor target of radiotherapy radiation exposure and the normal structure organ jeopardized and its radiotherapy planning step simulations
Corresponding pose keep high accuracy consistent, the central space physical location error such as tumor radiotherapy(Radiotherapy position precision)Less than 2 millis
Rice.
A kind of monitoring optical imaging system of dynamic vision in real time can be used only in the tumor radiotherapy pendulum position stage(Such as more binocular visions
Feel system)Real time dynamic tracing and high accuracy positioning radiation therapeutical bed(Carry patient)It is auxiliary in pose and therapeutic equipment, therapeutic room
The pose for the treatment of personnel and object is helped, two kinds, or two or more imaging systems can also be used in combination(Such as more binocular vision systems
System, structured-light system, laser radar system) real time dynamic tracing and high accuracy positioning radiation therapeutical bed(Carry patient)Pose and
The pose of auxiliary treatment personnel and object in therapeutic equipment, therapeutic room.
A kind of embodiment is as follows:
4-1)The foundation and demarcation of each coordinate system
This method is related to global coordinate system, radiation therapeutical bed local coordinate system, three kinds of space coordinates of patient's body coordinate system, is
Accurate three-dimensional position is obtained, it is necessary to be demarcated to coordinate systems at different levels, obtain the transformational relation between coordinate system.It is at different levels
The foundation rule of coordinate system is as follows:
(4-1-1)Global coordinate system
Using a certain fixing point in radiotherapy room as origin, Descartes's three-dimensional system of coordinate is established with right-handed coordinate system.It is by radiotherapy room
Interior radiation therapeutical bed, therapeutic equipment, airborne medical imaging devices and the optical monitoring imaging device for carrying patient etc. is unified to one
The world coordinate system risen
(4-1-2)Radiation therapeutical bed local coordinate system
Using a certain fixing point in radiation therapeutical bed as origin, the various kinds of equipment on radiation therapeutical bed is often corrected to this during task
Coordinate system, the equipment that optical imagery monitoring is used in therapeutic room are also demarcated under the coordinate system;
(4-1-3)Patient's body coordinate system
Using patient tumors radiotherapy planning isocenter point as origin, three-dimensional system of coordinate is established, X-axis represents left and right directions, and Y-axis represents head
Pin direction, Z axis represent abdomen back of the body direction, and horizontal plane is rotated in a clockwise direction vector value as just.
Wherein, the demarcation between global coordinate system and radiation therapeutical bed local coordinate system can enter rower using cooperation mark
It is fixed, some cooperation marks are set on space and radiation therapeutical bed indoors respectively, gathers 3 pairs using total powerstation and the above is not conllinear
The coordinate in two spaces coordinate system of index point, calculate rigid body transformational relation of the radiation therapeutical bed in global coordinate system.
Matched by threedimensional model, calculate the transformation relation between patient's body coordinate and radiation therapeutical bed local coordinate, together
When can have a local displacement due to patient, therefore the conversion needs to carry out real time correction, is set by Near-infrared Double item stereo vision
The standby position and attitude for obtaining patient in real time is simultaneously transformed under radiation therapeutical bed coordinate system;
4-2)First order radiotherapy position
The real-time pose of first order radiotherapy position by the scanning of the big field laser that is arranged in therapeutic room and(Or)Various visual angles are color
Color depth is imaged(RGBD)Equipment(System)Three-dimensional modeling determination is carried out to indoor static interesting target.Wherein, radiotherapy
The position of bed is obtained by demarcation, and the position of other static objects can be obtained by the matching of threedimensional model.Calculate patient's
Real-time pose simultaneously feeds back to Visual servoing control system, and radiation therapeutical bed is moved in the central areas such as Patients Treated by Radiotherapy plan;
4-3)Second level radiotherapy position
The real-time pose of second level radiotherapy position is determined by near-infrared multiple binocular stereovision imaging system.By to patient and
Radiation therapeutical bed region carries out real-time three-dimensional reconstruction, and is matched with static radiation therapeutic bed model, calculates the reality of patient
Shi Weizi, and Visual servoing control system is fed back to, radiation therapeutical bed is moved to the near central regions such as Patients Treated by Radiotherapy plan;
4-4)Third level radiotherapy position
Third level radiotherapy position pose passes through active binocular structure light scan equipment(System)It is determined that.Pass through real time scan patient
The central areas such as radiotherapy planning, and matched with entire patient's threedimensional model, calculate reality of the patient on radiation therapeutical bed
Shi Weizi, and result is fed back into Visual servoing control system, radiation therapeutical bed is moved to the central areas such as radiotherapy planning.
Reference picture 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, step 1)With step 2)Using multistage more Vision imaging systems come pair
The radiation therapeutical bed for carrying patient carries out the intelligently pendulum position control of dynamic vision monitoring and visual servo, specifically includes:
A) first order radiotherapy position:Various visual angles color depth imaging device is installed on the wall of therapeutic room's surrounding, for 10m*
Thick pendulum position in the range of 10m*10m therapeutic rooms physical space.In a 10m*10m*10m therapeutic room, more sets can be installed and regarded more
Angle color depth imaging device, to the radiation therapeutical bed in the therapeutic room(Carry patient), therapeutic equipment, imaging device, auxiliary
Treatment personnel and object carry out visual information collection.Using digital picture or video intelligent analyzing and processing algorithm, calculate in real time true
Determine radiation therapeutical bed(Carry patient), therapeutic equipment, imaging device, the current pose of auxiliary treatment personnel and object.Using regarding
Feel servo intelligent control algolithm to radiation therapeutical bed(Carry patient)It is controlled, control radiation therapeutical bed is moved by optimal path
The pose determined to radiotherapy planning, the positioning precision of patient reaches 0.5 meter when making radiotherapy position;
b)Second level radiotherapy position:The field range of the Near-infrared Double item stereo vision equipment used is 5m*5m*5m.Passing through
After the monitoring of various visual angles color depth imaging device puts position with servo intelligent control one-level, by covering Near-infrared Double item stereo vision more
System, to the radiation therapeutical bed in the therapeutic room(Carry patient), therapeutic equipment, imaging device, auxiliary treatment personnel and object
Carry out visual information collection.Using digital picture or video intelligent analyzing and processing algorithm, calculate determine radiation therapeutical bed in real time(Take
Carry patient), therapeutic equipment, imaging device, the current pose of auxiliary treatment personnel and object.Calculated using visual servo intelligent control
Method is to radiation therapeutical bed(Carry patient)It is controlled, control radiation therapeutical bed moves to radiotherapy planning determination by optimal path
Pose, the positioning precision of patient reaches 0.5 millimeter when making radiotherapy position;
c)Third level radiotherapy position:The field range of the active structural light three-dimensional scanning system used is 0.5m*0.5m*
0.5m, it can more accurately be put position.By active structural light three-dimensional scanning system, the radiation in the therapeutic room is controlled
Treat bed(Carry patient), therapeutic equipment, imaging device, auxiliary treatment personnel and object carry out visual information collection.Using numeral
Image or video intelligent analyzing and processing algorithm, calculate determine radiation therapeutical bed in real time(Carry patient), therapeutic equipment, imaging set
Standby, auxiliary treatment personnel and the current pose of object.Using visual servo intelligent control algorithm to radiation therapeutical bed(Carry and suffer from
Person)It is controlled, control radiation therapeutical bed is moved to the pose of radiotherapy planning determination by optimal path, patient when making radiotherapy position
Positioning precision reach 0.1 millimeter;Specially:Start airborne image system(CBCT)Intelligent control, and on radiation therapeutical bed
Patient carry out CBCT scannings, to determine the centers such as the current treatment of patient and actual tumor radiotherapy target area pose;Specific bag
Include following steps:
c-1):Reference picture 7, using near-infrared multiple binocular stereovision system, the airborne image system of monitoring(CBCT), radiation control
Treat bed(Carry patient), therapeutic equipment(Linear accelerator, multi-leaf optical grating), auxiliary treatment personnel and object current pose, and
By airborne imagery system(CBCT)Current pose and airborne imagery system(CBCT)Expected pose contrasted, caused mistake
Difference passes to airborne CBCT systems Visual servoing control device.The Visual servoing control device exports different according to different control information
Frequency control signal, control airborne image system(CBCT)Motor driver.Different frequency control signal controlled motors
Driver, produce different voltage and act on airborne imagery system(CBCT)Servomotor on so that servomotor turn
Speed is changed.Torque caused by the gear-box amplification servomotor being connected with servomotor, and act on airborne imagery system
(CBCT)On, make its motion, until airborne imagery system(CBCT)Current pose and expected pose error in permissible range
It is interior.DC power supply switch is sent to servomotor brake to instruct, pass through the power on servomotor by Visual servoing control device
Square controls servomotor stop motion, locks airborne imagery system(CBCT)Pose;
c-2) :Reference picture 8, in locking airborne imagery system(CBCT)After pose, by the expected pose of CBCT Image Plates and in real time
The Image Plate real-time pose of monitoring is contrasted, and caused error is transmitted to CBCT Image Plate servo controllers, the servo controller
CBCT Image Plate motor driver frequency control signals are exported, controlled motor driver exports different voltage, acts on CBCT
On Image Plate servomotor, so as to controlled motor rotating speed, after gear-box amplifies, certain moment loading is produced in Image Plate
On, it is moved, reaches image formation state;
c-3):Reference picture 9, in locking airborne imagery system(CBCT)After pose, by the expected pose and reality of CBCT X-ray machine
When the X-ray machine real-time pose that monitors contrasted, caused error is transmitted to servo controller, servo controller output X-ray machine
Motor driver frequency control signal, controlled motor driver export different voltage, acted on X-ray machine servomotor, from
And controlled motor rotating speed, after the amplification of gear-box, certain moment loading is produced on X-ray machine, is moved it, is arrived
Reach image position;
d)Fourth stage radiotherapy position:After by third level radiotherapy position, before starting treatment, pass through airborne imaging visual servo control
System control machine processed carries CBCT or CT scan before image system is treated to patient, and uses medical image intellectual analysis
Processing Algorithm analyzes and processes to corresponding CBCT or CT images, determines to treat isocentric current location in real time, will treat
Isocentric current location is contrasted with treating the centers such as the preceding radiotherapy planning determined, if error is not in permissible range
It is interior, then using visual servo intelligent control algorithm to radiation therapeutical bed and therapeutic equipment(Linear accelerator, multi-leaf optical grating)Carry out
The center such as motion control, the radiotherapy planning for ensureing to determine before the centers such as Current therapeutic and treatment is controlled in tolerance scope(2
Millimeter)It is interior, then carry out radiotherapy;
D) fourth stage pendulum position stage, using a kind of two-way consistent, high precision elastic method for registering based on Demons, meter is passed through
The Deformation Field calculated, radiotherapy planning step simulations are positioned into the target of prophylactic radiotherapy in CT images automatically and jeopardizes organ information and transforms to
On the CBCT images of radiotherapy position, the auto-changing for jeopardizing organ, target area profile and radiotherapy planning from plan CT to CBCT is realized.And
By adjusting radiation therapeutical bed, linear accelerator and multi-leaf optical grating in real time, ensure to jeopardize organ and target during per fractionated radiotherapy
Jeopardize organ and target area position and attitude error in allowed band in plan CT before the current pose in area and treatment, put when being treated with reducing
Treat the error of the centers such as target area and the center such as target of prophylactic radiotherapy in the radiotherapy planning of determination before treatment.Mobile system CBCT
Image is registering with the simulator locating CT images progress gathered during radiotherapy planning, and reference picture 13 comprises the following steps that:
101)Import fractionated radiotherapy process CBCT images and original scheme CT images;
102)Fractionated radiotherapy process CBCT images are registering with image progress in original scheme CT;
103)Generation target area and jeopardize this of organ and sketch outline by several times;
104)Before parameter, fractionated radiotherapy in original scheme CT CBCT images and this sketch outline recalculate dosage by several times
Distribution and dose volume histogram, judge whether to need to change radiotherapy planning with reference to original prescription constraint:Whether need, carry out
Step 1055) and step 106);If not, preserving original radiotherapy planning as this fractionated radiotherapy plan, this radiotherapy is completed
Journey;
105)The quick on-line amending of patient's radiotherapy planning is carried out with reference to clinical requirement, generates this fractionated radiotherapy plan;
106)Adjustment radiation therapeutical bed, linear accelerator and multi-leaf optical grating carry out this Patients During Radiotherapy.
During jeopardizing organ and delineating, present example uses a kind of automatic delineation method based on Atlas.First,
The N group CT images of collection are pre-processed, randomly selecting one group in the N group images of collection is used as reference picture, remaining work
It is for target image, reference picture is registering with all target images, count probability existing for each space pixel.Using repeatedly following
Ring iterative, create Atlas collection of illustrative plates.Probability graph time spectrum is established, suitable threshold value is chosen according to concrete application environment, collection of illustrative plates from general
Rate spatial transformation is to gray space.Afterwards, input patient this time plan CT images do target image, using affine registration algorithm with
Collection of illustrative plates carries out registration, by Deformation Field caused by registration process, will jeopardize organ contours information in collection of illustrative plates and is automatically transferred to patient
This time on plan CT.Finally, radiotherapy planning is formulated on this time plan CT of clinician or physics teacher after automatic segmentation, performed
Step 101).
In step 102)In registration process, the CBCT images of this radiotherapy and plan CT images are first subjected to Rigid Registration,
Translate and rotate;Carry out non-rigid registration again afterwards, i.e., it is affine with high-precision, the two-way consistent deformation based on Demons algorithms
Registration.
Driving floating image is obtained using Demons algorithmsI 0Each point corresponding to deformation vectors, you can obtain all pictures
The offset of vegetarian refreshments, realize floating image to the registration of reference picture.If floating image pointxThe grey scale pixel value at place is more than
The grey scale pixel value of respective point on reference picture, then this moves along positive gradient direction on floating image., whereas if float
The pixel grey scale of the point is less than the grey scale pixel value of respective point on reference picture on image, then the point on floating image will be along negative
Gradient direction moves, and displacement is by deformation vectorsd(x) determine.
Step 102)Image registration algorithm be iteration process, by constantly deformation field is optimized until algorithm receive
Hold back, realize accuracy registration.In order to ensure that the conversion is continuous in global scope, so as to keep the topological structure of image, each time
After iteration, Gaussian filter is usedGSkew obtained by smooth, improve computational efficiency.
To improve algorithm the convergence speed, the unidirectional force that driving force is spread from floating image pixel to reference picture pixel becomes
For the bidirectional force from two images gradient, i.e., the power of arbitrfary point not only drives floating image to be spread to reference picture on image,
Also reference picture is driven to be spread to floating image simultaneously.
Registration Algorithm flow such as Figure 14 based on demons.The selection of demons points is typically chosen current figure in step 204
As whole pixels.Step 205 tries to achieve Deformation Field using formula (3) and its improvement, and step 206 acts on the Deformation Field current floating
Motion video simultaneously carries out step 207 similarity measure, and searching makes the immediate Deformation Field of two images.
To ensure that registration result is two-way consistent, present example uses two-way consistent registration Algorithm.Two-way consistent registration
Algorithm is based on differomorphism demons, reference pictureI 1And floating imageI 0Deformation Field there is symmetrical characteristic, in differomorphism
Similarity measure criterion is ensured in transformation space.The deformable registration makes reference picture and floating image concurrent deformation, anyway
Select registration similarity measure and Optimal Parameters, the symmetric derivative homeomorphism ensure byI 1ArriveI 0Conversion with byI 0ArriveI 1Change
Change with uniformity.The two-way uniformity registration ensures the symmetry of two images transformed distances estimation, does not limit how to select
Still image and reference picture.By directly adding reversible limitation during optimization, this method ensures the registration of sub- voxel
The degree of accuracy and the reversible deformation of discrete domain.Meanwhile symmetrical deformation makes registration error minimum using invertibity.
This example is used by coarse and fine registration strategies, is constantly reduced the details or data volume of original image, is obtained one group
By coarse and fine image, rough registration first is carried out to the image with few details or small data quantity, after current layer registration is terminated
Deformation results do the initial value of next layer of registration parameter, as indicated in step 210, repeat this process, judged by step 208,
Until the registration between original image.Finally, step 209 output registration result.
In the present invention, step 203 forms image pyramid layering using direct extraction algorithm and does Multi-Resolution Registration framework,
Concretely comprise the following steps:
A. by reference pictureI 1And floating imageI 0Pretreatment obtains source imagesI 1 1And floating imageI 0 1;
B. by source imagesI 1 1And floating imageI 0 12 extract at equal intervals respectively, obtain image after original image extractsI 1 2WithI 0 2;
C. by imageI 1 2WithI 0 22 extract, obtain at equal intervals respectivelyI 1 2WithI 0 2Image after image contractI 1 4WithI 0 4;
D. by image I 1 4WithI 0 4Registration, obtain registration parameterθ 4;
E. withθ 4For primary iteration value, by imageI 1 2WithI 0 2Registration, obtain registration parameterθ 2;
F. withθ 2For primary iteration value, by source imagesI 1 1WithI 0 1Registration, obtain registration parameterθ 1。
In step 103)Target area and jeopardize the gradation CBCT navigational figures during organ delineates, obtained with reference to step 1) and
The deformation vector field that step 102) obtains, jeopardize organ based on the target area on plan CT and using what Atlas algorithms were delineated automatically
Profile, generate initial gradation and sketch outline.Initial gradation is sketched outline and sketches outline progress with original scheme CT
Contrast, modifies with reference to clinical requirement, makes it consistent with the anatomical structure of gradation navigational figure, and this gradation for generating target area hooks
Outline.
Parameter, fractionated radiotherapy image in original scheme CT and this sketch outline by several times recalculate dosage distribution and
Dose volume histogram, judge whether to need to change radiotherapy planning with reference to original prescription constraint;If desired for modification, then clinic is combined
It is required that carrying out the quick online modification of patient's radiotherapy planning, this fractionated radiotherapy plan is generated, so as to adjust radiation therapeutical bed, straight line
Accelerator and multi-leaf optical grating complete this Patients During Radiotherapy.As do not needed, then original radiotherapy planning is preserved as this fractionated radiotherapy meter
Draw and perform.
If in the radiotherapy position stage, the isocentric current location for the treatment of this moment can not be put with what is determined before treating all the time
The control errors of the position of center line such as plan are treated within allowed band, then corrects radiotherapy planning, establishes new radiotherapy meter again
Draw.It is measured with airborne imagery system again afterwards it is real-time wait center contrast, error is fed back into visual servo intelligent control
System, and by visual servo intelligent control algorithm, to radiation therapeutical bed and therapeutic equipment(Linear accelerator, multi-leaf optical grating)
Progress visual servo motion control, it is ensured that the center control such as the center such as treatment and the treatment preceding radiotherapy planning determined during per fractionated radiotherapy
System is in the range of clinical radiation therapy tolerance.
The radiotherapy planning stage, can be according only to a certain tumor radiotherapy simulator locating imaging image data(Such as CT)It is it is determined that swollen
Knurl target area and OAR poses, two kinds can also be used in combination, or two or more tumor radiotherapy simulator locating images determine tumor target
Area and OAR poses, such as combination tumor CT/MRI;Or CT/CBCT;Or PET/CT;Or PET/MRI;Or PET/CT/MRI.
In the radiotherapy position stage, a kind of monitoring optical imaging system of dynamic vision in real time can be used only(Such as binocular vision system
System)Real time dynamic tracing and high accuracy positioning radiation therapeutical bed(Carry patient)Aid in controlling in pose and therapeutic equipment, therapeutic room
Treatment personnel and the pose of object, two kinds, or two or more imaging systems can also be used in combination(Such as more binocular vision systems, knot
Structure photosystem, laser radar system) real time dynamic tracing and high accuracy positioning radiation therapeutical bed(Carry patient)Pose and treatment
The pose of auxiliary treatment personnel and object in equipment, therapeutic room.
Per fractionated radiotherapy during pendulum position, airborne CBCT systems, or CT system can be used to determine in high precision per interval procedure
Preceding target area and OAR current pose;Ultrasound or MRI system real time dynamic tracing and high accuracy can be used in therapeutic process really
Determine target area and OAR current pose in therapeutic process;The different times of therapeutic process, use different visual monitoring image intelligents
Analysis and processing method and visual servo intelligent control method carry out guiding in real time and control radiation therapeutical bed(Carry patient)With control
Equipment is treated, therapeutic process is carried out by radiotherapy planning, it is ensured that the quality of therapeutic process is effectively controlled.
As shown in figure 16, the visual monitoring in radiation therapy process of the present invention and visual servo intelligent control principle are as follows:
During tumour radiotherapy, pass through airborne imagery system(MRI or ultrasound)The image of real time scan collection, dynamic
Monitoring, tracking and high accuracy determine target area and OAR current pose in therapeutic process, by target area and OAR current pose with controlling
Treat front simulation be positioned to as system image determine radiotherapy planning tumor radiotherapy target area compared with OAR poses, if error is being held
Perhaps in the range of, then continue the therapeutic progresses of next step, if error has exceeded the scope allowed, error is transmitted to radiotherapy
Process vision servo intelligent control system.The present invention uses visual servo intelligent control algorithm, controls radiation therapeutical bed and treatment
Equipment(Linear accelerator, multi-leaf optical grating)Motion, the current pose of target area and OAR during ensureing per gradation tumor radiotherapy with
Radiotherapy planning tumor radiotherapy target area and the OAR position and attitude errors that treatment front simulation is positioned to determine as system image are in permissible range
Ensure the quality of therapeutic process.
If target area and OAR current pose and treatment front simulation can not be positioned to all the time in radiotherapy planning implementation procedure
The control errors of the radiotherapy planning tumor radiotherapy target area and the OAR poses that determine as system image in clinical radiation therapy permissible range,
It is then online as shown in figure 16 to correct radiotherapy planning in real time, then using Visual servoing control algorithm to radiation therapeutical bed and therapeutic equipment
(Linear accelerator, multi-leaf optical grating)Be controlled, make in Patients During Radiotherapy target area and OAR current pose and simulator locating this moment into
As the radiotherapy planning tumor radiotherapy target area that system image determines and OAR position and attitude errors are controlled in permissible range all the time, it is ensured that whole
The quality of individual radiation therapy process is guaranteed.
Intelligently controlled pendulum position with visual servo by the visual monitoring in tumour radiotherapy radiotherapy position stage, and complete the third level
Behind radiotherapy position, start airborne image system(CBCT)Intelligent control, and CBCT scannings are carried out to the patient on radiation therapeutical bed,
To determine the centers such as the current treatment of patient and tumor radiotherapy target area pose;Specifically comprise the following steps:
6-1)Using near-infrared multiple binocular stereovision system, the airborne image system of monitoring(CBCT), radiation therapeutical bed(Carry and suffer from
Person), therapeutic equipment(Linear accelerator, multi-leaf optical grating), auxiliary treatment personnel and object current pose, and by airborne imagery system
System(CBCT)Current pose and airborne imagery system(CBCT)Expected pose contrasted, caused error propagation is to airborne
CBCT system Visual servoing control devices.The Visual servoing control device exports different FREQUENCY CONTROLs according to different control information to be believed
Number, control airborne image system(CBCT)Motor driver.Different frequency control signal controlled motor drivers, produce not
Same voltage acts on airborne imagery system(CBCT)Servomotor on so that the driving servomotor of airborne imagery system
Rotating speed is changed.Torque caused by the gear-box amplification servomotor being connected with servomotor, and act on airborne imagery system
System(CBCT)On, make its motion, until airborne imagery system(CBCT)Current pose and the error of expected pose allowing model
In enclosing.Then, send DC power supply switch to servomotor brake by Visual servoing control device to instruct, and pass through servo electricity
Torque Control servomotor stop motion on machine, lock airborne imagery system(CBCT)Pose;
6-2) in locking airborne imagery system(CBCT)After pose, by the expected pose of CBCT Image Plates and the shadow that in real time monitors
As plate real-time pose is contrasted, caused error is transmitted to CBCT Image Plate servo controllers, servo controller output CBCT
Image version motor driver frequency control signal, controlled motor driver export different voltage, act on CBCT Image Plates and watch
Take on motor, so as to controlled motor rotating speed, after gear-box amplifies, produce certain moment loading on Image Plate, make it
Move to imaging pose;
6-3)In locking airborne imagery system(CBCT)After pose, by the expected pose of CBCT X-ray machine and the X that in real time monitors
Ray machine real-time pose is contrasted, and caused error is transmitted to servo controller, servo controller output X-ray machine motor driver
Frequency control signal, controlled motor driver export different voltage, acted on X-ray machine servomotor, so as to controlled motor
Rotating speed, after the amplification of gear-box, certain moment loading is produced on X-ray machine, causes it to move to imaging pose.
In the fourth stage radiotherapy position stage, pass through the height between simulator locating CT images when airborne CBCT images and radiotherapy planning
Precision registration, determines target of prophylactic radiotherapy and jeopardizes the Deformation Field of organ.Further, by the Deformation Field of determination, radiotherapy meter will be based on
The target of prophylactic radiotherapy and jeopardize organ information, radiotherapy planning information auto-changing that simulator locating CT is delineated when drawing, with holding and currently
CBCT images are consistent.And by SERVO CONTROL, radiation therapeutical bed, linear accelerator and multi-leaf optical grating pose are adjusted in real time, is ensured
Simulator locating CT images during radiotherapy planning before jeopardizing the current pose of organ and target of prophylactic radiotherapy during per fractionated radiotherapy and treating
In jeopardize organ and target area position and attitude error in allowed band, when being treated with reducing the center such as target of prophylactic radiotherapy with before treatment really
The error of the centers such as fixed radiotherapy planning.Comprise the following steps that:
7-1)Import fractionated radiotherapy process CBCT images and original scheme simulator locating CT images;
7-2)Fractionated radiotherapy process CBCT images carry out high registration accuracy with image in original scheme CT;
7-3)According to registering Deformation Field, automatic conversion target of prophylactic radiotherapy and jeopardize organ and sketch outline information;
7-4)Before parameter, fractionated radiotherapy in original scheme CT CBCT images and this sketch outline recalculate dosage by several times
Distribution and dose volume histogram, constrained with reference to original prescription dosage and radiotherapy clinical dosage, judge whether to need to change radiotherapy
Plan.Plan if desired for modification, then carry out 7-5) and 7-6);Plan need not be such as changed, then using former radiotherapy planning as currently
Fractionated radiotherapy plan, and implement current fractionated radiotherapy;
7-5)The quick on-line amending of patient's radiotherapy planning is carried out with reference to clinical requirement, automatically generates current fractionated radiotherapy plan;
7-6)According to amended radiotherapy planning, radiation therapeutical bed, linear accelerator and multi-leaf optical grating are adjusted by SERVO CONTROL
Implement current Patients During Radiotherapy after pose again.
A kind of image high registration accuracy algorithm is as follows:
8-1)The CBCT images of current fractionated radiotherapy and plan CT images are first subjected to Rigid Registration, that is, translates and rotates;It
Carry out non-rigid registration again afterwards.Present example is matched somebody with somebody using affine with the two-way consistent, high precision elastic based on Demons algorithms
Quasi- algorithm;
8-2)Image registration algorithm is multiple dimensioned, multiresolution, two-way consistent iteration registration process, by constantly to deformation field
Optimize until algorithmic statement, realize tumor imaging high precision elastic registration.The algorithm is converted using topography's information
Image, in order to ensure that the conversion is continuous in global scope, so as to keep the topological structure of image, after iteration each time, use
Skew obtained by Gaussian filter is smooth, improves computational efficiency.Further, to improve algorithm the convergence speed, by driving force
Bidirectional force from two images gradient is changed into from the unidirectional force that floating image pixel spreads to reference picture pixel, i.e., on image
The power of arbitrfary point not only drives floating image to be spread to reference picture, while also drives reference picture to be spread to floating image.
To ensure that registration result is two-way consistent, present example uses two-way consistent registration Algorithm.Two-way consistent registration
Algorithm is based on differomorphism demons, reference pictureI 1And floating imageI 0Deformation Field there is symmetrical characteristic, in differomorphism
Similarity measure criterion is ensured in transformation space.The deformable registration makes reference picture and floating image concurrent deformation, anyway
Select registration similarity measure and Optimal Parameters, the symmetric derivative homeomorphism ensure byI 1ArriveI 0Conversion with byI 0ArriveI 1Change
Change with uniformity.The two-way uniformity registration ensures the symmetry of two images transformed distances estimation, does not limit how to select
Still image and reference picture.By directly adding reversible limitation during optimization, this method ensures the registration of sub- voxel
The degree of accuracy and the reversible deformation of discrete domain.Meanwhile symmetrical deformation makes registration error minimum using invertibity.
Further, using Diagonal Recurrent Neural Network(DRNN)Visual servoing control algorithm is to radiation therapeutical bed, treatment
Equipment etc. carries out multistage visual servo intelligent control, and a kind of embodiment is as follows:
9-1)Diagonal Recurrent Neural Network structure
Diagonal Recurrent Neural Network(DRNN)With three layers:Input layer, hidden layer and output layer.Its structure chart is as shown in Figure 6.Its
It is characterised by:Between the neuron of Diagonal Recurrent Neural Network hidden layer or network be output to input between there is feedback connect
Connect, its feedback element also has time delayses characteristic, and network has short-term memory capability.The present invention a kind of embodiment be
The Patients During Radiotherapy of complexity is modeled using DRNN networks, recognized and control;
The expression formula of DRNN network states equation and output equation is:
(1)
(2)
(3)
In formula,It is the of networkIndividual input vector;It is in hidden layerThe summation of the input of individual neuron
Vector;It is hidden layerThe state vector of individual neuron;It is the output vector of network;It is Sigmoid letters
Number;
9-2)Diagonal Recurrent Neural Network identification structure
When being recognized using Diagonal Recurrent Neural Network to system,kFor the iterative steps of network,u(k) it is the defeated of identification objects
Incoming vector,y d (k) be identification objects output vector,y(k) be DRNN networks output vector.The object that needs are recognized
The difference of the input of input and output as DRNN networks, the output of identification objects and the output of DRNN networks is believed as DRNN adjustment
Number.And the parameter of identification network is adjusted according to the size of mean square error;
9-3)Diagonal Recurrent Neural Network learning algorithm
Present example carries out network using dynamic error back-propagation algorithm (Dynamic Back Propagation, DBP)
Training study;
9-4)The training flow of diagonal recurrent neural network
9-4-1):Initialization network parameter;
9-4-2):Forward-propagating, according to the expression formula of DRNN network states equation and output equation calculate the output of network to
Amount, the input vector of hidden layerAnd state vector;
9-4-3):The Identification Errors of calculating network;
9-4-4):Backpropagation, the connection weight between each layer of network is adjusted;
9-4-5):JudgeWhether performance requirement is reached;
9-4-6):If reaching requirement, weights are preserved, terminate training;If not up to, go to 9-4-2);
9-5)Radiation therapeutical bed Visual servoing control system based on DRNN
Desired image characteristic point is compared with the image characteristic point obtained by optical imagery detecting system, by error propagation
Give Visual servoing control device, the speed and Jacobean matrix of the radiation therapeutical bed mechanical arm tail end of output in three-dimensional Cartesian space
Battle array inverse matrix is multiplied, and obtains the speed in each joint of radiation therapeutical bed mechanical arm, and passes to each joint control of radiation therapeutical bed
Device, Diagonal Recurrent Neural Network and its corresponding, being inaccurately partly modified and distinguishing to radiation therapeutical bed kinetic model
Know, and pass the result to radiation therapeutical bed kinetic model, so as to control the pose in each joint of radiation therapeutical bed mechanical arm.
Compared with prior art, it is low, reliable effectively to solve at present clinically manual control Patients During Radiotherapy efficiency by the present invention
Property it is not high the problem of, improve radiotherapy implementation procedure operating efficiency and work quality, it is ensured that the quality of therapeutic process obtains
It is effectively ensured.
Above-described is only some embodiments of the present invention.For the person of ordinary skill of the art, not
On the premise of departing from the invention design, various modifications and improvements can be made, these belong to the protection model of the present invention
Enclose.
Claims (10)
1. tumour radiotherapy visual monitoring and visual servo intelligent control method, it is characterised in that:
Comprise the following steps:
1)Tumour radiotherapy overall process visual monitoring
1-1)By radiotherapy planning step simulations be positioned to as system acquisition patient CT, MRI, PET/CT and(Or)PET/MRI
Medical image, high accuracy determine tumor radiotherapy target area(GTV)With jeopardize organ(OAR)Pose;
1-2)By multiple color depth imaging systems in the phase treatment room of radiotherapy position, laser infrared radar imaging system, multigroup
The visible ray monitoring image that near infrared light binocular vision imaging system, structure light imaging system etc. gather in real time, in real time dynamic with
Track and high accuracy determine radiation therapeutical bed(Carry patient)With the position of auxiliary treatment personnel and object in therapeutic equipment and therapeutic room
Appearance;
1-3)Pass through radiotherapy position phase treatment equipment(Accelerator)CBCT the or CT medical images of airborne imaging system acquisition,
High accuracy determines the target of prophylactic radiotherapy before every interval procedure(GTV)With jeopardize organ(OAR)Current pose;
1-4)The ultrasound or MRI medical image data gathered in real time by radiotherapy planning execution stage airborne imaging system, it is real
When dynamic tracking and high accuracy determine radiotherapy planning implementation procedure in target of prophylactic radiotherapy(GTV)With jeopardize organ(OAR)It is current
Pose;
2)Tumour radiotherapy overall process visual servo intelligent control
2-1)According to being put in radiotherapy position stage optical Imaging: Monitoring system real time dynamic tracing and the therapeutic room determined in high precision
Penetrate therapeutic bed(Carry patient), therapeutic equipment, the spatial pose of auxiliary treatment personnel and object, pass through visual servo intelligent control
System(Algorithm), guide and control radiation therapeutical bed(Carry patient)Predetermined radiotherapy district is moved to, and makes the centre bits such as treatment
Put in the certain limits of center such as the radiotherapy planning for reaching the determination of radiotherapy planning stage;
2-2)Pass through automatic comparison therapeutic equipment(Accelerator)Every fractionated radiotherapy pendulum position stage that airborne imaging system image determines
The radiotherapy planning that target of prophylactic radiotherapy determines with jeopardizing current pose and the radiotherapy planning step simulations of organ to be positioned to as system image
Tumor radiotherapy target area and the pose for jeopardizing organ, automatic guiding and control, adjustment radiation therapeutical bed(Carry patient)Pose, make to work as
The centers such as the radiotherapy planning that the centers such as preceding treatment and radiotherapy planning stage determine are controlled in tolerance scope
It is interior, and make Pose Control when target of prophylactic radiotherapy pose and simulator locating plan in the range of clinical radiation therapy tolerance, then,
The radiotherapy planning determined by the radiotherapy planning stage carries out the treatment of this gradation;Otherwise, radiotherapy planning is corrected, and is put by revised
The plan for the treatment of carries out the treatment of this gradation, it is ensured that the quality of therapeutic process is effectively ensured;
2-3)Stage airborne imaging system image real time dynamic tracing and the treatment determined in high precision are performed according to radiotherapy planning
Target area and the current pose for jeopardizing organ in journey, guiding in real time and control radiation therapeutical bed(Carry patient)With the position of therapeutic equipment
Appearance, radiotherapy planning implementation procedure is set to be carried out by the radiotherapy planning determined before treatment;Otherwise, radiotherapy planning is corrected in real time, and by repairing
Radiotherapy planning after just carries out this interval procedure, it is ensured that the quality of therapeutic process is effectively ensured.
2. tumour radiotherapy visual monitoring according to claim 1 and visual servo intelligent control method, its feature exist
In:The present invention is realized by radiation treatment robot system;
Radiation treatment robot system is by tumour radiotherapy simulator locating medical image system, radiotherapy position optical imagery system
System, multiple degrees of freedom radiation therapeutical bed, the therapeutic equipment for carrying patient(Accelerator and multi-leaf optical grating), airborne medical image system etc.
Composition;
3)Further, visual monitoring and visual servo intelligence control system(Algorithm)By multistage more visual optics imagings, monitoring
Equipment(System)Composition, is specifically included:
(3-1)In radiotherapy room centered on the centers such as radiotherapy planning, to 10m*10m*10m(Wherein m:For long measure:Rice)
Interesting target in spatial dimension carries out vision positioning, and a kind of embodiment is as follows:
A certain fixing point establishes global space coordinates as origin using in room;
Using the big field laser radar that is fixedly mounted in quality room and(Or)The various visual angles color depth of multiple fixed installations into
Picture(RGBD)Equipment(System)The static three-dimensional for carrying out radiotherapy room is rebuild, and reconstructed results show as three-dimensional point cloud;According to this reconstruction
As a result, indoor static target is established(The interior space, radiation therapeutical bed, therapeutic equipment, airborne medical imaging devices etc.)Coarse positioning
Model, and their positions in global space coordinate system are determined, carried for first order radiotherapy position visual monitoring and visual servo
For the interesting target three dimensional local information of lower accuracy;
(3-2)In radiotherapy room centered on the centers such as radiotherapy planning, the interesting target in 5m*5m*5m spatial dimensions is entered
Row degree of precision vision positioning, a kind of embodiment are as follows:
Using the near-infrared multiple binocular stereovision equipment that multigroup radiotherapy apparatus is airborne(System), gather the radiation in radiation treatment region
Therapeutic bed and entire patient's realtime graphic;
It is based on(3-1)Coarse positioning model, determine radiation therapeutical bed static three-dimensional model, establish radiation therapeutical bed local coordinate
System;Using technique of binocular stereoscopic vision, moving target interested is monitored in real time(Radiation therapeutical bed, patient), determine that patient is complete
Pose under body three-dimensional space model and radiation therapeutical bed local coordinate system, watched for second level radiotherapy position visual monitoring and vision
Clothes provide the interesting target three-dimensional space position information of degree of precision;
(3-3)In radiotherapy room centered on the centers such as radiotherapy planning, to interested in 0.5m*0.5m*0.5m spatial dimensions
Target carries out high-precision vision positioning, and a kind of embodiment is as follows:
It is based on(3-2)Entire patient's three-dimensional space model of determination, establish patient's body coordinate system;
Using active structural light three-dimensional scanning device(System), Real-time modeling set is carried out to patient's local body surface marker,
And and(3-2)Entire patient's threedimensional model of middle determination carries out real-time matching, determines that patient's local volume surface marker position exists
Real time position and posture in patient's body coordinate system, provided for third level radiotherapy position visual monitoring and visual servo high-precision
Interesting target three-dimensional space position information.
3. tumour radiotherapy visual monitoring according to claim 2 and visual servo intelligent control method, its feature exist
In:It is as follows using the multistage pendulum position visual monitoring of tumour radiotherapy and visual servo, a kind of embodiment:
4-1)The foundation and demarcation of each coordinate system
This method is related to global coordinate system, radiation therapeutical bed local coordinate system, three kinds of space coordinates of patient's body coordinate system, is
Accurate three-dimensional position is obtained, it is necessary to be demarcated to coordinate systems at different levels, obtain the transformational relation between coordinate system;
The foundation rule of coordinate systems at different levels is as follows:
(4-1-1)Global coordinate system
Using a certain fixing point in radiotherapy room as origin, Descartes's three-dimensional system of coordinate is established with right-handed coordinate system;It is by radiotherapy room
Interior radiation therapeutical bed, therapeutic equipment, airborne medical imaging devices and the optical monitoring imaging device for carrying patient etc. is unified to one
The world coordinate system risen;
(4-1-2)Radiation therapeutical bed local coordinate system
Using a certain fixing point in radiation therapeutical bed as origin, the various kinds of equipment on radiation therapeutical bed is often corrected to this during task
Coordinate system, the equipment that optical imagery monitoring is used in therapeutic room are also demarcated under the coordinate system;
(4-1-3)Patient's body coordinate system
Using patient tumors radiotherapy planning isocenter point as origin, three-dimensional system of coordinate is established, X-axis represents left and right directions, and Y-axis represents head
Pin direction, Z axis represent abdomen back of the body direction, and horizontal plane is rotated in a clockwise direction vector value as just;
Wherein, the demarcation between global coordinate system and radiation therapeutical bed local coordinate system can be demarcated using cooperation mark, point
Some cooperation marks are not set on space and radiation therapeutical bed indoors, gathers 3 pairs using total powerstation and the above does not indicate collinearly
The coordinate in two spaces coordinate system of point, calculate rigid body transformational relation of the radiation therapeutical bed in global coordinate system;
Matched by threedimensional model, the transformation relation between calculating patient's body coordinate and radiation therapeutical bed local coordinate, while by
There can be local displacement in patient, therefore the conversion needs to carry out real time correction, it is real by Near-infrared Double item stereo vision equipment
When obtain patient position and attitude and be transformed under radiation therapeutical bed coordinate system;
Further, using following multistage radiotherapy position visual monitoring and visual servo intelligent control:
4-2)First order radiotherapy position
The real-time pose of first order radiotherapy position by the scanning of the big field laser that is arranged in therapeutic room and(Or)Various visual angles are color
Color depth is imaged(RGBD)Equipment(System)Three-dimensional modeling determination is carried out to indoor static interesting target;
Wherein, the position of radiation therapeutical bed is obtained by demarcation, and the position of other static objects can pass through of threedimensional model
With obtaining;
Calculate the real-time pose of patient and feed back to Visual servoing control system, radiation therapeutical bed is moved to Patients Treated by Radiotherapy plan
Deng in central area;
4-3)Second level radiotherapy position
The real-time pose of second level radiotherapy position is determined by near-infrared multiple binocular stereovision imaging system;
Carried out by carrying out real-time three-dimensional reconstruction to patient and radiation therapeutical bed region, and with static radiation therapeutic bed model
Match somebody with somebody, calculate the real-time pose of patient, and feed back to Visual servoing control system, radiation therapeutical bed is moved to Patients Treated by Radiotherapy meter
Draw and wait near central regions;
4-4)Third level radiotherapy position
Third level radiotherapy position pose passes through active binocular structure light scan equipment(System)It is determined that;Pass through real time scan patient
The central areas such as radiotherapy planning, and matched with entire patient's threedimensional model, calculate reality of the patient on radiation therapeutical bed
Shi Weizi, and result is fed back into Visual servoing control system, radiation therapeutical bed is moved to the central areas such as radiotherapy planning.
4. tumour radiotherapy visual monitoring according to claim 2 and visual servo intelligent control method, its feature exist
In:Step 1)With step 2)Dynamic vision prison is carried out to the radiation therapeutical bed for carrying patient using multistage more Vision imaging systems
Survey and visual servo intelligently puts level controlling system(Algorithm), a kind of embodiment is as follows:
5-1)First order radiotherapy position
Various visual angles color depth imaging device is installed on the wall of therapeutic room's surrounding, for 10m*10m*10m therapeutic rooms physics
Thick pendulum position in spatial dimension;
In a 10m*10m*10m therapeutic room, more set various visual angles color depth imaging devices can be installed, in the therapeutic room
Radiation therapeutical bed(Carry patient), therapeutic equipment, imaging device, auxiliary treatment personnel and object carry out visual information collection;
Using digital picture or video intelligent analyzing and processing algorithm, calculate determine radiation therapeutical bed in real time(Carry patient), treatment
Equipment, imaging device, the current pose of auxiliary treatment personnel and object;
Using visual servo intelligent control algorithm to radiation therapeutical bed(Carry patient)It is controlled, control radiation therapeutical bed is by most
Shortest path moves to the pose of radiotherapy planning determination, and the positioning precision of patient reaches 0.5 meter when making radiotherapy position;
5-2)Second level radiotherapy position
The field range of the Near-infrared Double item stereo vision equipment used is 5m*5m*5m;
After position is put with servo intelligent control one-level by the monitoring of various visual angles color depth imaging device, by covering Near-infrared Double more
Item stereo vision system, to the radiation therapeutical bed in the therapeutic room(Carry patient), therapeutic equipment, imaging device, auxiliary treatment
Personnel and object carry out visual information collection;
Using digital picture or video intelligent analyzing and processing algorithm, calculate determine radiation therapeutical bed in real time(Carry patient), treatment
Equipment, imaging device, the current pose of auxiliary treatment personnel and object;
Using visual servo intelligent control algorithm to radiation therapeutical bed(Carry patient)It is controlled, control radiation therapeutical bed is by most
Shortest path moves to the pose of radiotherapy planning determination, and the positioning precision of patient reaches 0.5 millimeter when making radiotherapy position;
5-3)Third level radiotherapy position
The field range of the active structural light three-dimensional scanning system-system used can be carried out more smart for 0.5m*0.5m*0.5m
True pendulum position;
By active structural light three-dimensional scanning system, to the radiation therapeutical bed in the therapeutic room(Carry patient), therapeutic equipment,
Imaging device, auxiliary treatment personnel and object carry out visual information collection;
Using digital picture or video intelligent analyzing and processing algorithm, calculate determine radiation therapeutical bed in real time(Carry patient), treatment
Equipment, imaging device, the current pose of auxiliary treatment personnel and object;
Using visual servo intelligent control algorithm to radiation therapeutical bed(Carry patient)It is controlled, control radiation therapeutical bed is by most
Shortest path moves to the pose of radiotherapy planning determination, and the positioning precision of patient reaches 0.1 millimeter when making radiotherapy position;
5-4)Fourth stage radiotherapy position
After by third level radiotherapy position, before starting treatment, shadow is carried by airborne imaging Visual servoing control system control machine
CBCT or CT scan before being treated as system to patient, and using medical image intellectual analysis Processing Algorithm to corresponding
CBCT or CT images are analyzed and processed, and are determined to treat isocentric current location in real time, will be treated isocentric current location
Contrasted with centers such as the radiotherapy plannings that is determined before treatment, if error not in permissible range, using visual servo
Intelligent control algorithm is to radiation therapeutical bed and therapeutic equipment(Linear accelerator, multi-leaf optical grating)Motion control is carried out, is ensured current
The centers such as the radiotherapy planning determined before the centers such as treatment and treatment are controlled in tolerance scope(2 millimeters)It is interior, then put
Penetrate treatment;
Further, all the time can not be true by the centers such as current treatment and radiotherapy planning stage if in the radiotherapy position stage
The control errors of the position of center line such as fixed radiotherapy planning are then corrected radiotherapy planning, established again within clinical allowed band
New radiotherapy planning;
It is measured with airborne imagery system again afterwards it is real-time wait center contrast, error is fed back into visual servo intelligent control
System, and by visual servo intelligent control algorithm, to radiation therapeutical bed and therapeutic equipment(Linear accelerator, multi-leaf optical grating)
Carry out visual servo motion control, it is ensured that true per the centers such as fractionated radiotherapy plan execution phase treatment and radiotherapy planning stage
The centers such as fixed radiotherapy planning are controlled in the range of clinical tolerance.
5. tumour radiotherapy visual monitoring according to claim 4 and visual servo intelligent control method, its feature exist
In:The radiotherapy planning stage, can be according only to a certain tumor radiotherapy simulator locating imaging image data(Such as CT)Determine that tumour is put
Treat target area(GTV)With jeopardize organ(OAR)Pose, two kinds can also be used in combination, or the simulation of two or more tumor radiotherapies is fixed
Position image determines tumor radiotherapy target area(GTV)With jeopardize organ(OAR)Pose, such as combination tumor CT/MRI;Or CT/CBCT;
Or PET/CT;Or PET/MRI;Or PET/CT/MRI;
In the radiotherapy position stage, a kind of monitoring optical imaging system of dynamic vision in real time can be used only(Such as binocular vision system)It is real
When dynamic monitoring, tracking and high accuracy positioning radiation therapeutical bed(Carry patient)Aid in controlling in pose and therapeutic equipment, therapeutic room
Treatment personnel and the pose of object, two kinds, or two or more imaging systems can also be used in combination(Such as more binocular vision systems, knot
Structure photosystem, laser radar system) real time dynamic tracing and high accuracy positioning radiation therapeutical bed(Carry patient)Pose and treatment
The pose of auxiliary treatment personnel and object in equipment, therapeutic room;
The pendulum position stage per fractionated radiotherapy, before airborne CBCT systems, or the every interval procedure of CT system high accuracy determination can be used
Tumor radiotherapy target area(GTV)With jeopardize organ(OAR)Current pose;In radiotherapy planning implementation procedure can use ultrasound or
MRI system real time dynamic tracing and high accuracy determine tumor radiotherapy target area in therapeutic process(GTV)With jeopardize organ(OAR)'s
The current pose of pose;The different times of therapeutic process, using different visual monitoring image intelligent analysis and processing methods and regard
Feel that servo intelligent control method carries out guiding in real time and control radiation therapeutical bed(Carry patient)And therapeutic equipment, make therapeutic process
Carried out by radiotherapy planning, it is ensured that the quality of Patients During Radiotherapy is effectively controlled.
6. tumour radiotherapy visual monitoring according to claim 5 and visual servo intelligent control method, its feature exist
In:Intelligently controlled pendulum position with visual servo by the visual monitoring in tumour radiotherapy radiotherapy position stage, and complete third level radiotherapy
After putting position, start airborne image system(CBCT)Intelligent control, and CBCT scannings are carried out to the patient on radiation therapeutical bed, with true
Determine the centers such as the current treatment of patient and tumor radiotherapy target area pose;Specifically comprise the following steps:
6-1)Using near-infrared multiple binocular stereovision system, airborne image system is monitored(CBCT), radiation therapeutical bed(Carry and suffer from
Person), therapeutic equipment(Linear accelerator, multi-leaf optical grating), auxiliary treatment personnel and object current pose, and by airborne imagery system
System(CBCT)Current pose and airborne imagery system(CBCT)Expected pose contrasted, caused error propagation is to airborne
CBCT system Visual servoing control devices;The Visual servoing control device exports different FREQUENCY CONTROLs according to different control information to be believed
Number, control airborne image system(CBCT)Motor driver;Different frequency control signal controlled motor drivers, produce not
Same voltage acts on airborne imagery system(CBCT)Servomotor on so that the driving servomotor of airborne imagery system
Rotating speed is changed;Torque caused by the gear-box amplification servomotor being connected with servomotor, and act on airborne imagery system
System(CBCT)On, make its motion, until airborne imagery system(CBCT)Current pose and the error of expected pose allowing model
In enclosing;Then, send DC power supply switch to servomotor brake by Visual servoing control device to instruct, and pass through servo electricity
Torque Control servomotor stop motion on machine, lock airborne imagery system(CBCT)Pose;
6-2) in locking airborne imagery system(CBCT)After pose, by the expected pose of CBCT Image Plates and the shadow that in real time monitors
As plate real-time pose is contrasted, caused error is transmitted to CBCT Image Plate servo controllers, servo controller output CBCT
Image version motor driver frequency control signal, controlled motor driver export different voltage, act on CBCT Image Plates and watch
Take on motor, so as to controlled motor rotating speed, after gear-box amplifies, produce certain moment loading on Image Plate, make it
Move to imaging pose;
6-3)In locking airborne imagery system(CBCT)After pose, by the expected pose of CBCT X-ray machine and the X-ray machine of monitoring
Real-time pose is contrasted, and caused error is transmitted to servo controller, servo controller output X-ray machine motor driver frequency
Control signal, controlled motor driver exports different voltage, acted on X-ray machine servomotor, so that controlled motor rotating speed,
After the amplification of gear-box, certain moment loading is produced on X-ray machine, causes it to move to imaging pose.
7. tumour radiotherapy visual monitoring according to claim 6 and visual servo intelligent control method, its feature exist
In:In the fourth stage radiotherapy position stage, pass through the high accuracy between simulator locating CT images when airborne CBCT images and radiotherapy planning
Registration, determine target of prophylactic radiotherapy and jeopardize the Deformation Field of organ;
Further, by the Deformation Field of determination, simulator locating CT is delineated during by based on radiotherapy planning target of prophylactic radiotherapy and jeopardize device
Official's information, radiotherapy planning information auto-changing, to keep consistent with current CBCT images;
And by SERVO CONTROL, radiation therapeutical bed, linear accelerator and multi-leaf optical grating pose are adjusted in real time, is ensured per fractionated radiotherapy
During jeopardize organ and target of prophylactic radiotherapy current pose and treatment before radiotherapy planning when simulator locating CT images in jeopardize organ
With target area position and attitude error in allowed band, when being treated by reduction in terms of radiotherapy of the center such as target of prophylactic radiotherapy with treating preceding determination
Draw etc. center error;
Comprise the following steps that:
7-1)Import fractionated radiotherapy process CBCT images and original scheme simulator locating CT images;
7-2)Fractionated radiotherapy process CBCT images carry out high registration accuracy with image in original scheme CT;
7-3)According to registering Deformation Field, automatic conversion target of prophylactic radiotherapy and jeopardize organ and sketch outline information;
7-4)Before parameter, fractionated radiotherapy in original scheme CT CBCT images and this sketch outline recalculate dosage by several times
Distribution and dose volume histogram, constrained with reference to original prescription dosage and radiotherapy clinical dosage, judge whether to need to change radiotherapy
Plan;
Plan if desired for modification, then carry out 7-5) and 7-6);Plan need not be such as changed, then is used as current point using former radiotherapy planning
Secondary radiotherapy planning, and implement current fractionated radiotherapy;
7-5)The quick on-line amending of patient's radiotherapy planning is carried out with reference to clinical requirement, automatically generates current fractionated radiotherapy plan;
7-6)According to amended radiotherapy planning, radiation therapeutical bed, linear accelerator and multi-leaf optical grating are adjusted by SERVO CONTROL
Implement current Patients During Radiotherapy after pose again.
8. tumour radiotherapy visual monitoring and visual servo intelligent control method according to claim 7, its feature
It is:A kind of image high registration accuracy algorithm is as follows:
8-1)The CBCT images of current fractionated radiotherapy and plan CT images are first subjected to Rigid Registration, that is, translates and rotates;It
Carry out non-rigid registration again afterwards;
Present example is using affine registering with the two-way consistent, high precision elastic based on Demons algorithms;
8-2)Image registration algorithm is multiple dimensioned, multiresolution, two-way consistent iteration registration process, by constantly to deformation
Field is optimized until algorithmic statement, realizes tumor imaging high precision elastic registration;
The algorithm using topography's information come changing image, in order to ensure that the conversion is continuous in global scope, so as to keep scheming
The topological structure of picture, after iteration each time, using skew of the Gaussian filter obtained by smooth, improve computational efficiency;
Further, for raising algorithm the convergence speed, the list that driving force is spread from floating image pixel to reference picture pixel
It is changed into the bidirectional force from two images gradient to power, i.e., the power of arbitrfary point not only drives floating image to reference picture on image
Diffusion, while also drive reference picture to be spread to floating image.
9. tumour radiotherapy visual monitoring according to claim 8 and visual servo intelligent control method, its feature exist
In:
To ensure that registration result is two-way consistent, present example uses two-way consistent registration Algorithm;
Two-way consistent registration Algorithm is based on differomorphism demons, reference pictureI 1And floating imageI 0Deformation Field have it is symmetrical
Characteristic, similarity measure criterion is ensured in differomorphism transformation space;
The deformable registration makes reference picture and floating image concurrent deformation, selects similarity measure and the optimization of registration anyway
Parameter, the symmetric derivative homeomorphism ensure byI 1ArriveI 0Conversion with byI 0ArriveI 1Conversion there is uniformity;
Two-way uniformity registration ensures the symmetry of two images transformed distances estimation, do not limit how to select still image and
Reference picture;By directly adding reversible limitation during optimization, this method ensure sub- voxel registration accuracy and from
Dissipate the reversible deformation in domain;Meanwhile symmetrical deformation makes registration error minimum using invertibity.
10. tumour radiotherapy visual monitoring according to claim 2 and visual servo intelligent control method, its feature exist
In:Using Diagonal Recurrent Neural Network(DRNN)Visual servoing control algorithm carries out multistage to radiation therapeutical bed, therapeutic equipment etc.
Visual servo intelligent control, a kind of embodiment are as follows:
9-1)Diagonal Recurrent Neural Network structure
Diagonal Recurrent Neural Network(DRNN)With three layers:Input layer, hidden layer and output layer;
Its structure chart is as shown in Figure 6;
It is characterized in that:Between the neuron of Diagonal Recurrent Neural Network hidden layer or network be output to input between there is
Feedback link, its feedback element also have time delayses characteristic, and network has short-term memory capability;
A kind of embodiment of the present invention is that the Patients During Radiotherapy of complexity is modeled using DRNN networks, recognized and control;
The expression formula of DRNN network states equation and output equation is:
(1)
(2)
(3)
In formula,It is the of networkIndividual input vector;It is in hidden layerThe summation of the input of individual neuron to
Amount;It is hidden layerThe state vector of individual neuron;It is the output vector of network;It is Sigmoid functions;
9-2)Diagonal Recurrent Neural Network identification structure
When being recognized using Diagonal Recurrent Neural Network to system,kFor the iterative steps of network,u(k) it is the defeated of identification objects
Incoming vector,y d (k) be identification objects output vector,y(k) be DRNN networks output vector;
Using the input and output for the object for needing to recognize as the input of DRNN networks, the output of identification objects and DRNN networks are defeated
Adjustment signal of the difference gone out as DRNN;
And the parameter of identification network is adjusted according to the size of mean square error;
9-3)Diagonal Recurrent Neural Network learning algorithm
Present example carries out network using dynamic error back-propagation algorithm (Dynamic Back Propagation, DBP)
Training study;
9-4)The training flow of diagonal recurrent neural network
9-4-1):Initialization network parameter;
9-4-2):Forward-propagating, according to the expression formula of DRNN network states equation and output equation calculate the output of network to
Amount, the input vector of hidden layerAnd state vector;
9-4-3):The Identification Errors of calculating network;
9-4-4):Backpropagation, the connection weight between each layer of network is adjusted;
9-4-5):JudgeWhether performance requirement is reached;
9-4-6):If reaching requirement, weights are preserved, terminate training;If not up to, go to 9-4-2);
9-5)Radiation therapeutical bed Visual servoing control system based on DRNN
Desired image characteristic point is compared with the image characteristic point obtained by optical imagery detecting system, by error propagation
Give Visual servoing control device, the speed and Jacobean matrix of the radiation therapeutical bed mechanical arm tail end of output in three-dimensional Cartesian space
Battle array inverse matrix is multiplied, and obtains the speed in each joint of radiation therapeutical bed mechanical arm, and passes to each joint control of radiation therapeutical bed
Device, Diagonal Recurrent Neural Network and its corresponding, being inaccurately partly modified and distinguishing to radiation therapeutical bed kinetic model
Know, radiation therapeutical bed kinetic model is passed the result to, so as to control the pose in each joint of radiation therapeutical bed mechanical arm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710688862.1A CN107358607A (en) | 2017-08-13 | 2017-08-13 | Tumour radiotherapy visual monitoring and visual servo intelligent control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710688862.1A CN107358607A (en) | 2017-08-13 | 2017-08-13 | Tumour radiotherapy visual monitoring and visual servo intelligent control method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107358607A true CN107358607A (en) | 2017-11-17 |
Family
ID=60287705
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710688862.1A Pending CN107358607A (en) | 2017-08-13 | 2017-08-13 | Tumour radiotherapy visual monitoring and visual servo intelligent control method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107358607A (en) |
Cited By (40)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107823807A (en) * | 2017-11-24 | 2018-03-23 | 北京新核医疗科技有限公司 | Treat bed system and control method |
CN107875526A (en) * | 2017-11-27 | 2018-04-06 | 温州医科大学附属第医院 | The accurate control method of radiotherapy instrument during a kind of ocular tumor adaptive radiation therapy |
CN108272513A (en) * | 2018-01-26 | 2018-07-13 | 珠海市俊凯机械科技有限公司 | Clinical orientation method, apparatus, computer equipment and storage medium |
CN108543234A (en) * | 2018-05-22 | 2018-09-18 | 太丛信息科技(上海)有限公司 | A kind of novel patient real-time three-dimensional optical positioning system |
CN108619621A (en) * | 2018-05-23 | 2018-10-09 | 郑向鹏 | A kind of system and method for being used for tumor patient radiotherapy precise positioning and putting position |
CN108663989A (en) * | 2018-04-25 | 2018-10-16 | 中国科学院合肥物质科学研究院 | A kind of inverse solution engineering and trajectory planning algorithm of five axis therapeutic bed |
CN108765547A (en) * | 2018-04-23 | 2018-11-06 | 北京林业大学 | A kind of method and its application of leaf morphology free-air correction |
CN109009211A (en) * | 2018-06-22 | 2018-12-18 | 联想(北京)有限公司 | Smart machine, the method and device based on ultrasound examination |
WO2019127539A1 (en) * | 2017-12-29 | 2019-07-04 | Shenzhen United Imaging Healthcare Co., Ltd. | Systems and methods for determining region of interest in medical imaging |
CN110364253A (en) * | 2018-04-11 | 2019-10-22 | 西门子医疗有限公司 | The system and method that patient for auxiliary positions |
CN110517242A (en) * | 2019-08-23 | 2019-11-29 | 强联智创(北京)科技有限公司 | A kind of aneurysmal analysis method and device |
CN110689942A (en) * | 2019-08-28 | 2020-01-14 | 广州普天云健康科技发展有限公司 | Fishing net algorithm implementation method and device based on 3D accurate radiotherapy platform |
CN110911002A (en) * | 2019-11-20 | 2020-03-24 | 广州科莱瑞迪医疗器材股份有限公司 | Monitoring method and device for radiotherapy fixing device and monitoring system for radiotherapy fixing device |
CN111223575A (en) * | 2020-01-14 | 2020-06-02 | 于金明 | Radiotherapy auxiliary display method and system based on virtual intelligent medical platform |
CN111223576A (en) * | 2020-01-14 | 2020-06-02 | 于金明 | Patient propaganda and education system based on virtual intelligent medical platform |
WO2020125261A1 (en) * | 2018-12-21 | 2020-06-25 | 苏州雷泰医疗科技有限公司 | Radiotherapy auxiliary system and method thereof employing three-dimensional reconstruction of infrared and visible light |
CN111557676A (en) * | 2020-05-13 | 2020-08-21 | 山东省肿瘤防治研究院(山东省肿瘤医院) | System and equipment for dynamically adjusting target area position according to change of tumor in radiotherapy process |
CN111584034A (en) * | 2020-04-14 | 2020-08-25 | 四川省肿瘤医院 | Radiation therapy implementation quality control method and system based on artificial intelligence |
US10789498B2 (en) | 2017-12-29 | 2020-09-29 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for patient positioning |
CN111870825A (en) * | 2020-07-31 | 2020-11-03 | 于金明 | Radiotherapy precise field-by-field positioning method based on virtual intelligent medical platform |
CN112263788A (en) * | 2020-11-02 | 2021-01-26 | 浙江省肿瘤医院 | Quantitative detection system for morphological change in radiotherapy process |
CN112316318A (en) * | 2020-11-06 | 2021-02-05 | 中国科学院近代物理研究所 | Positioning guide system and method for image-guided radiotherapy |
CN112438805A (en) * | 2019-08-28 | 2021-03-05 | 美国发现集团有限公司 | Magnetic therapy body nanometer robot control system |
WO2021056446A1 (en) * | 2019-09-27 | 2021-04-01 | 西安大医集团股份有限公司 | Method, device, and system for detecting patient movement state |
CN112604187A (en) * | 2020-12-31 | 2021-04-06 | 河南省肿瘤医院 | Remote radiotherapy cooperation system and method |
CN112843503A (en) * | 2021-01-29 | 2021-05-28 | 中国人民解放军陆军军医大学第二附属医院 | Radiotherapy in-vivo dose monitoring method |
CN113230548A (en) * | 2017-12-21 | 2021-08-10 | 瓦里安医疗系统国际股份公司 | Virtual portal direction viewing imaging for patient setup in radiation therapy |
CN113274654A (en) * | 2021-06-29 | 2021-08-20 | 程明霞 | Real-time monitoring system for ultrasound images of tumor and visceral organ positions for radiotherapy |
CN113476755A (en) * | 2021-07-19 | 2021-10-08 | 迈胜医疗设备有限公司 | Transfer method, device, system, electronic device, and storage medium |
CN113491845A (en) * | 2021-08-02 | 2021-10-12 | 上海联影医疗科技股份有限公司 | Radiotherapy system, radiotherapy device and storage medium |
CN113975661A (en) * | 2021-11-16 | 2022-01-28 | 神州医疗科技股份有限公司 | Quality control method, device and system for monitoring treatment equipment and storage medium |
CN114073827A (en) * | 2020-08-15 | 2022-02-22 | 中硼(厦门)医疗器械有限公司 | Radiation irradiation system and control method thereof |
CN114177545A (en) * | 2022-01-17 | 2022-03-15 | 中国科学院合肥物质科学研究院 | Non-contact respiratory rhythm monitoring device and method used in radiotherapy |
CN114247061A (en) * | 2021-12-07 | 2022-03-29 | 苏州雷泰医疗科技有限公司 | Tumor dynamic tracking control method and device and radiotherapy equipment |
CN114452548A (en) * | 2021-09-27 | 2022-05-10 | 西安大医集团股份有限公司 | Radiotherapy monitoring system and radiotherapy system |
CN114452546A (en) * | 2021-09-26 | 2022-05-10 | 西安大医集团股份有限公司 | Radiotherapy monitoring system and control method thereof |
WO2023115261A1 (en) * | 2021-12-20 | 2023-06-29 | 西安大医集团股份有限公司 | Radiotherapy device control method and apparatus, device, and storage medium |
CN116872216A (en) * | 2023-08-28 | 2023-10-13 | 安徽工业大学 | Robot vision servo operation method based on finite time control |
CN117745732A (en) * | 2024-02-21 | 2024-03-22 | 江苏康众数字医疗科技股份有限公司 | Auxiliary positioning guide method for medical examination and image desensitization processing method |
RU2820456C1 (en) * | 2020-08-15 | 2024-06-04 | Нойборон Терапи Систем Лтд. | System for therapeutic exposure to radioactive beam and method for controlling it |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN203220692U (en) * | 2013-04-06 | 2013-10-02 | 成都威铭科技有限公司 | Real-time image automatic tracking system |
CN106139414A (en) * | 2016-06-23 | 2016-11-23 | 深圳市奥沃医学新技术发展有限公司 | A kind of position monitoring method for radiotherapy system, device and radiotherapy system |
CN106310544A (en) * | 2016-09-30 | 2017-01-11 | 上海联影医疗科技有限公司 | Real-time monitoring method and device for tumors and radiotherapy system |
CN106943678A (en) * | 2017-04-06 | 2017-07-14 | 北京大学第医院 | A kind of method and device of automatic radiotherapy position |
-
2017
- 2017-08-13 CN CN201710688862.1A patent/CN107358607A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN203220692U (en) * | 2013-04-06 | 2013-10-02 | 成都威铭科技有限公司 | Real-time image automatic tracking system |
CN106139414A (en) * | 2016-06-23 | 2016-11-23 | 深圳市奥沃医学新技术发展有限公司 | A kind of position monitoring method for radiotherapy system, device and radiotherapy system |
CN106310544A (en) * | 2016-09-30 | 2017-01-11 | 上海联影医疗科技有限公司 | Real-time monitoring method and device for tumors and radiotherapy system |
CN106943678A (en) * | 2017-04-06 | 2017-07-14 | 北京大学第医院 | A kind of method and device of automatic radiotherapy position |
Cited By (61)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107823807A (en) * | 2017-11-24 | 2018-03-23 | 北京新核医疗科技有限公司 | Treat bed system and control method |
CN107875526B (en) * | 2017-11-27 | 2020-01-24 | 温州医科大学附属第一医院 | Accurate control method of radiotherapy instrument in self-adaptive radiotherapy of eye tumor |
CN107875526A (en) * | 2017-11-27 | 2018-04-06 | 温州医科大学附属第医院 | The accurate control method of radiotherapy instrument during a kind of ocular tumor adaptive radiation therapy |
CN113230548A (en) * | 2017-12-21 | 2021-08-10 | 瓦里安医疗系统国际股份公司 | Virtual portal direction viewing imaging for patient setup in radiation therapy |
US11964171B2 (en) | 2017-12-21 | 2024-04-23 | Siemens Healthineers International Ag | Virtual beam's-eye view imaging in radiation therapy for patient setup |
US10825170B2 (en) | 2017-12-29 | 2020-11-03 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for determining a region of interest in medical imaging |
US11295153B2 (en) | 2017-12-29 | 2022-04-05 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for patient positioning |
WO2019127539A1 (en) * | 2017-12-29 | 2019-07-04 | Shenzhen United Imaging Healthcare Co., Ltd. | Systems and methods for determining region of interest in medical imaging |
US11532083B2 (en) | 2017-12-29 | 2022-12-20 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for determining a region of interest in medical imaging |
US10789498B2 (en) | 2017-12-29 | 2020-09-29 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for patient positioning |
US11730396B2 (en) | 2017-12-29 | 2023-08-22 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for patient positioning |
CN108272513A (en) * | 2018-01-26 | 2018-07-13 | 珠海市俊凯机械科技有限公司 | Clinical orientation method, apparatus, computer equipment and storage medium |
CN110364253B (en) * | 2018-04-11 | 2024-03-01 | 西门子医疗有限公司 | System and method for assisted patient positioning |
CN110364253A (en) * | 2018-04-11 | 2019-10-22 | 西门子医疗有限公司 | The system and method that patient for auxiliary positions |
CN108765547A (en) * | 2018-04-23 | 2018-11-06 | 北京林业大学 | A kind of method and its application of leaf morphology free-air correction |
CN108663989B (en) * | 2018-04-25 | 2021-04-06 | 中国科学院合肥物质科学研究院 | Inverse solution engineering and trajectory planning algorithm of five-axis treatment bed |
CN108663989A (en) * | 2018-04-25 | 2018-10-16 | 中国科学院合肥物质科学研究院 | A kind of inverse solution engineering and trajectory planning algorithm of five axis therapeutic bed |
CN108543234A (en) * | 2018-05-22 | 2018-09-18 | 太丛信息科技(上海)有限公司 | A kind of novel patient real-time three-dimensional optical positioning system |
CN108619621A (en) * | 2018-05-23 | 2018-10-09 | 郑向鹏 | A kind of system and method for being used for tumor patient radiotherapy precise positioning and putting position |
CN108619621B (en) * | 2018-05-23 | 2020-08-21 | 郑向鹏 | System for be used for accurate location of tumour patient radiation therapy and pendulum position |
CN109009211A (en) * | 2018-06-22 | 2018-12-18 | 联想(北京)有限公司 | Smart machine, the method and device based on ultrasound examination |
WO2020125261A1 (en) * | 2018-12-21 | 2020-06-25 | 苏州雷泰医疗科技有限公司 | Radiotherapy auxiliary system and method thereof employing three-dimensional reconstruction of infrared and visible light |
CN110517242A (en) * | 2019-08-23 | 2019-11-29 | 强联智创(北京)科技有限公司 | A kind of aneurysmal analysis method and device |
CN110689942A (en) * | 2019-08-28 | 2020-01-14 | 广州普天云健康科技发展有限公司 | Fishing net algorithm implementation method and device based on 3D accurate radiotherapy platform |
CN112438805A (en) * | 2019-08-28 | 2021-03-05 | 美国发现集团有限公司 | Magnetic therapy body nanometer robot control system |
CN112867537B (en) * | 2019-09-27 | 2023-02-21 | 西安大医集团股份有限公司 | Method, device and system for detecting movement state of patient |
WO2021056446A1 (en) * | 2019-09-27 | 2021-04-01 | 西安大医集团股份有限公司 | Method, device, and system for detecting patient movement state |
CN112867537A (en) * | 2019-09-27 | 2021-05-28 | 西安大医集团股份有限公司 | Method, device and system for detecting movement state of patient |
CN110911002A (en) * | 2019-11-20 | 2020-03-24 | 广州科莱瑞迪医疗器材股份有限公司 | Monitoring method and device for radiotherapy fixing device and monitoring system for radiotherapy fixing device |
CN111223575A (en) * | 2020-01-14 | 2020-06-02 | 于金明 | Radiotherapy auxiliary display method and system based on virtual intelligent medical platform |
CN111223576A (en) * | 2020-01-14 | 2020-06-02 | 于金明 | Patient propaganda and education system based on virtual intelligent medical platform |
CN111584034A (en) * | 2020-04-14 | 2020-08-25 | 四川省肿瘤医院 | Radiation therapy implementation quality control method and system based on artificial intelligence |
CN111557676A (en) * | 2020-05-13 | 2020-08-21 | 山东省肿瘤防治研究院(山东省肿瘤医院) | System and equipment for dynamically adjusting target area position according to change of tumor in radiotherapy process |
CN111557676B (en) * | 2020-05-13 | 2023-12-19 | 山东省肿瘤防治研究院(山东省肿瘤医院) | System and equipment for dynamically adjusting target area position by tumor change in radiotherapy process |
CN111870825A (en) * | 2020-07-31 | 2020-11-03 | 于金明 | Radiotherapy precise field-by-field positioning method based on virtual intelligent medical platform |
CN111870825B (en) * | 2020-07-31 | 2023-08-18 | 于金明 | Radiation therapy accurate field-by-field positioning method based on virtual intelligent medical platform |
WO2022037468A1 (en) * | 2020-08-15 | 2022-02-24 | 中硼(厦门)医疗器械有限公司 | Radioactive ray radiation system and control method therefor |
CN114073827A (en) * | 2020-08-15 | 2022-02-22 | 中硼(厦门)医疗器械有限公司 | Radiation irradiation system and control method thereof |
RU2820456C1 (en) * | 2020-08-15 | 2024-06-04 | Нойборон Терапи Систем Лтд. | System for therapeutic exposure to radioactive beam and method for controlling it |
CN114073827B (en) * | 2020-08-15 | 2023-08-04 | 中硼(厦门)医疗器械有限公司 | Radiation irradiation system and control method thereof |
CN112263788A (en) * | 2020-11-02 | 2021-01-26 | 浙江省肿瘤医院 | Quantitative detection system for morphological change in radiotherapy process |
CN112316318A (en) * | 2020-11-06 | 2021-02-05 | 中国科学院近代物理研究所 | Positioning guide system and method for image-guided radiotherapy |
CN112604187A (en) * | 2020-12-31 | 2021-04-06 | 河南省肿瘤医院 | Remote radiotherapy cooperation system and method |
CN112843503A (en) * | 2021-01-29 | 2021-05-28 | 中国人民解放军陆军军医大学第二附属医院 | Radiotherapy in-vivo dose monitoring method |
CN113274654A (en) * | 2021-06-29 | 2021-08-20 | 程明霞 | Real-time monitoring system for ultrasound images of tumor and visceral organ positions for radiotherapy |
CN113274654B (en) * | 2021-06-29 | 2024-03-01 | 程明霞 | Tumor and organ position ultrasonic image real-time monitoring system for radiotherapy |
CN113476755B (en) * | 2021-07-19 | 2023-11-21 | 迈胜医疗设备有限公司 | Transfer method, device, system, electronic equipment and storage medium |
CN113476755A (en) * | 2021-07-19 | 2021-10-08 | 迈胜医疗设备有限公司 | Transfer method, device, system, electronic device, and storage medium |
CN113491845A (en) * | 2021-08-02 | 2021-10-12 | 上海联影医疗科技股份有限公司 | Radiotherapy system, radiotherapy device and storage medium |
CN114452546A (en) * | 2021-09-26 | 2022-05-10 | 西安大医集团股份有限公司 | Radiotherapy monitoring system and control method thereof |
CN114452548A (en) * | 2021-09-27 | 2022-05-10 | 西安大医集团股份有限公司 | Radiotherapy monitoring system and radiotherapy system |
CN113975661A (en) * | 2021-11-16 | 2022-01-28 | 神州医疗科技股份有限公司 | Quality control method, device and system for monitoring treatment equipment and storage medium |
CN113975661B (en) * | 2021-11-16 | 2024-03-01 | 神州医疗科技股份有限公司 | Quality control method, device and system for monitoring treatment equipment and storage medium |
CN114247061A (en) * | 2021-12-07 | 2022-03-29 | 苏州雷泰医疗科技有限公司 | Tumor dynamic tracking control method and device and radiotherapy equipment |
WO2023115261A1 (en) * | 2021-12-20 | 2023-06-29 | 西安大医集团股份有限公司 | Radiotherapy device control method and apparatus, device, and storage medium |
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 |
CN114177545A (en) * | 2022-01-17 | 2022-03-15 | 中国科学院合肥物质科学研究院 | Non-contact respiratory rhythm monitoring device and method used in radiotherapy |
CN116872216A (en) * | 2023-08-28 | 2023-10-13 | 安徽工业大学 | Robot vision servo operation method based on finite time control |
CN116872216B (en) * | 2023-08-28 | 2023-12-08 | 安徽工业大学 | Robot vision servo operation method based on finite time control |
CN117745732A (en) * | 2024-02-21 | 2024-03-22 | 江苏康众数字医疗科技股份有限公司 | Auxiliary positioning guide method for medical examination and image desensitization processing method |
CN117745732B (en) * | 2024-02-21 | 2024-05-10 | 江苏康众数字医疗科技股份有限公司 | Auxiliary positioning guide method for medical examination and image desensitization processing method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107358607A (en) | Tumour radiotherapy visual monitoring and visual servo intelligent control method | |
JP6960509B2 (en) | Neural network for generating synthetic medical images | |
Bert et al. | A phantom evaluation of a stereo‐vision surface imaging system for radiotherapy patient setup | |
CN111627521B (en) | Enhanced utility in radiotherapy | |
CN106920234B (en) | Combined automatic radiotherapy planning method | |
US7835500B2 (en) | Multi-phase registration of 2-D X-ray images to 3-D volume studies | |
CN110378881B (en) | Tumor positioning system based on deep learning | |
CN105031833A (en) | Dosage verification system for radiotherapy apparatus | |
EP3285639B1 (en) | Monitoring a patient's position using a planning image and subsequent thermal imaging | |
CN106408509A (en) | Registration method and apparatus | |
CN105167788B (en) | Slur is as C arm systems | |
CN102622775B (en) | A kind of real-time dynamic reconstruction technology of heart compensated based on model interpolation | |
CN113041516B (en) | Method, system, processing equipment and storage medium for guiding positioning of three-dimensional image | |
CN110706779B (en) | Automatic generation method of accurate target function of radiotherapy plan | |
CN104867104A (en) | Method for obtaining anatomical structural atlas for target mouse based on XCT image non-rigid registration | |
CN104636618B (en) | Radiotherapy planning system | |
Fan et al. | Verification of the machine delivery parameters of a treatment plan via deep learning | |
CN116958443A (en) | SMPLX-based digital human quantitative detection model reconstruction method and application | |
CN108460829B (en) | A kind of 3-D image register method for AR system | |
CN113041515A (en) | Three-dimensional image guided moving organ positioning method, system and storage medium | |
Calow et al. | Photogrammetric measurement of patients in radiotherapy | |
Samadi Miandoab et al. | Calculation of inter-and intra-fraction motion errors at external radiotherapy using a markerless strategy based on image registration combined with correlation model | |
CN113744320B (en) | Intelligent ion beam self-adaptive radiotherapy system, storage medium and equipment | |
CN111408066A (en) | Tumor position calibration system and equipment based on magnetic resonance image | |
Steininger et al. | A novel class of machine-learning-driven real-time 2D/3D tracking methods: texture model registration (TMR) |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171117 |