WO2022222502A1 - 一种肿瘤的运动估计方法、装置、终端设备和存储介质 - Google Patents

一种肿瘤的运动估计方法、装置、终端设备和存储介质 Download PDF

Info

Publication number
WO2022222502A1
WO2022222502A1 PCT/CN2021/138010 CN2021138010W WO2022222502A1 WO 2022222502 A1 WO2022222502 A1 WO 2022222502A1 CN 2021138010 W CN2021138010 W CN 2021138010W WO 2022222502 A1 WO2022222502 A1 WO 2022222502A1
Authority
WO
WIPO (PCT)
Prior art keywords
tumor
organ
motion estimation
velocity vector
vector field
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.)
Ceased
Application number
PCT/CN2021/138010
Other languages
English (en)
French (fr)
Inventor
胡颖
赵保亮
雷隆
唐华杰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2022222502A1 publication Critical patent/WO2022222502A1/zh
Priority to US18/366,704 priority Critical patent/US12537106B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G06T7/251—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments involving models
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/40—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
    • G16H40/63—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices for local operation
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10076—4D tomography; Time-sequential 3D tomography
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06T2207/30056—Liver; Hepatic
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06T2207/30061—Lung
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06T2207/30092—Stomach; Gastric
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30204—Marker
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06T2207/30241—Trajectory

Definitions

  • the present application relates to the technical field of image processing, and in particular, to a method, apparatus, terminal device and storage medium for motion estimation of tumors.
  • Percutaneous medical image-guided interventional surgery is a common method for cancer diagnosis and treatment, but during the actual operation, the movement of the tumor and organs caused by the patient's physiological respiration increases the puncture needle accurately without damaging important structures such as blood vessels around the tumor. Difficulty inserting into the tumor site.
  • a breathing movement model is usually established to estimate the breathing movement of the target in real time.
  • the method mainly uses the strong correlation between the easily obtained proxy signal and the internal target motion to establish the correlation model between the two, and estimates the target motion by detecting the proxy signal in real time during the operation.
  • this method can usually only simulate the respiration motion of the tumor or a single anatomical feature point, and is not applicable to the whole organ, and cannot achieve accurate localization of the tumor and its surrounding important anatomical structures.
  • the embodiments of the present application provide a tumor motion estimation method, device, terminal device and storage medium, which can estimate the respiratory motion of the organ where the tumor is located in real time, and improve the accuracy of locating the tumor and its surrounding important anatomical structures. sex.
  • a first aspect of the embodiments of the present application provides a method for estimating tumor motion, including:
  • the respiration-related signal reflecting the motion trajectory characteristics of the patient's designated organ with a tumor under different respiration states
  • the current value of the respiration-related signal is input into the tumor motion estimation model to obtain the estimated current position of the tumor, and the tumor motion estimation model is constructed with a priori tumor location data set and the respiration-related signal as prior knowledge.
  • the prior tumor location data set is determined according to the pre-collected image data set of the designated organ, and includes the position of the tumor in the different breathing states, and the image data set includes the designated organ in the the three-dimensional images of the different breathing states;
  • the organ motion estimation model is constructed using a priori velocity vector field and the prior tumor location dataset as prior knowledge, and the prior velocity vector field is determined from the image dataset and includes the specified organ at The velocity vector field in each of the different breathing states.
  • the embodiment of the present application proposes a hierarchical estimation framework from tumor to whole organ motion under free breathing motion, the framework includes a tumor motion estimation model and an organ motion estimation model, wherein the tumor motion estimation model is based on a prior tumor position data set
  • the respiration-related signal is constructed as prior knowledge
  • the organ motion estimation model is constructed with prior velocity vector field and prior tumor location dataset as prior knowledge.
  • the patient's respiration-related signal can be input into the tumor motion estimation model as a proxy signal to obtain an estimated tumor position; then, the estimated tumor position can be input into the organ motion estimation model , to obtain the estimated velocity vector field of the whole organ, thereby realizing the estimation of the respiratory motion of the whole organ, and improving the accuracy of locating the tumor and its surrounding important anatomical structures.
  • the respiration-related signal is a motion track signal of an optical marker disposed on a specified part of the patient's body
  • the motion track signal includes the spatial position of the optical marker at each time point , before inputting the current value of the respiration-related signal into the tumor motion estimation model, it may also include:
  • a position paired data set is constructed according to the prior tumor position data set and the motion trajectory signal, and the position paired data set includes the tumor position and the position of the optical marker in the motion track signal;
  • the tumor motion estimation model is constructed and obtained.
  • the tumor motion estimation model which may include:
  • a corresponding tumor motion estimation model is established in each preset spatial coordinate direction.
  • the method before inputting the current value of the respiration-related signal into the tumor motion estimation model, the method may further include:
  • a breathing state is selected from the different breathing states as a reference state, and a differential homeomorphic deformation registration process is performed on the volume data in the reference state and the volume data in other states to obtain the prior velocity vector field , the other states are other breathing states in the different breathing states except the reference state;
  • the organ motion estimation model is constructed and obtained.
  • the update value of the velocity vector field corresponding to the breathing state is calculated by adopting an alternate optimization strategy, and The updated value and the initial value are added to obtain the velocity vector field of the specified organ in the breathing state.
  • inputting the estimated current position of the tumor into an organ motion estimation model to obtain the estimated current velocity vector field of the specified organ may include:
  • the velocity vector of the position point is used as a function of the estimated current position of the tumor, and according to the estimated current position of the tumor, the method of spatial interpolation is used to move to the position point Interpolate between the velocity vectors in the different breathing states to obtain the estimated current velocity vector of the position point, wherein the velocity vectors of the position point in the different breathing states are based on the prior velocity vector field Sure.
  • the method may further include:
  • a preset Gaussian kernel function is used to normalize the current velocity vector field, and the current dense displacement field of the designated organ is obtained by means of group exponential transformation, and the current dense displacement field includes each current displacement field of the designated organ.
  • the displacement vectors corresponding to the position points respectively.
  • a second aspect of the embodiments of the present application provides a tumor motion estimation device, including:
  • a respiratory-related signal acquisition module configured to acquire the current value of the patient's respiratory-related signal, where the respiratory-related signal reflects the motion trajectory characteristics of the patient's designated organ with a tumor under different respiratory states;
  • a tumor motion estimation module configured to input the current value of the respiration-related signal into a tumor motion estimation model to obtain an estimated current position of the tumor, the tumor motion estimation model is associated with the respiration with a prior tumor position data set
  • the signal is constructed as a priori knowledge, and the prior tumor location data set is determined according to the pre-collected image data set of the designated organ, and includes the location of the tumor in the different breathing states, and the image data the set contains three-dimensional images of the specified organ at the respective different respiratory states;
  • Organ motion estimation module for inputting the estimated current position of the tumor into the organ motion estimation model, to obtain the estimated current velocity vector field of the designated organ, the current velocity vector field containing the current positions of the designated organ
  • the velocity vectors corresponding to the points respectively, the organ motion estimation model is constructed by using the prior velocity vector field and the prior tumor position data set as prior knowledge, and the prior velocity vector field is determined according to the image dataset, and contains the velocity vector field of the specified organ in the respective different breathing states.
  • a third aspect of the embodiments of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the computer program
  • the method for estimating the motion of a tumor provided by the first aspect of the embodiments of the present application is implemented.
  • a fourth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the implementation of the first aspect of the embodiments of the present application is implemented Motion estimation method of tumor.
  • a fifth aspect of the embodiments of the present application provides a computer program product that, when the computer program product runs on a terminal device, enables the terminal device to execute the tumor motion estimation method described in the first aspect of the embodiments of the present application.
  • FIG. 1 is a flowchart of an embodiment of a tumor motion estimation method provided by an embodiment of the present application
  • FIG. 2 is a schematic diagram of an operation principle of the tumor motion estimation method provided by the embodiment of the present application.
  • FIG. 3 is a structural diagram of an embodiment of an apparatus for estimating tumor motion provided by an embodiment of the present application
  • FIG. 4 is a schematic diagram of a terminal device provided by an embodiment of the present application.
  • the present application provides a tumor motion estimation method, device, terminal device and storage medium, which can estimate the respiratory motion of the organ where the tumor is located in real time, and improve the accuracy of locating the tumor and its surrounding important anatomical structures.
  • the execution bodies of each method embodiment of the present application are various types of terminal devices or servers, such as mobile phones, tablet computers, notebook computers, desktop computers, wearable devices, and various medical devices.
  • FIG. 1 shows a tumor motion estimation method provided by an embodiment of the present application, including:
  • a patient refers to a patient or animal with a tumor in a designated organ (eg, liver, lung, stomach, etc.) in the body.
  • the respiration-related signal refers to any conveniently measurable signal capable of distinguishing the difference between periods of respiratory motion and within periods, and the signal reflects the motion trajectory characteristics of a designated organ with a tumor in different respiratory states.
  • an optical marker can be pasted on the surface of the chest and abdomen of the patient.
  • the optical tracking and capturing device is used to collect the motion trajectory of the optical marker as the respiratory-related signal.
  • the current value of the respiratory-related signal is The current spatial position of the optical marker.
  • the terminal device executing this embodiment of the method acquires the respiration-related signal in real time by docking with the optical tracking and capturing device.
  • the current value of the respiration-related signal is obtained, it is input into a pre-built tumor motion estimation model to obtain the estimated current position of the patient's tumor.
  • the tumor motion estimation model is constructed with a priori tumor location dataset and the respiratory-related signal as prior knowledge, and the prior tumor location dataset is determined according to a pre-collected image dataset of the designated organ, and includes the tumor
  • the image data set contains the three-dimensional images of the designated organ at the positions in different breathing states, and the construction process of the tumor motion estimation model will be described below.
  • the respiration-related signal is a motion track signal of an optical marker disposed on a specified part of the patient's body
  • the motion track signal includes the spatial position of the optical marker at each time point , before inputting the current value of the respiration-related signal into the tumor motion estimation model, it may also include:
  • the tumor motion estimation model is a machine learning model constructed in advance using a priori tumor position data set and a respiratory correlation signal as prior knowledge, wherein the prior knowledge refers to the acquired data for establishing the machine learning model.
  • the prior knowledge refers to the acquired data for establishing the machine learning model.
  • one or more optical markers can be fixed on a designated part of the patient's body (the part can be determined according to the position of the organ with the tumor in the patient's body, generally the skin surface of the chest and abdomen) by means of sticking or the like.
  • the optical tracking system is used to collect the motion trajectory signal of the optical marker as the respiratory correlation signal, which can be denoted as ⁇ M(t)
  • M (t) represents the spatial position of the optical marker recorded by the optical tracking system at time t
  • T is the data collection time.
  • each optical marker has a corresponding motion track signal, and the present application only needs to use the motion track signal of one of the optical markers during implementation.
  • the advantage of setting multiple optical markers is that each optical marker can obtain a corresponding tumor motion estimation result, and the subsequent process can select the smallest error from these tumor motion estimation results as the final result.
  • the prior tumor location dataset is a dataset that records the tumor location determined according to the pre-collected image dataset of the specified organ, and the image dataset can be a CT image dataset or an ultrasound image dataset, including the specified organ in each 3D images of different breathing states.
  • a four-dimensional CT scanner can be used to obtain N s three-dimensional CT volume data contained in the chest and abdomen during the patient's free breathing process, that is, four-dimensional CT volume data is obtained as the image data set.
  • N s represents the number of breathing states (which can be set according to the breathing cycle and phase of the patient, for example, it can take a value of 16), and different CT volume data come from different breathing cycles and phases.
  • the prior tumor location data set can be determined according to it. Specifically, the tumor location can be marked from each volume data in the image data set by manual marking or automatic identification, and then The tumor position corresponding to each volume data is extracted from the image data set, and since each volume data corresponds to a different breathing state, each extracted tumor position is also in a one-to-one correspondence with each breathing state. Therefore, the prior tumor location dataset can be denoted as ⁇ T j
  • T j ⁇ ⁇ 3 ; j 1,...,N s ⁇ , where T j represents the image coordinates of the tumor corresponding to any breathing state j Since a total of N s volume data in respiratory states are acquired before surgery, the prior tumor location dataset also includes N s tumor location data.
  • the prior tumor location dataset After the prior tumor location dataset is determined, it can be phase-matched with the motion trajectory signal to construct a location paired dataset, which includes the corresponding prior tumor location datasets in each breathing state. Tumor location and optical marker location in motion trajectory signals. While collecting the image data set, the optical tracking system is used to collect the motion track signal of the optical marker. Each volume data in the image data set has a corresponding time stamp of acquisition, and the motion track signal also has a time stamp corresponding to the time of acquisition, while Each tumor location in the prior tumor location dataset is extracted from each volume data in the image dataset, so it is only necessary to pair the time stamp of the motion track signal with the time stamp of the volume data to obtain the same Corresponding tumor location and optical marker location in respiratory state.
  • the position paired dataset can be expressed as ⁇ M(t j ), T j ⁇
  • j 1,...,N s ⁇ , where M(t j ) represents the breathing state j in the motion trajectory signal
  • M(t j ) represents the breathing state j in the motion trajectory signal
  • T j represents the tumor position corresponding to the respiratory state j in the prior tumor location data set
  • N s respiratory states there are N s respiratory states in total, that is, the position paired data set contains N s optical marker positions and tumor position pairings The data.
  • the location paired dataset After the location paired dataset is constructed, it can be used as prior knowledge to construct a corresponding tumor motion estimation model.
  • the tumor motion estimation model which may include:
  • a corresponding tumor motion estimation model is established in each preset spatial coordinate direction.
  • a tumor motion estimation model can be established in each spatial coordinate direction of the tumor, using the motion trajectory signal as a proxy signal. For example, if there are three spatial coordinate directions, a corresponding tumor motion estimation sub-model is established in each coordinate direction, that is, three tumor motion estimation sub-models are obtained. A tumor motion estimation sub-model is formed, which are respectively used for estimating the components of the tumor moving in different spatial coordinate directions.
  • a machine learning model based on ⁇ -SVR reference may be made to the prior art, and details are not described herein again.
  • the construction process of the tumor motion estimation model is completed before the operation of the patient, and during the operation of the patient, the respiratory correlation signal obtained in real time can be The current value of is input into the tumor motion estimation model, so as to obtain the estimated current position of the tumor, which can be recorded as
  • the respiration-related signal is the motion track signal of the optical marker
  • the current value of the motion track signal that is, the current position of the optical marker
  • the input of the model can also include the breathing direction.
  • j 1,...,N s ⁇
  • the model is trained to obtain model parameters.
  • D(t j ) is as follows:
  • ⁇ t is the sampling time interval of the optical tracking system, which is determined by the sampling frequency;
  • the function of ⁇ is to distinguish different breathing states such as expiratory phase, inspiratory phase, end-expiration and end-inspiration.
  • the current position is then input into a pre-built organ motion estimation model, which is used for motion estimation of the entire organ of the patient with the tumor.
  • organ motion estimation model the current velocity vector field of the designated organ can be estimated, and the current velocity vector field includes the velocity vectors corresponding to each position point of the designated organ at present.
  • the organ motion estimation model is constructed by using the prior velocity vector field and the prior tumor location data set described above as prior knowledge, wherein the prior velocity vector field can be based on the aforementioned prior knowledge.
  • the image data set is determined and contains velocity vector fields for the specified organ under various respiratory states.
  • the method before inputting the current value of the respiration-related signal into the tumor motion estimation model, the method may further include:
  • the volume data of the specified organ under different breathing states are segmented from the image data set, for example, the volume data of the liver under different breathing states can be segmented therefrom.
  • the volume data obtained by segmentation can be expressed as: ⁇ I j : ⁇
  • j 1,...,N s ⁇ , where ⁇ represents the organ area, N s is the number of respiratory states, and this formula represents the number of respiration states in the organ area.
  • one of the N s breathing states is selected as a reference state, and its corresponding volume data can be called reference volume data, which can be expressed as I 0 ⁇ ⁇ I j
  • j 1,...,N s ⁇
  • the differential homeomorphic deformation registration process is performed on the reference volume data and the volume data in all other breathing states (that is, other breathing states in the N s breathing states except the reference state), respectively, so that other breathing states can be established relative to each other.
  • the dense displacement vector field of the reference state and the corresponding velocity vector field in the Lie algebraic space are used to describe the patient-specific prior knowledge of respiratory motion.
  • This dense displacement vector field can be expressed as Its essence is the displacement of each position point in the specified organ from the reference state to each other breathing state, that is, a vector field composed of displacement vectors of a large number of points.
  • the prior velocity vector field can be expressed as It contains the velocity vector field of the specified organ in different breathing states, and the velocity vector field and the dense displacement vector field can be converted to each other by means of group exponential mapping.
  • the organ motion estimation model is constructed. For example, a dataset ⁇ T j , v j ⁇
  • j 1, . Organ motion estimation model.
  • the differential homeomorphic deformation registration process is performed on the volume data in the reference state and the volume data in other states to obtain the prior velocity vector field, which may include:
  • an initial value to the velocity vector field of the specified organ in the reference state for example, it can be assigned as 0 (that is, the initial velocity vector of each position point in the organ is 0, and it is initially in a stationary state).
  • the update of the velocity vector field corresponding to any breathing state can be calculated by adopting an alternate optimization strategy according to the volume data in the breathing state and the volume data in the reference state. value, and superimpose the updated value and the initial value to obtain the velocity vector field corresponding to any breathing state. In the same way, the velocity vector field corresponding to each breathing state in the other states can be obtained.
  • d j represents the displacement vector field corresponding to any other state
  • v j represents the velocity vector field corresponding to any other state
  • E dj represents the energy function corresponding to the displacement vector field d j
  • Id is the consistent transformation of the volume data I 0
  • the double vertical line in the formula represents the L2 norm
  • ⁇ i represents the image similarity weight
  • ⁇ x represents the spatial uncertainty weight of the transformation
  • the symbol o represents applying the transformation to the image.
  • inputting the estimated current position of the tumor into an organ motion estimation model to obtain the estimated current velocity vector field of the specified organ may include:
  • the velocity vector of the position point is used as a function of the estimated current position of the tumor, and according to the estimated current position of the tumor, the method of spatial interpolation is used to move to the position point Interpolate between the velocity vectors in the different breathing states to obtain the estimated current velocity vector of the position point, wherein the velocity vectors of the position point in the different breathing states are based on the prior velocity vector field Sure.
  • the interpolation of the velocity vector in the Lie algebra space can ensure that the differential homeomorphic deformation field is finally obtained.
  • Any spatial interpolation method can be used when performing interpolation. Taking Kriging interpolation method as an example, this method can not only consider the positional relationship between the estimated point and the observation point, but also consider the positional relationship between the observation points, so as to realize The optimal unbiased estimation of the target velocity vector can achieve the ideal interpolation effect when there are few observation points.
  • z is the velocity vector at any anatomical point of the specified organ after normalization (specifically, it can be a commonly used operation in data processing, such as the operation of subtracting the mean value from the original value and dividing by the standard deviation) in a certain coordinate direction
  • the coordinate value of that is, the value of z needs to be estimated, is the normalized tumor position under the same breathing state (that is, the estimated current tumor position described above), then z can be regarded as a combination of a regression model F and a random process e, that is e is used to describe the approximation error.
  • the regression model F can be taken as a constant ⁇ .
  • the mean of the random process e is 0 and the covariance is where ⁇ 2 is the process variance
  • It is a model describing the correlation between variables z under the tumor location corresponding to any two breathing states in the N s breathing states.
  • the model can reflect both the spatial structure characteristics of variable z and the random distribution of variable z. characteristics, the specific parameters of the model can be obtained from the preoperative observational data set It is obtained by fitting based on the least squares method.
  • the estimation of z can be performed by weighted summation of the preoperative observations (that is, the velocity vector of the position point in different breathing states, which can be determined according to the prior velocity vector field). get, as the following formula:
  • F is a column vector whose elements are all 1s, is the correlation matrix of the variable z in the N s respiratory states obtained before surgery, is the correlation vector between the real-time estimated tumor location and the variable z in the N s respiratory states obtained preoperatively.
  • the correlation vector r can be calculated according to the real-time estimated tumor position obtained during the operation, so that z can be calculated by using this formula group. , complete the estimation of the target.
  • each anatomical point in the specified organ can be estimated to obtain the current corresponding velocity vector in the same way, so as to obtain the current velocity vector field of the specified organ, that is, to obtain the current anatomical points of the specified organ respectively.
  • Corresponding velocity vector can be estimated to obtain the current corresponding velocity vector in the same way, so as to obtain the current velocity vector field of the specified organ, that is, to obtain the current anatomical points of the specified organ respectively.
  • the method may further include:
  • a preset Gaussian kernel function is used to normalize the current velocity vector field, and the current dense displacement field of the designated organ is obtained by means of group exponential transformation, and the current dense displacement field includes each current displacement field of the designated organ.
  • the displacement vectors corresponding to the position points respectively.
  • a preset Gaussian kernel function can be used to regularize the current velocity vector field, and then the final velocity vector field can be obtained through group exponential mapping transformation.
  • the reconstruction of the motion shape of the organ in the new breathing state can finally be realized based on the dense deformation displacement vector field.
  • the embodiment of the present application proposes a hierarchical estimation framework from tumor to whole organ motion under free breathing motion, the framework includes a tumor motion estimation model and an organ motion estimation model, wherein the tumor motion estimation model is based on a prior tumor position data set
  • the respiration-related signal is constructed as prior knowledge
  • the organ motion estimation model is constructed with prior velocity vector field and prior tumor location dataset as prior knowledge.
  • the patient's respiration-related signal can be input into the tumor motion estimation model as a proxy signal to obtain an estimated tumor position; then, the estimated tumor position can be input into the organ motion estimation model , to obtain the estimated velocity vector field of the whole organ, thereby realizing the estimation of the respiratory motion of the whole organ, and improving the accuracy of locating the tumor and its surrounding important anatomical structures.
  • FIG. 2 it is a schematic diagram of an operation principle of the tumor motion estimation method proposed in the present application.
  • the motion estimation method can be divided into two stages: preoperative and intraoperative:
  • a four-dimensional CT image data set of the patient is obtained, and the image data set contains the corresponding three-dimensional CT images of the designated organ of the patient under different breathing states; on the one hand, the designated organ is segmented from the image data set.
  • Volume data organ mask
  • the tumor position extracted from the image data set that is, the prior tumor position data set
  • the motion trajectory signals of the pre-collected optical markers are obtained after phase matching.
  • a tumor motion estimation model is obtained by training; based on the prior velocity vector field and the prior tumor position data Set, train to obtain an organ motion estimation model.
  • the motion signal of the optical marker on the body surface is detected in real time, and the current optical marker position is input into the preoperatively constructed tumor motion estimation model to obtain the estimated tumor position; then, the estimated tumor position is input into the preoperatively constructed organ
  • the motion estimation model estimates the breathing motion of the entire organ to obtain the corresponding dense velocity vector field; finally, by performing group exponential mapping processing on the dense velocity vector field, the corresponding differential homeomorphic dense deformation field (ie, the displacement vector field) can be obtained. ), and then reconstruct the motion shape of the organ in real time to achieve hierarchical motion estimation.
  • the hierarchical estimation framework from the tumor to the whole organ under free breathing motion proposed in this application can achieve a higher accuracy estimation of the tumor on the basis of realizing the accurate motion estimation of the whole organ;
  • the estimation of the respiratory motion of the whole organ does not require iterative optimization calculation, so it can have better real-time processing; Global one-to-one, differentiable and reversible advantages.
  • a method for estimating the motion of a tumor is mainly described above, and an apparatus for estimating a motion of a tumor will be described below.
  • an embodiment of a tumor motion estimation apparatus in the embodiment of the present application includes:
  • Respiration-related signal acquisition module 301 configured to acquire a current value of a patient's respiration-related signal, where the respiration-related signal reflects the motion trajectory characteristics of a designated organ with a tumor in the patient under different breathing states;
  • a tumor motion estimation module 302 configured to input the current value of the respiration-related signal into a tumor motion estimation model to obtain an estimated current position of the tumor, the tumor motion estimation model using a priori tumor position data set and the respiration
  • the correlation signal is constructed as a priori knowledge, and the prior tumor location data set is determined according to the pre-collected image data set of the specified organ, and includes the location of the tumor in each of the different breathing states, and the image the data set includes three-dimensional images of the specified organ in the respective different respiratory states;
  • the organ motion estimation module 303 is configured to input the estimated current position of the tumor into the organ motion estimation model, and obtain the estimated current velocity vector field of the specified organ, where the current velocity vector field includes each of the current specified organs.
  • the velocity vectors corresponding to the position points respectively, the organ motion estimation model is constructed with a priori velocity vector field and the prior tumor location dataset as prior knowledge, and the prior velocity vector field is determined according to the image dataset , and contains the velocity vector fields of the specified organ at the respective different respiration states.
  • the respiration-related signal is a motion track signal of an optical marker disposed on a specified part of the patient's body
  • the motion track signal includes the spatial position of the optical marker at each time point
  • the motion estimation apparatus may further include:
  • a paired data set construction module configured to construct a position paired data set according to the prior tumor position data set and the motion trajectory signal, and the position paired data set includes the prior corresponding to each of the breathing states The tumor location in the tumor location dataset and the optical marker location in the motion trajectory signal;
  • the tumor motion estimation model building module is configured to use the position paired data set as prior knowledge to construct and obtain the tumor motion estimation model.
  • the tumor motion estimation model building module can be specifically used for: based on the ⁇ -SVR machine learning model and the position paired data set, using the motion trajectory signal as a proxy signal, in each preset spatial coordinate direction, respectively. A corresponding tumor motion estimation model was established.
  • the motion estimation apparatus may further include:
  • volume data acquisition module configured to segment the volume data of the designated organ under the different breathing states from the image data set
  • a differential homeomorphic deformation registration module used to select a breathing state from the different breathing states as a reference state, and perform differential homeomorphic deformation registration on the volume data in the reference state and the volume data in other states processing to obtain the prior velocity vector field, and the other states are other breathing states except the reference state in the different breathing states;
  • the organ motion estimation model building module is configured to use the prior velocity vector field and the prior tumor position data set as prior knowledge to construct the organ motion estimation model.
  • differential homeomorphic deformation registration module may include:
  • a velocity vector field assignment unit configured to assign a preset initial value to the velocity vector field of the designated organ in the reference state
  • the velocity vector field update unit is used for each breathing state in the other states, according to the volume data in the breathing state and the volume data in the reference state, using an alternate optimization strategy to calculate the corresponding breathing state.
  • the updated value of the velocity vector field, and the updated value and the initial value are added to obtain the velocity vector field of the specified organ in the breathing state.
  • the organ motion estimation module may include:
  • a spatial interpolation unit for each position point of the specified organ, using the velocity vector of the position point as a function of the estimated current position of the tumor, and using spatial interpolation according to the estimated current position of the tumor.
  • the motion estimation apparatus may further include:
  • a group exponential transformation module configured to use a preset Gaussian kernel function to regularize the current velocity vector field, and obtain the current dense displacement field of the specified organ by means of group exponential transformation, where the current dense displacement field includes The displacement vector corresponding to each position point of the specified organ at present.
  • Embodiments of the present application further provide a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements any tumor motion estimation method as shown in FIG. 1 .
  • Embodiments of the present application further provide a computer program product, which, when the computer program product runs on a terminal device, enables the terminal device to execute any tumor motion estimation method as shown in FIG. 1 .
  • FIG. 4 is a schematic diagram of a terminal device provided by an embodiment of the present application.
  • the terminal device 4 of this embodiment includes: a processor 40 , a memory 41 , and a computer program 42 stored in the memory 41 and running on the processor 40 .
  • the processor 40 executes the computer program 42
  • the steps in the above embodiments of each tumor motion estimation method are implemented, for example, steps 101 to 103 shown in FIG. 1 .
  • the processor 40 executes the computer program 42
  • the functions of the modules/units in each of the foregoing apparatus embodiments such as the functions of the modules 301 to 303 shown in FIG. 3, are implemented.
  • the computer program 42 may be divided into one or more modules/units, which are stored in the memory 41 and executed by the processor 40 to complete the present application.
  • the one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 42 in the terminal device 4 .
  • the so-called processor 40 may be a central processing unit (Central Processing Unit, CPU), or other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
  • the memory 41 may be an internal storage unit of the terminal device 4 , such as a hard disk or a memory of the terminal device 4 .
  • the memory 41 may also be an external storage device of the terminal device 4, such as a plug-in hard disk equipped on the terminal device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) Card, Flash Card, etc.
  • the memory 41 may also include both an internal storage unit of the terminal device 4 and an external storage device.
  • the memory 41 is used for storing the computer program and other programs and data required by the terminal device.
  • the memory 41 can also be used to temporarily store data that has been output or will be output.
  • the disclosed apparatus and method may be implemented in other manners.
  • the system embodiments described above are only illustrative.
  • the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods.
  • multiple units or components may be Combinations can either be integrated into another system, or some features can be omitted, or not implemented.
  • the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit.
  • the above-mentioned integrated units may be implemented in the form of hardware, or may be implemented in the form of software functional units.
  • the integrated unit if implemented in the form of a software functional unit and sold or used as an independent product, may be stored in a computer-readable storage medium.
  • the present application can implement all or part of the processes in the methods of the above embodiments, and can also be completed by instructing the relevant hardware through a computer program.
  • the computer program can be stored in a computer-readable storage medium, and the computer When the program is executed by the processor, the steps of the foregoing method embodiments can be implemented.
  • the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form, and the like.
  • the computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, removable hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory) , Random Access Memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.
  • ROM Read-Only Memory
  • RAM Random Access Memory
  • electric carrier signal telecommunication signal and software distribution medium, etc.
  • the content contained in the computer-readable media may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer-readable media Excluded are electrical carrier signals and telecommunication signals.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Biomedical Technology (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Data Mining & Analysis (AREA)
  • Pathology (AREA)
  • Databases & Information Systems (AREA)
  • Radiology & Medical Imaging (AREA)
  • Multimedia (AREA)
  • Quality & Reliability (AREA)
  • Software Systems (AREA)
  • Urology & Nephrology (AREA)
  • Surgery (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Business, Economics & Management (AREA)
  • General Business, Economics & Management (AREA)
  • Image Analysis (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

本申请涉及图像处理技术领域，提出一种肿瘤的运动估计方法、装置、终端设备和存储介质。该运动估计方法包括：获取患者的呼吸关联信号的当前值，所述呼吸关联信号反映所述患者带有肿瘤的指定器官在各个不同呼吸状态下的运动轨迹特征；将所述呼吸关联信号的当前值输入肿瘤运动估计模型，得到估计的所述肿瘤的当前位置；将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场。采用该运动估计方法，能够实现对整个器官的呼吸运动估计，提高对肿瘤及其周围重要解剖结构进行定位的准确性。

Description

一种肿瘤的运动估计方法、装置、终端设备和存储介质 技术领域
本申请涉及图像处理技术领域，尤其涉及一种肿瘤的运动估计方法、装置、终端设备和存储介质。
背景技术
经皮医学影像引导介入手术是癌症诊断和治疗的常用手段，但在实际手术过程中，患者生理呼吸导致的肿瘤及器官运动增加了在不损伤肿瘤周围血管等重要结构的同时将穿刺针精确地插入到肿瘤位置的难度。
针对上述问题，现有技术通常会建立一个呼吸运动模型来实时估计目标的呼吸运动。该方法主要利用容易获得的代理信号和内部目标运动之间的强相关性建立二者之间的关联模型，在手术过程中通过实时检测代理信号来估计目标运动。然而，由于建模的复杂性，该方法通常只能模拟肿瘤或者单个解剖特征点的呼吸运动，不适用于整个器官，无法实现对肿瘤及其周围重要解剖结构的准确定位。
发明内容
有鉴于此，本申请实施例提供了一种肿瘤的运动估计方法、装置、终端设备和存储介质，能够实时估计肿瘤所处器官的呼吸运动，提高对肿瘤及其周围重要解剖结构进行定位的准确性。
本申请实施例的第一方面提供了一种肿瘤的运动估计方法，包括：
获取患者的呼吸关联信号的当前值，所述呼吸关联信号反映所述患者带有肿瘤的指定器官在各个不同呼吸状态下的运动轨迹特征；
将所述呼吸关联信号的当前值输入肿瘤运动估计模型，得到估计的所述肿 瘤的当前位置，所述肿瘤运动估计模型以先验肿瘤位置数据集和所述呼吸关联信号作为先验知识构建得到，所述先验肿瘤位置数据集根据预采集的所述指定器官的图像数据集确定，且包含所述肿瘤在所述各个不同呼吸状态下的位置，所述图像数据集包含所述指定器官在所述各个不同呼吸状态下的三维图像；
将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，所述当前速度向量场包含当前所述指定器官的各个位置点分别对应的速度向量，所述器官运动估计模型以先验速度向量场和所述先验肿瘤位置数据集作为先验知识构建得到，所述先验速度向量场根据所述图像数据集确定，且包含所述指定器官在所述各个不同呼吸状态下的速度向量场。
本申请实施例提出一个自由呼吸运动下从肿瘤到整个器官运动的分级估计框架，该框架包含一个肿瘤运动估计模型和一个器官运动估计模型，其中，该肿瘤运动估计模型以先验肿瘤位置数据集和呼吸关联信号作为先验知识构建得到，该器官运动估计模型以先验速度向量场和先验肿瘤位置数据集作为先验知识构建得到。在对肿瘤患者进行手术的过程中，可以先将该患者的呼吸关联信号作为代理信号输入该肿瘤运动估计模型，得到估计的肿瘤位置；然后，再将该估计的肿瘤位置输入该器官运动估计模型，得到估计的整个器官的速度向量场，从而实现对整个器官的呼吸运动估计，提高对肿瘤及其周围重要解剖结构进行定位的准确性。
在本申请的一个实施例中，所述呼吸关联信号为设置于所述患者的身体指定部位的光学标记的运动轨迹信号，所述运动轨迹信号包含所述光学标记在各个时间点下的空间位置，在将所述呼吸关联信号的当前值输入肿瘤运动估计模型之前，还可以包括：
根据所述先验肿瘤位置数据集和所述运动轨迹信号构建位置配对数据集，所述位置配对数据集包含每个所述呼吸状态下分别对应的所述先验肿瘤位置数据集中的肿瘤位置和所述运动轨迹信号中的光学标记位置；
以所述位置配对数据集作为先验知识，构建得到所述肿瘤运动估计模型。
进一步的，以所述位置配对数据集作为先验知识，构建得到所述肿瘤运动估计模型，可以包括：
基于ε-SVR机器学习模型和所述位置配对数据集，以所述运动轨迹信号作为代理信号，在各个预设空间坐标方向上分别建立对应的肿瘤运动估计模型。
在本申请的一个实施例中，在将所述呼吸关联信号的当前值输入肿瘤运动估计模型之前，还可以包括：
从所述图像数据集中分割出所述指定器官在所述各个不同呼吸状态下的体数据；
从所述各个不同呼吸状态中选取一个呼吸状态作为参考状态，并对所述参考状态下的体数据和其它状态下的体数据执行微分同胚变形配准处理，得到所述先验速度向量场，所述其它状态为所述各个不同呼吸状态中除所述参考状态外的其它呼吸状态；
以所述先验速度向量场和所述先验肿瘤位置数据集作为先验知识，构建得到所述器官运动估计模型。
进一步的，对所述参考状态下的体数据和其它状态下的体数据执行微分同胚变形配准处理，得到所述先验速度向量场，可以包括：
为所述指定器官在所述参考状态下的速度向量场赋予预设的初始值；
针对所述其它状态中的每个呼吸状态，根据该呼吸状态下的体数据以及所述参考状态下的体数据，采用交替优化的策略计算得到该呼吸状态对应的速度向量场的更新值，并将所述更新值与所述初始值相加，得到所述指定器官在该呼吸状态下的速度向量场。
在本申请的一个实施例中，将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，可以包括：
针对所述指定器官的每个位置点，将该位置点的速度向量作为估计的所述肿瘤的当前位置的函数，并根据估计的所述肿瘤的当前位置，采用空间插值的方法往该位置点在所述各个不同呼吸状态下的速度向量之间插值，得到估计的 该位置点的当前速度向量，其中，该位置点在所述各个不同呼吸状态下的速度向量根据所述先验速度向量场确定。
在本申请的一个实施例中，在得到估计的所述指定器官的当前速度向量场之后，还可以包括：
采用预设的高斯核函数对所述当前速度向量场进行正则化，并通过群指数变换的方式获得所述指定器官的当前稠密位移场，所述当前稠密位移场包含当前所述指定器官的各个位置点分别对应的位移向量。
本申请实施例的第二方面提供了一种肿瘤的运动估计装置，包括：
呼吸关联信号获取模块，用于获取患者的呼吸关联信号的当前值，所述呼吸关联信号反映所述患者带有肿瘤的指定器官在各个不同呼吸状态下的运动轨迹特征；
肿瘤运动估计模块，用于将所述呼吸关联信号的当前值输入肿瘤运动估计模型，得到估计的所述肿瘤的当前位置，所述肿瘤运动估计模型以先验肿瘤位置数据集和所述呼吸关联信号作为先验知识构建得到，所述先验肿瘤位置数据集根据预采集的所述指定器官的图像数据集确定，且包含所述肿瘤在所述各个不同呼吸状态下的位置，所述图像数据集包含所述指定器官在所述各个不同呼吸状态下的三维图像；
器官运动估计模块，用于将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，所述当前速度向量场包含当前所述指定器官的各个位置点分别对应的速度向量，所述器官运动估计模型以先验速度向量场和所述先验肿瘤位置数据集作为先验知识构建得到，所述先验速度向量场根据所述图像数据集确定，且包含所述指定器官在所述各个不同呼吸状态下的速度向量场。
本申请实施例的第三方面提供了一种终端设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现如本申请实施例的第一方面提供的肿瘤的运动估计方 法。
本申请实施例的第四方面提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行时实现如本申请实施例的第一方面提供的肿瘤的运动估计方法。
本申请实施例的第五方面提供了一种计算机程序产品，当计算机程序产品在终端设备上运行时，使得终端设备执行本申请实施例的第一方面所述的肿瘤的运动估计方法。
可以理解的是，上述第二方面至第五方面的有益效果可以参见上述第一方面中的相关描述，在此不再赘述。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例提供的肿瘤的运动估计方法的一个实施例的流程图；
图2是本申请实施例提供的肿瘤的运动估计方法的一种操作原理示意图；
图3是本申请实施例提供的肿瘤的运动估计装置的一个实施例的结构图；
图4是本申请实施例提供的一种终端设备的示意图。
具体实施方式
以下描述中，为了说明而不是为了限定，提出了诸如特定系统结构、技术之类的具体细节，以便透彻理解本申请实施例。然而，本领域的技术人员应当清楚，在没有这些具体细节的其它实施例中也可以实现本申请。在其它情况中，省略对众所周知的系统、装置、电路以及方法的详细说明，以免不必要的细节妨碍本申请的描述。另外，在本申请说明书和所附权利要求书的描述中，术语 “第一”、“第二”、“第三”等仅用于区分描述，而不能理解为指示或暗示相对重要性。
本申请提出一种肿瘤的运动估计方法、装置、终端设备和存储介质，能够实时估计肿瘤所处器官的呼吸运动，提高对肿瘤及其周围重要解剖结构进行定位的准确性。应当理解，本申请各个方法实施例的执行主体为各种类型的终端设备或服务器，比如手机、平板电脑、笔记本电脑、台式电脑、可穿戴设备和各类医疗设备等。
请参阅图1，示出了本申请实施例提供的一种肿瘤的运动估计方法，包括：
101、获取患者的呼吸关联信号的当前值；
在本申请实施例中，患者是指体内的指定器官(例如肝、肺、胃等)带有肿瘤的病人或者动物。呼吸关联信号是指任何方便测量的能够区分呼吸运动周期间以及周期内差异性的信号，该信号反映患者带有肿瘤的指定器官在各个不同呼吸状态下的运动轨迹特征。例如，可以在患者的胸腹体表粘贴光学标记，在患者的呼吸运动过程中，采用光学跟踪捕捉设备采集该光学标记的运动轨迹作为呼吸关联信号，此时该呼吸关联信号的当前值即为该光学标记当前所处的空间位置。与此同时，执行本方法实施例的终端设备通过对接该光学跟踪捕捉设备以实时获取该呼吸关联信号。
102、将所述呼吸关联信号的当前值输入肿瘤运动估计模型，得到估计的肿瘤的当前位置；
在获取到呼吸关联信号的当前值后，将其输入一个预先构建的肿瘤运动估计模型，从而得到估计的患者肿瘤的当前位置。其中，该肿瘤运动估计模型以先验肿瘤位置数据集和该呼吸关联信号作为先验知识构建得到，该先验肿瘤位置数据集根据预采集的该指定器官的图像数据集确定，且包含该肿瘤在各个不同呼吸状态下的位置，该图像数据集包含该指定器官在各个不同呼吸状态下的三维图像，以下先说明该肿瘤运动估计模型的构建过程。
在本申请的一个实施例中，所述呼吸关联信号为设置于所述患者的身体指 定部位的光学标记的运动轨迹信号，所述运动轨迹信号包含所述光学标记在各个时间点下的空间位置，在将所述呼吸关联信号的当前值输入肿瘤运动估计模型之前，还可以包括：
(1)根据所述先验肿瘤位置数据集和所述运动轨迹信号构建位置配对数据集，所述位置配对数据集包含每个所述呼吸状态下分别对应的所述先验肿瘤位置数据集中的肿瘤位置和所述运动轨迹信号中的光学标记位置；
(2)以所述位置配对数据集作为先验知识，构建得到所述肿瘤运动估计模型。
该肿瘤运动估计模型是以先验肿瘤位置数据集和呼吸关联信号作为先验知识预先构建得到的机器学习模型，其中先验知识是指获得的用于建立机器学习模型的数据。特别的，可以在患者的身体指定部位(该部位可以根据带有肿瘤的器官在患者体内的位置确定，一般为胸腹部的皮肤表面)采用粘贴等方式固定一个或多个光学标记，在对患者进行手术之前，采用光学跟踪系统采集该光学标记的运动轨迹信号作为呼吸关联信号，可以记为{M(t)|M(t)∈□ 3,t∈[-T,0]}，其中M(t)表示t时刻光学跟踪系统记录的该光学标记的空间位置，T为数据采集时长。显然，每个光学标记都有对应的一个运动轨迹信号，而本申请在实施时只需采用其中一个光学标记的运动轨迹信号即可。设置多个光学标记的优点为：每个光学标记都能获得对应的一个肿瘤运动估计结果，后续过程可以从这些肿瘤运动估计结果中挑选出误差最小的作为最终结果。
先验肿瘤位置数据集是根据预采集的该指定器官的图像数据集确定的记录肿瘤位置的数据集合，而该图像数据集可以是CT图像数据集或者超声图像数据集，包含该指定器官在各个不同呼吸状态下的三维图像。例如，在对患者进行手术之前，可以采用四维的CT扫描器获得患者自由呼吸过程中胸腹部包含的N s个三维的CT体数据，也即得到四维的CT体数据作为该图像数据集。其中，N s表示呼吸状态的数量(可以根据患者的呼吸周期和相位设定，例如可以取值16)，不同的CT体数据来自于不同的呼吸周期和相位。在获得该图像数 据集之后，可以根据其确定先验肿瘤位置数据集，具体的，可以通过人工标记或者自动识别的方式从该图像数据集中的每个体数据中标记出肿瘤的位置，然后即可从该图像数据集中分别提取出每个体数据对应的肿瘤位置，而由于每个体数据分别对应不同的呼吸状态，故提取出的各个肿瘤位置也是和各个呼吸状态一一对应的。因此，该先验肿瘤位置数据集可以记为{T j|T j∈□ 3；j＝1,...,N s}，其中，T j表示任意一个呼吸状态j对应的肿瘤在图像坐标系下的位置，由于术前总共获取N s个呼吸状态下的体数据，故该先验肿瘤位置数据集也包含N s个肿瘤位置数据。
在确定先验肿瘤位置数据集之后，可以将其和该运动轨迹信号进行相位匹配，构建出位置配对数据集，该位置配对数据集包含每个呼吸状态下分别对应的先验肿瘤位置数据集中的肿瘤位置和运动轨迹信号中的光学标记位置。在采集该图像数据集的同时采用光学跟踪系统采集光学标记的运动轨迹信号，该图像数据集中的每个体数据都有对应的采集时间戳，该运动轨迹信号同样有采集时对应的时间戳，而该先验肿瘤位置数据集中的各个肿瘤位置又是从该图像数据集的各个体数据中提取得到的，故只需要将运动轨迹信号的时间戳和体数据的时间戳进行配对，就可以获得同一呼吸状态下，对应的肿瘤位置和光学标记位置。因此，该位置配对数据集可以表示为{{M(t j),T j}|j＝1,...,N s}，其中，M(t j)表示该运动轨迹信号中呼吸状态j对应的光学标记位置，T j表示该先验肿瘤位置数据集中呼吸状态j对应的肿瘤位置，总共有N s个呼吸状态，也即该位置配对数据集中包含N s个光学标记位置和肿瘤位置配对的数据。
在构建出该位置配对数据集之后，可以将其作为先验知识，构建得到对应的肿瘤运动估计模型。
进一步的，以所述位置配对数据集作为先验知识，构建得到所述肿瘤运动估计模型，可以包括：
基于ε-SVR机器学习模型和所述位置配对数据集，以所述运动轨迹信号作为代理信号，在各个预设空间坐标方向上分别建立对应的肿瘤运动估计模型。
在实际操作中，可以基于ε-SVR机器学习模型和该位置配对数据集，在肿瘤的各个空间坐标方向上分别建立以该运动轨迹信号为代理信号的肿瘤运动估计模型。例如，若有三个空间坐标方向，则分别在每个坐标方向上建立对应的肿瘤运动估计子模型，即得到3个肿瘤运动估计子模型，本申请实施例中的肿瘤运动估计模型可以由该3个肿瘤运动估计子模型构成，分别用于估计肿瘤往各个不同空间坐标方向进行运动的分量。另外，基于ε-SVR构建机器学习模型的具体方式，可以参照现有技术，在此不再赘述。
以上所述为肿瘤运动估计模型的构建过程，一般来说，在对患者进行手术之前完成该肿瘤运动估计模型的构建，而在对患者进行手术的过程中，即可将实时获取的呼吸关联信号的当前值输入该肿瘤运动估计模型，从而得到估计的肿瘤的当前位置，可以记为
Figure PCTCN2021138010-appb-000001
假设该呼吸关联信号为光学标记的运动轨迹信号，则此时将该运动轨迹信号的当前值，也即该光学标记的当前位置作为该肿瘤运动估计模型的输入。具体的，模型的输入还可以包括呼吸方向，假设该光学标记的当前位置为M(t j)，当前的呼吸方向为D(t j)，则该模型的输入特征向量可以表示为Y j＝(M(t j) T,D(t j)) T∈□ 4，记y j为肿瘤T j某一个坐标方向下的坐标值，可以构建数据集{{Y j,y j}|j＝1,...,N s}对该模型进行训练得到模型参数。D(t j)的表达式如下：
Figure PCTCN2021138010-appb-000002
其中，
Figure PCTCN2021138010-appb-000003
是标记点在t j时刻的主运动分量；Δt是光学跟踪系统采样的时间间隔，由采样频率决定；δ是判别阈值，可以根据实际采集的光学标记的运动轨迹信号特征确定其取值。δ的作用是区分呼气相、吸气相、呼气末和吸气末等各个不同的呼吸状态，具体的取值可以根据患者的实际呼吸运动信号确定，例如可以取值为0.05。
103、将估计的肿瘤的当前位置输入器官运动估计模型，得到估计的指定器官的当前速度向量场。
在通过该肿瘤运动估计模型获得估计的肿瘤的当前位置之后，接下来要将该当前位置输入一个预先构建的器官运动估计模型，该模型用于患者带有肿瘤的整个器官的运动估计。通过该器官运动估计模型，能够估计得到该指定器官的当前速度向量场，该当前速度向量场包含当前该指定器官的各个位置点分别对应的速度向量。
以下说明该器官运动估计模型的构建方法：该器官运动估计模型以先验速度向量场和前文所述的先验肿瘤位置数据集作为先验知识构建得到，其中该先验速度向量场可以根据前文所述的图像数据集确定，且包含该指定器官在各个不同呼吸状态下的速度向量场。
在本申请的一个实施例中，在将所述呼吸关联信号的当前值输入肿瘤运动估计模型之前，还可以包括：
(1)从所述图像数据集中分割出所述指定器官在所述各个不同呼吸状态下的体数据；
(2)从所述各个不同呼吸状态中选取一个呼吸状态作为参考状态，并对所述参考状态下的体数据和其它状态下的体数据执行微分同胚变形配准处理，得到所述先验速度向量场，所述其它状态为所述各个不同呼吸状态中除所述参考状态外的其它呼吸状态；
(3)以所述先验速度向量场和所述先验肿瘤位置数据集作为先验知识，构建得到所述器官运动估计模型。
首先，从该图像数据集中分割出该指定器官在各个不同呼吸状态下的体数据，例如，可以从中分割出各个不同呼吸状态下肝脏的体数据。分割得到的体数据可以表示为：{I j:Ω→□|j＝1,...,N s}，其中Ω表示器官区域，N s为呼吸状态的数量，该式表示器官区域中的像素位置到实数空间像素值的映射。然后，从这N s个呼吸状态中选取一个作为参考状态，其对应的体数据可以称作参考体数据， 可以表示为I 0∈{I j|j＝1,...,N s}，将该参考体数据与所有的其他呼吸状态(即该N s个呼吸状态中除参考状态外的其它呼吸状态)下的体数据分别执行微分同胚变形配准处理，可以建立其它呼吸状态相对于该参考状态的稠密位移向量场以及对应的李代数空间中的速度向量场(即本申请中的先验速度向量场)，以此描述患者特异的呼吸运动先验知识。该稠密位移向量场可以表示为
Figure PCTCN2021138010-appb-000004
其本质为从参考状态到各个其它呼吸状态时该指定器官中各个位置点的位移，也即由大量点的位移向量构成的向量场。该先验速度向量场可以表示为
Figure PCTCN2021138010-appb-000005
其包含该指定器官在各个不同呼吸状态下的速度向量场，速度向量场和稠密位移向量场之间可以通过群指数映射的方式相互转换。最后，以该先验速度向量场和该先验肿瘤位置数据集作为先验知识，构建得到该器官运动估计模型。例如，可以构建数据集{{T j,v j}|j＝1,...,N s}，其中T j表示肿瘤位置，v j表示先验速度向量场，并基于该数据集建立该器官运动估计模型。
具体的，对参考状态下的体数据和其它状态下的体数据执行微分同胚变形配准处理，得到所述先验速度向量场，可以包括：
(1)为所述指定器官在所述参考状态下的速度向量场赋予预设的初始值；
(2)针对所述其它状态中的每个呼吸状态，根据该呼吸状态下的体数据以及所述参考状态下的体数据，采用交替优化的策略计算得到该呼吸状态对应的速度向量场的更新值，并将所述更新值与所述初始值相加，得到所述指定器官在该呼吸状态下的速度向量场。
首先，为该指定器官在参考状态下的速度向量场赋予一个初始值，例如可以赋值为0(即器官中每个位置点的初始速度向量都是0，初始处于静止状态)。然后，针对该其它状态中的任意一个呼吸状态，可以根据该呼吸状态下的体数据和该参考状态下的体数据，采用交替优化的策略计算得到该任意一个呼吸状态对应的速度向量场的更新值，并将该更新值和该初始值叠加，从而得到该任 意一个呼吸状态对应的速度向量场。采用相同的方式，可以获得该其它状态中每个呼吸状态分别对应的速度向量场。
例如，给定两个体数据I 0(参考状态下的体数据)和I j(任意一个其它状态下的体数据)，可以采用Demons交替优化策略，先在当前变换d j＝exp(v j)的基础上通过在李代数空间上优化以下公式得到速度向量场的更新值u，以此来更新当前变换v j←v j+u，即用u更新v j：
Figure PCTCN2021138010-appb-000006
其中，d j表示任意一个其它状态对应的位移向量场，v j表示任意一个其它状态对应的速度向量场，E dj表示位移向量场d j对应的能量函数，Id为体数据I 0的一致变换，公式中的双竖线表示L2范数，σ i表示图像相似性权重，σ x表示变换的空间不确定性权重，符号o表示将变换作用到图像上。
接着，通过该变换与一个高斯核进行卷积运算实现正则化，可以得到该任意一个其它状态对应的李代数空间中的速度向量场，再经过群指数映射变换后可以得到对应的微分同胚稠密位移向量场d j＝exp(v j)。
在本申请的一个实施例中，将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，可以包括：
针对所述指定器官的每个位置点，将该位置点的速度向量作为估计的所述肿瘤的当前位置的函数，并根据估计的所述肿瘤的当前位置，采用空间插值的方法往该位置点在所述各个不同呼吸状态下的速度向量之间插值，得到估计的该位置点的当前速度向量，其中，该位置点在所述各个不同呼吸状态下的速度向量根据所述先验速度向量场确定。
在获得对数域的先验速度向量场{v j:Ω→□ 3|j＝1,...,N s}、先验肿瘤位置数据集{T j|T j∈□ 3；j＝1,...,N s}以及估计的肿瘤的当前位置
Figure PCTCN2021138010-appb-000007
的基础上，通过将该指定器官中任意位置点(解剖点)处的速度向量看作是相同呼吸状态下(当前时刻) 肿瘤位置的函数，那么当前该指定器官中任意位置点处的速度向量，便可以根据术前获得的该位置点在各个不同呼吸状态下的速度向量之间插值获得。
具体的，由于李代数空间具有双线性特性，在李代数空间中对速度向量进行插值能够保证最终获得微分同胚变形场。在进行插值时可以采用任意的空间插值方法，以Kriging插值方法为例，该方法不仅能够考虑被估计点跟观测点之间的位置关系，而且能够考虑各观测点之间的位置关系，从而实现对目标速度向量的最优无偏估计，在观测点较少的时候能够取得理想的插值效果。
由于肝脏等器官的呼吸运动具有各向异性，故本申请将对李代数空间中速度向量三个坐标方向上的分量分别进行估计。假设z为经标准化(具体可以是数据处理中常用的操作，例如将原数值减去均值后除以标准差的操作)后的指定器官的任意一个解剖点处的速度向量在某一个坐标方向上的坐标值，即需要估计得到该z的数值，
Figure PCTCN2021138010-appb-000008
为标准化后的相同呼吸状态下的肿瘤位置(即前文所述的估计的肿瘤当前位置)，那么可以将z看作一个回归模型F和一个随机过程e的组合，即
Figure PCTCN2021138010-appb-000009
e用于描述近似误差。为了简化运算，可以将回归模型F取为一个常数β。另外，可以假设随机过程e的均值为0且协方差为
Figure PCTCN2021138010-appb-000010
其中σ 2为过程方差，
Figure PCTCN2021138010-appb-000011
是描述该N s个呼吸状态中任意的两个呼吸状态对应的肿瘤位置下变量z之间的关联性的模型，该模型既可以反映变量z的空间结构特性，又可以反映变量z的随机分布特性，该模型的具体参数可以通过在术前的观测数据集
Figure PCTCN2021138010-appb-000012
上基于最小二乘法等方式拟合获得。
在对肿瘤位置进行实时估计的基础上，z的估计可以通过对术前观测值(即该位置点在各个不同呼吸状态下的速度向量，可以根据该先验速度向量场确定)进行加权求和得到，如以下公式：
z＝c TZ
其中，
Figure PCTCN2021138010-appb-000013
为术前获得的该位置点在N s个呼吸状态下分别对应的 速度向量的坐标值，
Figure PCTCN2021138010-appb-000014
是满足无偏估计且使均方误差最小的权重系数向量。
通过对公式z＝c TZ进行推导，可以得到以下公式组：
z＝a+r T(Z-aF)
a＝(F TR -1F) -1F TR -1Z
其中，F为元素全为1的列向量，
Figure PCTCN2021138010-appb-000015
是术前获得的N s个呼吸状态下变量z的关联矩阵，
Figure PCTCN2021138010-appb-000016
是实时估计的肿瘤位置与术前获得的N s个呼吸状态下变量z的关联向量。在该公式组中，除了关联向量r之外，其余的量都可以在术前确定，根据术中获取的实时估计的肿瘤位置可以计算得到该关联向量r，从而可以采用该公式组计算得到z，完成对目标的估计。
另外，该任意解剖点处的速度向量在别的坐标方向上的坐标值x和y的估计，可以参照上述对z进行估计的方法实现，从而获得该任意解剖点处当前的速度向量(x，y，z)。以此类推，该指定器官中的每个解剖点都可以采用相同的方式估计得到当前对应的速度向量，从而获得该指定器官的当前速度向量场，也即获得当前该指定器官的各个解剖点分别对应的速度向量。
在本申请的一个实施例中，在得到估计的所述指定器官的当前速度向量场之后，还可以包括：
采用预设的高斯核函数对所述当前速度向量场进行正则化，并通过群指数变换的方式获得所述指定器官的当前稠密位移场，所述当前稠密位移场包含当前所述指定器官的各个位置点分别对应的位移向量。
在得到整个器官的当前速度向量场的基础上，为了保证最终得到变换的光滑性，可以采用一个预设的高斯核函数对该当前速度向量场进行正则化，然后通过群指数映射变换得到最终的相对于参考状态的稠密变形位移向量场的估计，最后可以基于该稠密变形位移向量场实现新的呼吸状态下器官运动形态的重构。
本申请实施例提出一个自由呼吸运动下从肿瘤到整个器官运动的分级估计框架，该框架包含一个肿瘤运动估计模型和一个器官运动估计模型，其中，该肿瘤运动估计模型以先验肿瘤位置数据集和呼吸关联信号作为先验知识构建得到，该器官运动估计模型以先验速度向量场和先验肿瘤位置数据集作为先验知识构建得到。在对肿瘤患者进行手术的过程中，可以先将该患者的呼吸关联信号作为代理信号输入该肿瘤运动估计模型，得到估计的肿瘤位置；然后，再将该估计的肿瘤位置输入该器官运动估计模型，得到估计的整个器官的速度向量场，从而实现对整个器官的呼吸运动估计，提高对肿瘤及其周围重要解剖结构进行定位的准确性。
如图2所示，为本申请提出的肿瘤的运动估计方法的一种操作原理示意图。该运动估计方法可以分为术前和术中两个阶段：
在术前阶段，获取患者的四维的CT图像数据集，该图像数据集包含该患者的指定器官在各个不同呼吸状态下对应的三维CT图像；一方面，从该图像数据集中分割出指定器官的体数据(器官掩模)，根据这些体数据执行微分同胚的变形配准处理，得到各个不同呼吸状态下分别对应的速度向量场(即先验速度向量场)；另一方面，根据该CT图像数据集中提取出的肿瘤位置(即先验肿瘤位置数据集)以及预采集的光学标记(该光学标记可以粘贴于患者胸腹部位的皮肤表面)的运动轨迹信号，经过相位匹配后得到各个不同呼吸状态下对应的外部标记点-肿瘤位置配对数据集；然后，基于该外部标记点-肿瘤位置配对数据集，训练得到肿瘤运动估计模型；基于该先验速度向量场和该先验肿瘤位置数据集，训练得到器官运动估计模型。
在术中阶段，实时检测体表光学标记的运动信号，将当前的光学标记位置输入术前构建的肿瘤运动估计模型，得到估计的肿瘤位置；然后，将估计的肿瘤位置输入术前构建的器官运动估计模型，对整个器官的呼吸运动进行估计，得到对应的稠密速度向量场；最后，通过对稠密速度向量场执行群指数映射处理，可以得到对应的微分同胚稠密变形场(即位移向量场)，进而实时重构器 官的运动形态，实现分级运动估计。
总的来说，本申请提出的自由呼吸运动下从肿瘤到整个器官运动的分级估计框架在实现整个器官准确运动估计的基础上，能够实现对肿瘤更高精度的估计；而且，在术中对整个器官呼吸运动的估计不需要进行迭代优化计算，因而可以具有更好的处理实时性；另外，本申请采用器官运动估计模型估计得到的整个器官的稠密变形场是微分同胚的，也即具有全局一对一，可微和可逆的优点。
应理解，上述各个实施例中各步骤的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。
上面主要描述了一种肿瘤的运动估计方法，下面将对一种肿瘤的运动估计装置进行描述。
请参阅图3，本申请实施例中一种肿瘤的运动估计装置的一个实施例包括：
呼吸关联信号获取模块301，用于获取患者的呼吸关联信号的当前值，所述呼吸关联信号反映所述患者带有肿瘤的指定器官在各个不同呼吸状态下的运动轨迹特征；
肿瘤运动估计模块302，用于将所述呼吸关联信号的当前值输入肿瘤运动估计模型，得到估计的所述肿瘤的当前位置，所述肿瘤运动估计模型以先验肿瘤位置数据集和所述呼吸关联信号作为先验知识构建得到，所述先验肿瘤位置数据集根据预采集的所述指定器官的图像数据集确定，且包含所述肿瘤在所述各个不同呼吸状态下的位置，所述图像数据集包含所述指定器官在所述各个不同呼吸状态下的三维图像；
器官运动估计模块303，用于将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，所述当前速度向量场包含当前所述指定器官的各个位置点分别对应的速度向量，所述器官运动估计模型以先验速度向量场和所述先验肿瘤位置数据集作为先验知识构建得到，所 述先验速度向量场根据所述图像数据集确定，且包含所述指定器官在所述各个不同呼吸状态下的速度向量场。
在本申请的一个实施例中，所述呼吸关联信号为设置于所述患者的身体指定部位的光学标记的运动轨迹信号，所述运动轨迹信号包含所述光学标记在各个时间点下的空间位置，所述运动估计装置还可以包括：
配对数据集构建模块，用于根据所述先验肿瘤位置数据集和所述运动轨迹信号构建位置配对数据集，所述位置配对数据集包含每个所述呼吸状态下分别对应的所述先验肿瘤位置数据集中的肿瘤位置和所述运动轨迹信号中的光学标记位置；
肿瘤运动估计模型构建模块，用于以所述位置配对数据集作为先验知识，构建得到所述肿瘤运动估计模型。
进一步的，所述肿瘤运动估计模型构建模块具体可以用于：基于ε-SVR机器学习模型和所述位置配对数据集，以所述运动轨迹信号作为代理信号，在各个预设空间坐标方向上分别建立对应的肿瘤运动估计模型。
在本申请的一个实施例中，所述运动估计装置还可以包括：
体数据获取模块，用于从所述图像数据集中分割出所述指定器官在所述各个不同呼吸状态下的体数据；
微分同胚变形配准模块，用于从所述各个不同呼吸状态中选取一个呼吸状态作为参考状态，并对所述参考状态下的体数据和其它状态下的体数据执行微分同胚变形配准处理，得到所述先验速度向量场，所述其它状态为所述各个不同呼吸状态中除所述参考状态外的其它呼吸状态；
器官运动估计模型构建模块，用于以所述先验速度向量场和所述先验肿瘤位置数据集作为先验知识，构建得到所述器官运动估计模型。
进一步的，所述微分同胚变形配准模块可以包括：
速度向量场赋值单元，用于为所述指定器官在所述参考状态下的速度向量场赋予预设的初始值；
速度向量场更新单元，用于针对所述其它状态中的每个呼吸状态，根据该呼吸状态下的体数据以及所述参考状态下的体数据，采用交替优化的策略计算得到该呼吸状态对应的速度向量场的更新值，并将所述更新值与所述初始值相加，得到所述指定器官在该呼吸状态下的速度向量场。
在本申请的一个实施例中，所述器官运动估计模块可以包括：
空间插值单元，用于针对所述指定器官的每个位置点，将该位置点的速度向量作为估计的所述肿瘤的当前位置的函数，并根据估计的所述肿瘤的当前位置，采用空间插值的方法往该位置点在所述各个不同呼吸状态下的速度向量之间插值，得到估计的该位置点的当前速度向量，其中，该位置点在所述各个不同呼吸状态下的速度向量根据所述先验速度向量场确定。
在本申请的一个实施例中，所述运动估计装置还可以包括：
群指数变换模块，用于采用预设的高斯核函数对所述当前速度向量场进行正则化，并通过群指数变换的方式获得所述指定器官的当前稠密位移场，所述当前稠密位移场包含当前所述指定器官的各个位置点分别对应的位移向量。
本申请实施例还提供一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行时实现如图1表示的任意一种肿瘤的运动估计方法。
本申请实施例还提供一种计算机程序产品，当该计算机程序产品在终端设备上运行时，使得终端设备执行实现如图1表示的任意一种肿瘤的运动估计方法。
图4是本申请一实施例提供的终端设备的示意图。如图4所示，该实施例的终端设备4包括：处理器40、存储器41以及存储在所述存储器41中并可在所述处理器40上运行的计算机程序42。所述处理器40执行所述计算机程序42时实现上述各个肿瘤的运动估计方法的实施例中的步骤，例如图1所示的步骤101至103。或者，所述处理器40执行所述计算机程序42时实现上述各装置实施例中各模块/单元的功能，例如图3所示模块301至303的功能。
所述计算机程序42可以被分割成一个或多个模块/单元，所述一个或者多个模块/单元被存储在所述存储器41中，并由所述处理器40执行，以完成本申请。所述一个或多个模块/单元可以是能够完成特定功能的一系列计算机程序指令段，该指令段用于描述所述计算机程序42在所述终端设备4中的执行过程。
所称处理器40可以是中央处理单元(Central Processing Unit，CPU)，还可以是其他通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现成可编程门阵列(Field-Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
所述存储器41可以是所述终端设备4的内部存储单元，例如终端设备4的硬盘或内存。所述存储器41也可以是所述终端设备4的外部存储设备，例如所述终端设备4上配备的插接式硬盘，智能存储卡(Smart Media Card,SMC)，安全数字(Secure Digital,SD)卡，闪存卡(Flash Card)等。进一步地，所述存储器41还可以既包括所述终端设备4的内部存储单元也包括外部存储设备。所述存储器41用于存储所述计算机程序以及所述终端设备所需的其他程序和数据。所述存储器41还可以用于暂时地存储已经输出或者将要输出的数据。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。实施例中的各功能单元、模块可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中，上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。另外，各功能单元、模块的具体名称也只是为了便于相互区分，并不用于限制本申请的保护范围。上述系统中单元、模块的具体工作过程，可以参考前述方法实施例中的对应过程， 在此不再赘述。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统，装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述或记载的部分，可以参见其它实施例的相关描述。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
在本申请所提供的实施例中，应该理解到，所揭露的装置和方法，可以通过其它的方式实现。例如，以上所描述的系统实施例仅仅是示意性的，例如，所述模块或单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通讯连接可以是通过一些接口，装置或单元的间接耦合或通讯连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请实现上述实施例方法中的全部或部分流程，也可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一计算机可读存储介质中，该计算机程序在被处理器执行时，可实现上述各个方法实施例的步骤。其中，所述计算机程序包括计算机程序代码，所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读介质可以包括：能够携带所述计算机程序代码的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、电载波信号、电信信号以及软件分发介质等。需要说明的是，所述计算机可读介质包含的内容可以根据司法管辖区内立法和专利实践的要求进行适当的增减，例如在某些司法管辖区，根据立法和专利实践，计算机可读介质不包括是电载波信号和电信信号。
以上所述实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围，均应包含在本申请的保护范围之内。

Claims (10)

  1. 一种肿瘤的运动估计方法，其特征在于，包括：
    获取患者的呼吸关联信号的当前值，所述呼吸关联信号反映所述患者带有肿瘤的指定器官在各个不同呼吸状态下的运动轨迹特征；
    将所述呼吸关联信号的当前值输入肿瘤运动估计模型，得到估计的所述肿瘤的当前位置，所述肿瘤运动估计模型以先验肿瘤位置数据集和所述呼吸关联信号作为先验知识构建得到，所述先验肿瘤位置数据集根据预采集的所述指定器官的图像数据集确定，且包含所述肿瘤在所述各个不同呼吸状态下的位置，所述图像数据集包含所述指定器官在所述各个不同呼吸状态下的三维图像；
    将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，所述当前速度向量场包含当前所述指定器官的各个位置点分别对应的速度向量，所述器官运动估计模型以先验速度向量场和所述先验肿瘤位置数据集作为先验知识构建得到，所述先验速度向量场根据所述图像数据集确定，且包含所述指定器官在所述各个不同呼吸状态下的速度向量场。
  2. 如权利要求1所述的方法，其特征在于，所述呼吸关联信号为设置于所述患者的身体指定部位的光学标记的运动轨迹信号，所述运动轨迹信号包含所述光学标记在各个时间点下的空间位置，在将所述呼吸关联信号的当前值输入肿瘤运动估计模型之前，还包括：
    根据所述先验肿瘤位置数据集和所述运动轨迹信号构建位置配对数据集，所述位置配对数据集包含每个所述呼吸状态下分别对应的所述先验肿瘤位置数据集中的肿瘤位置和所述运动轨迹信号中的光学标记位置；
    以所述位置配对数据集作为先验知识，构建得到所述肿瘤运动估计模型。
  3. 如权利要求2所述的方法，其特征在于，以所述位置配对数据集作为先验知识，构建得到所述肿瘤运动估计模型，包括：
    基于ε-SVR机器学习模型和所述位置配对数据集，以所述运动轨迹信号作 为代理信号，在各个预设空间坐标方向上分别建立对应的肿瘤运动估计模型。
  4. 如权利要求1所述的方法，其特征在于，在将所述呼吸关联信号的当前值输入肿瘤运动估计模型之前，还包括：
    从所述图像数据集中分割出所述指定器官在所述各个不同呼吸状态下的体数据；
    从所述各个不同呼吸状态中选取一个呼吸状态作为参考状态，并对所述参考状态下的体数据和其它状态下的体数据执行微分同胚变形配准处理，得到所述先验速度向量场，所述其它状态为所述各个不同呼吸状态中除所述参考状态外的其它呼吸状态；
    以所述先验速度向量场和所述先验肿瘤位置数据集作为先验知识，构建得到所述器官运动估计模型。
  5. 如权利要求4所述的方法，其特征在于，对所述参考状态下的体数据和其它状态下的体数据执行微分同胚变形配准处理，得到所述先验速度向量场，包括：
    为所述指定器官在所述参考状态下的速度向量场赋予预设的初始值；
    针对所述其它状态中的每个呼吸状态，根据该呼吸状态下的体数据以及所述参考状态下的体数据，采用交替优化的策略计算得到该呼吸状态对应的速度向量场的更新值，并将所述更新值与所述初始值相加，得到所述指定器官在该呼吸状态下的速度向量场。
  6. 如权利要求1所述的方法，其特征在于，将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，包括：
    针对所述指定器官的每个位置点，将该位置点的速度向量作为估计的所述肿瘤的当前位置的函数，并根据估计的所述肿瘤的当前位置，采用空间插值的方法往该位置点在所述各个不同呼吸状态下的速度向量之间插值，得到估计的该位置点的当前速度向量，其中，该位置点在所述各个不同呼吸状态下的速度向量根据所述先验速度向量场确定。
  7. 如权利要求1至6中任一项所述的方法，其特征在于，在得到估计的所述指定器官的当前速度向量场之后，还包括：
    采用预设的高斯核函数对所述当前速度向量场进行正则化，并通过群指数变换的方式获得所述指定器官的当前稠密位移场，所述当前稠密位移场包含当前所述指定器官的各个位置点分别对应的位移向量。
  8. 一种肿瘤的运动估计装置，其特征在于，包括：
    呼吸关联信号获取模块，用于获取患者的呼吸关联信号的当前值，所述呼吸关联信号反映所述患者带有肿瘤的指定器官在各个不同呼吸状态下的运动轨迹特征；
    肿瘤运动估计模块，用于将所述呼吸关联信号的当前值输入肿瘤运动估计模型，得到估计的所述肿瘤的当前位置，所述肿瘤运动估计模型以先验肿瘤位置数据集和所述呼吸关联信号作为先验知识构建得到，所述先验肿瘤位置数据集根据预采集的所述指定器官的图像数据集确定，且包含所述肿瘤在所述各个不同呼吸状态下的位置，所述图像数据集包含所述指定器官在所述各个不同呼吸状态下的三维图像；
    器官运动估计模块，用于将估计的所述肿瘤的当前位置输入器官运动估计模型，得到估计的所述指定器官的当前速度向量场，所述当前速度向量场包含当前所述指定器官的各个位置点分别对应的速度向量，所述器官运动估计模型以先验速度向量场和所述先验肿瘤位置数据集作为先验知识构建得到，所述先验速度向量场根据所述图像数据集确定，且包含所述指定器官在所述各个不同呼吸状态下的速度向量场。
  9. 一种终端设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现如权利要求1至7中任一项所述的肿瘤的运动估计方法。
  10. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至7中任 一项所述的肿瘤的运动估计方法。
PCT/CN2021/138010 2021-04-20 2021-12-14 一种肿瘤的运动估计方法、装置、终端设备和存储介质 Ceased WO2022222502A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US18/366,704 US12537106B2 (en) 2021-04-20 2023-08-08 Motion estimation method and apparatus for tumor, terminal device, and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110424937.1A CN113112486B (zh) 2021-04-20 2021-04-20 一种肿瘤的运动估计方法、装置、终端设备和存储介质
CN202110424937.1 2021-04-20

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/366,704 Continuation US12537106B2 (en) 2021-04-20 2023-08-08 Motion estimation method and apparatus for tumor, terminal device, and storage medium

Publications (1)

Publication Number Publication Date
WO2022222502A1 true WO2022222502A1 (zh) 2022-10-27

Family

ID=76718938

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/138010 Ceased WO2022222502A1 (zh) 2021-04-20 2021-12-14 一种肿瘤的运动估计方法、装置、终端设备和存储介质

Country Status (3)

Country Link
US (1) US12537106B2 (zh)
CN (1) CN113112486B (zh)
WO (1) WO2022222502A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119868822A (zh) * 2024-11-28 2025-04-25 西安大医集团股份有限公司 肿瘤定位方法、电子设备和存储介质

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113112486B (zh) * 2021-04-20 2022-11-29 中国科学院深圳先进技术研究院 一种肿瘤的运动估计方法、装置、终端设备和存储介质
CN113761230B (zh) * 2021-09-08 2023-08-15 深圳市大数据研究院 一种测算全国各地文书公开率的方法
KR102486818B1 (ko) * 2022-01-04 2023-01-10 주식회사 클라리파이 의료영상의 움직임 보상처리 장치 및 방법
CN115317011A (zh) * 2022-07-12 2022-11-11 上海精劢医疗科技有限公司 体内目标区域呼吸运动估计方法、系统、介质及设备
CN115463352B (zh) * 2022-09-20 2025-06-06 同济人工智能研究院(苏州)有限公司 一种基于遗传算法的放疗机器人交付方向优化方法和系统
CN116072253A (zh) * 2023-03-16 2023-05-05 江苏铁人科技有限公司 一种人体数据实时捕捉系统
CN117115221B (zh) * 2023-09-14 2026-03-03 中国科学院合肥物质科学研究院 一种肺肿瘤位置和形态实时估计方法、系统和存储介质
CN119055972B (zh) * 2024-09-23 2025-08-15 陕西省肿瘤医院(陕西省肿瘤防治研究所)(陕西省第三人民医院) 一种在放疗过程追踪肿瘤位置变化的定位系统及方法
CN119318493A (zh) * 2024-10-11 2025-01-17 四川大学华西医院 一种基于肌电信号的肿瘤运动预测方法和系统
CN119280713B (zh) * 2024-11-26 2025-08-19 南通市肿瘤医院（南通市第五人民医院） 一种放疗机器人肿瘤运动估计预测系统及方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100198101A1 (en) * 2007-09-24 2010-08-05 Xubo Song Non-invasive location and tracking of tumors and other tissues for radiation therapy
CN103761745A (zh) * 2013-07-31 2014-04-30 深圳大学 一种肺部运动模型估计方法及系统
CN104268895A (zh) * 2014-10-24 2015-01-07 山东师范大学 一种联合空域和时域信息的4d-ct形变配准方法
CN111161333A (zh) * 2019-12-12 2020-05-15 中国科学院深圳先进技术研究院 一种肝脏呼吸运动模型的预测方法、装置及存储介质
CN113112486A (zh) * 2021-04-20 2021-07-13 中国科学院深圳先进技术研究院 一种肿瘤的运动估计方法、装置、终端设备和存储介质

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101623198A (zh) * 2008-07-08 2010-01-13 深圳市海博科技有限公司 动态肿瘤实时跟踪方法
US8824756B2 (en) * 2009-06-30 2014-09-02 University Of Utah Research Foundation Image reconstruction incorporating organ motion
CN101972515B (zh) * 2010-11-02 2012-05-09 华中科技大学 图像和呼吸引导的辅助放疗床垫系统
KR102070427B1 (ko) * 2012-08-08 2020-01-28 삼성전자주식회사 종양의 위치를 추적하는 방법 및 장치
CN106563210B (zh) * 2016-11-10 2020-07-10 苏州大学 基于ut变换的放疗机器人肿瘤呼吸运动估计及预测方法
CN110473440B (zh) * 2019-07-09 2021-08-17 中国科学院深圳先进技术研究院 肿瘤呼吸运动模拟平台及肿瘤位置估计方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100198101A1 (en) * 2007-09-24 2010-08-05 Xubo Song Non-invasive location and tracking of tumors and other tissues for radiation therapy
CN103761745A (zh) * 2013-07-31 2014-04-30 深圳大学 一种肺部运动模型估计方法及系统
CN104268895A (zh) * 2014-10-24 2015-01-07 山东师范大学 一种联合空域和时域信息的4d-ct形变配准方法
CN111161333A (zh) * 2019-12-12 2020-05-15 中国科学院深圳先进技术研究院 一种肝脏呼吸运动模型的预测方法、装置及存储介质
CN113112486A (zh) * 2021-04-20 2021-07-13 中国科学院深圳先进技术研究院 一种肿瘤的运动估计方法、装置、终端设备和存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
YU YONGHUA , ZHAO YUEHUAN, LUO LIMIN, YU JINMING, LI BAOSHENG, LIANG CHAOQIAN: "Establishment of a Mathematical Model for the Influence of Respiratory Movement upon the Position of an Intrahepatic Space-occupying Lesion", CHINESE JOURNAL OF RADIATION ONCOLOGY, vol. 11, no. 4, 25 December 2002 (2002-12-25), pages 245 - 247, XP055979942 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119868822A (zh) * 2024-11-28 2025-04-25 西安大医集团股份有限公司 肿瘤定位方法、电子设备和存储介质

Also Published As

Publication number Publication date
US12537106B2 (en) 2026-01-27
US20230377758A1 (en) 2023-11-23
CN113112486A (zh) 2021-07-13
CN113112486B (zh) 2022-11-29

Similar Documents

Publication Publication Date Title
WO2022222502A1 (zh) 一种肿瘤的运动估计方法、装置、终端设备和存储介质
US11890063B2 (en) System and methods for a trackerless navigation system
CN112885453B (zh) 用于标识后续医学图像中的病理变化的方法和系统
Han et al. A nonlinear biomechanical model based registration method for aligning prone and supine MR breast images
Haouchine et al. Image-guided simulation of heterogeneous tissue deformation for augmented reality during hepatic surgery
US20120296202A1 (en) Method and System for Registration of Ultrasound and Physiological Models to X-ray Fluoroscopic Images
CN116485850B (zh) 基于深度学习的手术导航影像的实时非刚性配准方法及系统
Williamson et al. Ultrasound-based liver tracking utilizing a hybrid template/optical flow approach
JP2015531607A (ja) ３次元物体を追跡するための方法
CN113870324A (zh) 多模态图像的配准方法及其配准装置和计算机可读存储介质
Vásquez Osorio et al. Accurate CT/MR vessel‐guided nonrigid registration of largely deformed livers
WO2022227597A1 (zh) 肝脏内部组织的位移预测方法、装置、系统及电子设备
WO2023092959A1 (zh) 图像分割方法及其模型的训练方法及相关装置、电子设备
Spinczyk et al. Methods for abdominal respiratory motion tracking
Ozkan et al. Robust motion tracking in liver from 2D ultrasound images using supporters
US20240206907A1 (en) System and Method for Device Tracking in Magnetic Resonance Imaging Guided Inerventions
CN116612166A (zh) 一种多模态影像的配准融合算法
CN114943690A (zh) 医学图像处理方法、装置、计算机设备及可读存储介质
Zheng Cross-modality medical image detection and segmentation by transfer learning of shapel priors
Lei et al. Diffeomorphic respiratory motion estimation of thoracoabdominal organs for image‐guided interventions
Krüger et al. Simulation of mammographic breast compression in 3D MR images using ICP-based B-spline deformation for multimodality breast cancer diagnosis
CN119919434A (zh) 基于深度学习的肝胆部位分割方法、装置、设备及介质
Puyol-Anton et al. Towards a multimodal cardiac motion atlas
Wei et al. A model that predicts a real-time tumour surface using intra-treatment skin surface and end-of-expiration and end-of-inhalation planning CT images
Meng et al. Weighted local mutual information for 2D-3D registration in vascular interventions

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21937724

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21937724

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 27.05.2024)

122 Ep: pct application non-entry in european phase

Ref document number: 21937724

Country of ref document: EP

Kind code of ref document: A1