WO2024002360A1 - 配准方法、装置、计算机设备和可读存储介质 - Google Patents

配准方法、装置、计算机设备和可读存储介质 Download PDF

Info

Publication number
WO2024002360A1
WO2024002360A1 PCT/CN2023/105062 CN2023105062W WO2024002360A1 WO 2024002360 A1 WO2024002360 A1 WO 2024002360A1 CN 2023105062 W CN2023105062 W CN 2023105062W WO 2024002360 A1 WO2024002360 A1 WO 2024002360A1
Authority
WO
WIPO (PCT)
Prior art keywords
registration
point set
image model
transformation matrix
point
Prior art date
Application number
PCT/CN2023/105062
Other languages
English (en)
French (fr)
Inventor
吴博
曹乐
刘双龙
Original Assignee
武汉联影智融医疗科技有限公司
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 武汉联影智融医疗科技有限公司 filed Critical 武汉联影智融医疗科技有限公司
Publication of WO2024002360A1 publication Critical patent/WO2024002360A1/zh

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • 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
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • 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
    • 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; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Definitions

  • the present application relates to the technical field of medical image processing, and in particular to a registration method, device, computer equipment and readable storage medium.
  • This application provides a registration method, device, computer equipment and readable storage medium.
  • an embodiment of the present application provides a registration method, which method includes: obtaining the first image model of the object to be registered used in the last registration and the first transformation matrix obtained from the last registration; executing The current registration includes: adjusting the first image model based on the first transformation matrix to obtain the second image model; projecting the first registration point set on the surface of the object to be registered to the second image model surface to obtain the projection point set ; Register the first registration point set and the projection point set to obtain the second transformation matrix.
  • each projection point in the projection point set is the closest projection point to each corresponding registration point in the first registration point set.
  • the first registration point set and the projection point set are registered to obtain a second transformation matrix, which includes: based on the pose coordinates of each registration point in the first registration point set and the projection The pose coordinates of each projection point in the point set determine the second transformation matrix.
  • determining the second transformation matrix based on the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set includes: converting each registration point in the first registration point set The pose coordinates of the registration point and the pose coordinates of each projection point in the projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity and substituted into the first objective function to solve the second transformation matrix; where, the first objective The function is configured to minimize the pose difference between the first conversion point set and the first registration point set.
  • the first conversion point set is obtained by converting the projection point set based on the second conversion matrix.
  • registering the first registration point set and the projection point set to obtain the second transformation matrix includes: obtaining the registration weight corresponding to each registration point in the first registration point set; wherein, for the first registration point set, For each registration point in the accurate point set, the registration weight is obtained based on the number of registrations for the current registration, the number of preset registrations, and at least one difference in the pose between the registration point and the corresponding projection point of the registration point; the first The registration weight corresponding to each registration point in the first registration point set, the pose coordinates of each registration point in the first registration point set, and the pose coordinates of each projection point in the projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity, Substitute into the second objective function to solve the second transformation matrix.
  • the second objective function is configured to minimize the pose difference between the first transformation point set and the first registration point set.
  • the first transformation point set is based on the first transformation point set.
  • the two transformation matrices set the projection points into Obtained after row conversion.
  • each registration point in the first registration point set is a physiological anatomical feature point on the surface of the object to be registered.
  • the registration method also includes: determining whether the number of registrations for the current registration has reached the first iteration threshold; if the number of registrations for the current registration has not reached the first iteration threshold, then using the second image model as the first The image model uses the second transformation matrix as the first transformation matrix and returns to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the registration method further includes: if the number of registrations for the current registration reaches a first iteration threshold, using the second image model as the first image model and the second transformation matrix as the first Transformation matrix, use the second registration point set on the surface of the object to be registered as the first registration point set, return to the steps of adjusting the first image model based on the first transformation matrix to obtain the second image model, until the current registration The number of registrations reaches the second iteration number threshold; wherein, each registration point in the second registration point set is a point other than the physiological and anatomical feature points on the surface of the object to be registered.
  • each registration point in the first registration point set is a point other than the physiological anatomical feature point on the surface of the object to be registered.
  • the first registration point set is obtained.
  • the registration method also includes: determining whether the number of registrations for the current registration has reached the third iteration threshold; if the number of registrations for the current registration has not reached the third iteration threshold, then the second The image model is used as the first image model, the second transformation matrix is used as the first transformation matrix, and the step of adjusting the first image model based on the first transformation matrix to obtain the second image model is returned.
  • obtaining the first transformation matrix obtained from the last registration includes: registering the registration point set on the surface of the first image model with the first registration point set to obtain the first transformation matrix.
  • the method of obtaining the registration point set of the first image model surface includes: obtaining a template image model, the template image model surface has a template point set; matching the template image model with the first image model, so as to match the template image model with the first image model.
  • the point set is mapped to the first image model surface to obtain a registration point set of the first image model surface.
  • the registration method further includes: determining whether the second transformation matrix obtained by the current registration reaches a preset transformation matrix threshold. If the second transformation matrix does not reach the preset transformation matrix threshold, Then the second image model is used as the first image model, the second conversion matrix is used as the first conversion matrix, and the adjustment of the first image model based on the first conversion matrix is performed back to obtain the first image model. The step of two image models is until the second transformation matrix obtained by the current registration reaches the preset transformation matrix threshold.
  • the second image model is moved and rotated based on the displacement and rotation angle in the second transformation matrix to obtain a matching image model.
  • an embodiment of the present application provides a registration device, including: an acquisition module, configured to acquire the first image model of the object to be registered used in the last registration and the first transformation obtained from the last registration. Matrix; adjustment module, for the current registration, adjusting the first image model based on the first transformation matrix to obtain the second image model; projection module, used to project the first registration point set on the surface of the object to be registered to the surface of the second image model to obtain the projection point set; the registration module is used to register the first registration point set and the projection point set to obtain the second transformation matrix.
  • an embodiment of the present application provides a computer device, including a memory and a processor.
  • the memory stores a computer program.
  • the processor executes the computer program, it implements the steps of the registration method provided in the first aspect.
  • an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored.
  • the computer program is executed by a processor, the steps of the registration method provided in the first aspect are implemented.
  • an embodiment of the present application further provides a computer program product, which includes a computer program.
  • a computer program product which includes a computer program.
  • Figure 1 is an application environment diagram of a registration method provided by an embodiment
  • Figure 2 is a schematic flow chart of the steps of a registration method provided by an embodiment
  • Figure 3 is a schematic flowchart of steps of a registration method provided by another embodiment
  • Figure 4 is a schematic flowchart of steps of a registration method provided by another embodiment
  • Figure 5 is a schematic flowchart of steps of a registration method provided by another embodiment
  • Figure 6 is a schematic flowchart of steps of a registration method provided by another embodiment
  • Figure 7 is a schematic diagram of the distribution of coarse registration points provided by an embodiment
  • Figure 8 is a schematic diagram of the distribution of fine registration points provided by an embodiment
  • Figure 9 is a schematic flowchart of steps of a registration method provided by another embodiment.
  • Figure 10 is a schematic flow chart of the steps of a registration method provided by another embodiment
  • Figure 11 is a schematic structural diagram of a registration device provided by an embodiment
  • Figure 12 is a schematic structural diagram of a computer device provided by an embodiment of the present application.
  • the registration method provided by the embodiment of the present application can be applied in an application environment as shown in Figure 1 , which includes a terminal 100 and a medical scanning device 200.
  • the terminal 100 can communicate with the medical scanning device 200 through the network.
  • the terminal 100 can be, but is not limited to, various personal computers, notebook computers, and tablet computers.
  • the medical scanning equipment 200 may be, but is not limited to, CT (Computed Tomography) equipment and PET (Positron Emission Computed Tomography).
  • a registration method is provided. Taking the method as applied to the terminal in Figure 1 as an example, the method includes the following steps:
  • Step 200 Obtain the first image model of the object to be registered used in the last registration and the first transformation matrix obtained from the last registration.
  • the first image model of the object to be registered used in the last registration refers to the image model that was registered with the object to be registered in the last registration process.
  • the first transformation matrix is used to characterize the transformation relationship between the image space and the physical space of the object to be registered determined in the last registration. Before the terminal performs the current registration, it first obtains the first image model of the object to be registered used in the previous registration, and the first transformation matrix obtained from the previous registration. This embodiment does not limit the specific method used for the last registration, as long as its function can be realized.
  • Step 210 Adjust the first image model based on the first transformation matrix to obtain a second image model.
  • the terminal adjusts the first image model according to the first transformation matrix obtained from the previous registration to obtain the second image model.
  • the first conversion matrix includes the displacement size that needs to be moved and the rotation angle that needs to be rotated. That is to say, the terminal moves and rotates the first image model according to the displacement size and rotation angle in the first conversion matrix to obtain the second image. Model.
  • Step 220 Project the first registration point set on the surface of the object to be registered to the second image model surface to obtain a projection point set.
  • Each registration point in the first set of registration points on the surface of the object to be registered may be a registration point preset by the staff on the surface of the object to be registered. This embodiment does not limit the specific number of each registration point in the first registration point set, as well as the specific location of each registration point, as long as its function can be realized.
  • the terminal After obtaining the second image model, the terminal projects the first registration point set on the surface of the object to be registered onto the second image model surface to obtain a projection point set.
  • the terminal projects the first registration point set on the surface of the object to be registered onto the second image model surface to obtain a projection point set.
  • the terminal projects the registration point onto the second image model surface to obtain the corresponding projection point of the registration point on the second image model surface. , thereby being able to obtain projection points corresponding to each registration point in the first registration point set on the surface of the second image model, that is, a projection point set.
  • Step 230 Register the first registration point set and the projection point set to obtain a second transformation matrix.
  • the terminal After obtaining the projection point set, the terminal registers the first registration point set of the object to be registered with the projection point set to obtain the second transformation matrix.
  • the second transformation matrix is used to represent the transformation relationship between the image space and the physical space of the object to be registered determined by the current secondary registration.
  • the second transformation matrix includes the displacement size that needs to be moved and the rotation angle that needs to be turned. This embodiment does not limit the specific process of registering the first registration point set and the projection point set, as long as its function can be realized.
  • the terminal moves and rotates the second image model according to the displacement and rotation angle in the second transformation matrix, so as to obtain an image that matches the object to be registered. Model.
  • the registration method provided by the embodiment of the present application obtains the first image model of the object to be registered used in the last registration and the transformation matrix obtained from the last registration; for the current registration, based on the first transformation matrix, the first image model of the object to be registered is obtained. Adjust an image model to obtain a second image model; project the first registration point set on the surface of the object to be registered onto the surface of the second image model to obtain a projection point set; register the first registration point set with the projection point set , get the second transformation matrix.
  • This embodiment uses the projection point set projected from the first registration point set onto the surface of the second image model to register with the first registration point set, and the registration point set on the second image model is directly matched with the first registration point set.
  • each projection point in the projection point set is the closest projection point to each corresponding registration point in the first registration point set. That is to say, there are multiple initial projections of each registration point in the first registration point set onto the surface of the second image model. point, the projection point in the projection point set corresponding to the registration point is the projection point closest to the registration point among multiple initial projection points.
  • each projection point in the projection point set is set as the closest projection point to each corresponding registration point in the first registration point set.
  • a possible implementation method involving registering the first registration point set and the projection point set to obtain the second transformation matrix includes:
  • the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity and substituted into the first objective function to solve the second transformation matrix; among them, the optimization goal of the first objective function is: minimizing the pose difference between the first conversion point set and the first registration point set.
  • the first conversion point set converts the projection point set based on the second conversion matrix. obtained later.
  • the second transformation matrix includes Displacement t 2 and rotation angle R 2 .
  • the first conversion point set can be expressed as: (R 2 *p' i +t 2 ), then the first objective function can be expressed as: Among them, N 1 is the number of registration points in the first registration point set.
  • the terminal substitutes the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set into the first objective function. If the displacement and rotation angle in the second transformation matrix can be obtained, the second transformation matrix can be obtained. Transformation matrix.
  • the first transformation matrix can be obtained by solving the first objective function using the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set.
  • a possible implementation method involving registering the first registration point set and the projection point set to obtain the second transformation matrix includes:
  • Step 300 Obtain the registration weight corresponding to each registration point in the first registration point set; wherein, for each registration point in the first registration point set, the registration weight is based on the number of registrations of the current registration and the preset registration. The times and the pose difference between the registration point and the corresponding projection point of the registration point are obtained.
  • the registration weight may be based on the number of registrations for the current registration, the number of preset registrations, and the distance between the registration point and the corresponding projection point of the registration point. At least one of the pose differences is obtained. For example, the registration weight of the registration point is obtained based on the pose difference between the registration point and the corresponding projection point of the registration point.
  • the number of registrations is different, and the registration weight corresponding to each registration point is different.
  • This embodiment does not limit the specific method of determining the registration weight, as long as its function can be realized.
  • the terminal obtains the registration weight corresponding to each registration point in the first registration point set.
  • the registration weight can be calculated using the following formula:
  • l represents the number of registrations for the current registration
  • M is the preset number of registrations
  • w i is the registration weight corresponding to the i-th registration point
  • is the i-th registration point
  • the projection point p i ' corresponding to the accurate point is registered with the i-th
  • Step 310 Use the registration weight corresponding to each registration point in the first registration point set, the pose coordinate of each registration point in the first registration point set, and the pose coordinate of each projection point in the projection point set as known quantities, and use the second transformation As an unknown quantity, the matrix is substituted into the second objective function to solve the second transformation matrix.
  • the optimization goal of the second objective function is to minimize the pose difference between the first conversion point set and the first registration point set.
  • the first conversion point set is obtained by converting the projection point set based on the second conversion matrix. of.
  • the second objective function can be expressed as:
  • the terminal substitutes the known registration weight corresponding to each registration point in the first registration point set, the pose coordinates of each registration point in the first registration point set, and the pose coordinates of each projection point in the projection point set into the second objective function, and can solve the problem By obtaining the displacement and rotation angle in the second transformation matrix, the second transformation matrix can be obtained.
  • the calculation method of the second conversion matrix in this embodiment has simple logic, is easy to implement, and has high calculation efficiency.
  • each registration point in the first registration point set is a physiological anatomical feature point on the surface of the object to be registered, that is, a coarse registration point.
  • the coarse registration point setting method can be in the form of a probe, and the positioning form of the probe can be divided into visual tracking or magnetic navigation positioning. This embodiment does not limit this. Based on this, as shown in Figure 4, after registering the first registration point set and the projection point set to obtain the second transformation matrix, the steps of the registration method also include:
  • Step 400 Determine whether the number of registrations for the current registration reaches the first iteration number threshold.
  • the first iteration number threshold may be preset by the staff and stored in the terminal. After obtaining the second conversion matrix, the terminal obtains the registration number of the current registration of the second conversion matrix and the first iteration number threshold, and compares the registration number of the current registration with the first iteration number threshold, and determines Whether the registration number of the current registration reaches the first iteration number threshold, that is, whether the registration number of the current registration is equal to the first iteration number threshold.
  • Step 410 If the number of registrations for the current registration does not reach the first iteration threshold, use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to execute the method based on the first transformation matrix. The step of adjusting the first image model to obtain the second image model.
  • the model used for the current registration As the first image model, use the second transformation matrix obtained by the current registration as the first transformation matrix, and return to step 210, step 220, step 230, step 400 and step 410.
  • the second image model is used as the first image model
  • the second transformation matrix is used as the first transformation matrix
  • the execution step 210 is returned.
  • step 220, step 230, step 400 and step 410 multiple registrations are implemented based on rough registration points, which can improve the accuracy of registration.
  • the condition for stopping the current registration also includes whether the second transformation matrix obtained by the current registration reaches a preset transformation matrix threshold. That is to say, if the second transformation matrix obtained by the current registration does not reach the preset transformation matrix threshold; then use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to step 210. , step 220, step 230, step 400 and step 410, until the second conversion obtained by the current registration The matrix reaches the preset conversion threshold.
  • the preset transformation threshold may include, for example, thresholds for displacement and rotation angles in the second transformation matrix.
  • the conditions for stopping the current registration may include whether the number of registrations for the current registration reaches the first iteration threshold or whether the second transformation matrix obtained by the current registration reaches the preset transformation matrix. threshold. That is to say, as long as one of the two conditions for stopping the current registration is met, the registration will stop.
  • the conditions for stopping the current registration may include whether the number of registrations for the current registration reaches the first iteration threshold and whether the second transformation matrix obtained by the current registration reaches the preset value. Transformation matrix threshold. Registration stops only when both conditions are met.
  • the steps of the registration method further include:
  • Step 420 If the number of registrations for the current registration reaches the first iteration number threshold, use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and use the second image model on the surface of the object to be registered.
  • the second registration point set is used as the first registration point set, and the step of adjusting the first image model based on the first transformation matrix to obtain the second image model is returned until the number of registrations for the current registration reaches the second iteration number threshold;
  • each registration point in the second registration point set is a point other than the physiological and anatomical feature points on the surface of the object to be registered.
  • the second set of registration points on the surface of the object to be registered are points other than the physiological anatomical feature points of the surface of the object to be registered, that is, each registration point in the second set of registration points is different from each registration point in the first set of registration points.
  • Each registration point in the second registration point set can also be called a fine registration point.
  • the fine registration point can be in the form of a probe.
  • the positioning form of the probe can be divided into visual tracking or magnetic navigation positioning. This embodiment does not limit this.
  • the terminal uses the model (second image model) used in the current registration as the first image.
  • Model use the second transformation matrix obtained from the current registration as the first transformation matrix, use the second registration point set on the surface of the object to be registered as the first registration point set, and return to step 210, step 220, step 230, and step 400 and step 410 until it is determined that the current number of registrations reaches the second iteration number threshold.
  • the second registration point set is used for coarse registration. Performing multiple registrations on two sets of registration points (that is, using precise registration points for precise registration) can improve the accuracy of registration.
  • each registration point in the first registration point set is a point other than the physiological anatomical feature point on the surface of the object to be registered, that is, a fine registration point.
  • a fine registration point For a detailed description of the precise alignment point, reference may be made to the description in the above embodiment, and details will not be described again here.
  • Step 500 Determine whether the number of registrations for the current registration reaches the third iteration number threshold.
  • the third iteration number threshold may be preset by the staff and stored in the memory of the terminal.
  • the third iteration number threshold may be the same as the first iteration number threshold, or may be different.
  • the terminal After obtaining the second conversion matrix, the terminal obtains the registration number of the current registration of the second conversion matrix and the third iteration number threshold, and compares the registration number of the current registration with the third iteration number threshold, and determines Whether the registration number of the current registration reaches the third iteration number threshold, that is, whether the registration number of the current registration is equal to the third iteration number threshold.
  • Step 510 If the number of registrations for the current registration does not reach the third iteration threshold, use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to execute the method based on the first transformation matrix. The step of adjusting the first image model to obtain the second image model.
  • the model used currently (the second image model) is used as the first image model, and the second transformation matrix obtained by the current registration is used as The first transformation matrix returns to step 210, step 220, step 230, step 500 and step 510.
  • Step 220, step 230, step 500 and step 510 implement multiple registrations based on precise registration points, which can improve the accuracy of registration.
  • the last registration is the initial registration, that is, the first image model used in the first registration is image data obtained by scanning the object to be registered with a CT device, and the image data is sent to The terminal uses a three-dimensional surface reconstruction algorithm to perform three-dimensional reconstruction of the image data to obtain an image model of the object to be registered.
  • a possible implementation method of obtaining the first transformation matrix obtained from the last registration includes:
  • the terminal After acquiring the first image model, the terminal selects corresponding points on the surface of the object to be registered through the registration point set on the surface of the first image model to form a first registration point set, and combines the registration point set on the surface of the first image model with the first registration point set.
  • the first transformation matrix can be obtained. This embodiment does not limit the specific method of registering the registration point set on the surface of the first image model with the first registration point set, as long as the first transformation matrix can be obtained.
  • the terminal may use a point-pair registration algorithm to register the registration point set on the surface of the first image model with the first registration point set to obtain the first transformation matrix.
  • the terminal may use the pose coordinates of each registration point in the registration point set on the first image model surface and the pose coordinates of each registration point in the first registration point set as known quantities, and convert the first transformation As an unknown quantity, the matrix is substituted into the third objective function to solve the first transformation matrix.
  • the third objective function can be expressed as: The terminal substitutes the pose coordinates of each registration point in the registration point set on the surface of the first image model and the pose coordinates of each registration point in the first registration point set into the third objective function to obtain the displacement and rotation angle in the first transformation matrix. , then the first transformation matrix can be obtained.
  • a registration method is provided that obtains the first transformation matrix by registering the registration point set on the surface of the first image model with the first registration point set. This method is simple, easy to understand, and easy to implement. . Moreover, when using the registration method provided by this embodiment, one registration is performed first, and then the current registration is performed, which can improve the efficiency and accuracy of the current registration.
  • Figure 6 involves a possible implementation of obtaining the registration point set of the first image model surface.
  • the implementation includes:
  • Step 600 Obtain a template image model, which has a template point set on its surface.
  • the template image model refers to the general model used for registration. For example, a universal three-dimensional digital model of the pelvis.
  • the template image model can be stored in the memory of the terminal in advance by the staff.
  • the terminal obtains the template image model directly from the memory.
  • the surface of the template image model has a template point set, and the template point set may include physiological anatomical feature points (coarse registration points), and points other than physiological anatomical feature points (fine registration points).
  • the setting of physiological and anatomical feature points facilitates the staff to find the corresponding points on the object to be registered.
  • the number of physiological anatomical feature points in the template point set is at least 3 non-collinear point.
  • the coarse registration points are the physiological and anatomical feature points of the pelvis
  • the fine registration points are set on the bone surface of the area to be processed.
  • the number of fine registration points is at least 3, and the optimal number is more than 10.
  • Step 610 Match the template image model with the first image model to map the template point set to the first image model surface to obtain a registration point set of the first image model surface.
  • the terminal After acquiring the template image model, the terminal matches the template image model with the first image model, and can map the template point set on the template image model to the first image model surface to obtain a registration point set on the first image model surface.
  • a registration point set of the first image model is obtained, that is, an individualized registration of the object to be detected can be obtained. Accurate point, thereby more intuitively guiding the staff to determine the registration point set corresponding to the registration point set on the surface of the object to be detected and the first image model surface.
  • the distribution of coarse registration points on the surface of the first image model is shown in Figure 7
  • the distribution of fine registration points on the surface of the first image model is shown in Figure 8.
  • this application provides a registration method.
  • the steps of the method include:
  • Step 900 Obtain a template image model; the surface of the template image model has a template point set; wherein the template point set includes a first template point set and a second template point set, and each template point in the first template point set is an image of a common template object. Physiological anatomical feature points on the model surface, each template point in the second template point set is a point other than the physiological anatomical feature points on the image model surface of the general template object;
  • Step 910 Match the template image model with the first image model of the object to be registered to map the template point set to the first image model surface to obtain the registration point set of the first image model surface; wherein, the first image model
  • the registration point set of the surface includes a first sub-registration point set corresponding to the first template point set, and a second sub-registration point set corresponding to the second template point set.
  • Each sub-registration point in the first sub-registration point set is to be matched.
  • Physiological and anatomical feature points on the surface of the image model of the quasi-object, and each sub-registration point in the second sub-registration point set is a point other than the physiological and anatomical feature points on the surface of the image model of the object to be registered;
  • Step 920 Select corresponding points on the surface of the object to be registered through the first sub-registration point set on the surface of the first image model to form a first registration point set, and use the second sub-registration point set on the surface of the first image model on the object to be registered. Select corresponding points on the object surface to form a second registration point set.
  • the first registration point set is the physiological anatomical feature points selected on the surface of the object to be registered
  • the second registration point set is the physiological anatomical feature points selected on the surface of the object to be registered. points outside;
  • Step 930 Register the first sub-registration point set on the surface of the first image model with the first registration point set on the surface of the object to be registered to obtain a first transformation matrix
  • Step 940 Adjust the first image model based on the first transformation matrix to obtain a second image model
  • Step 950 Project the first registration point set on the surface of the object to be registered to the second image model surface to obtain a projection point set
  • Step 960 Register the first registration point set and the projection point set to obtain the second transformation matrix
  • the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity and substituted into the first objective function to solve the second transformation matrix;
  • Step 970 Determine whether the number of registrations for the current registration reaches the first iteration number threshold
  • Step 980 If the number of registrations for the current registration does not reach the first iteration threshold, use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to step 940 to step 980. ;
  • Step 990 If the number of registrations for the current registration reaches the first iteration number threshold, use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and use the second image model on the surface of the object to be registered.
  • the second set of registration points serves as the A registration point set, return to step 940 to step 990 until the number of registrations for the current registration reaches the second iteration number threshold; wherein, each registration point in the second registration point set is a physiological anatomical feature point on the surface of the object to be registered points outside.
  • this application provides a registration method.
  • the steps of the method include:
  • Step 101 Obtain a template image model; the template image model has a template point set on its surface; wherein, the template point set includes a first template point set and a second template point set, and each template point in the first template point set is an image of a universal template object.
  • Step 102 Match the template image model with the first image model of the object to be registered to map the template point set to the first image model surface to obtain the registration point set of the first image model surface; wherein, the first image model
  • the registration point set of the surface includes a first sub-registration point set corresponding to the first template point set, and a second sub-registration point set corresponding to the second template point set.
  • Each sub-registration point in the first sub-registration point set is to be matched.
  • Physiological and anatomical feature points on the surface of the image model of the quasi-object, and each sub-registration point in the second sub-registration point set is a point other than the physiological and anatomical feature points on the surface of the image model of the object to be registered;
  • Step 103 Select corresponding points on the surface of the object to be registered through the first sub-registration point set on the first image model surface to form a first registration point set, and use the second sub-registration point set on the first image model surface on the object to be registered. Select corresponding points on the object surface to form a second registration point set.
  • the first registration point set is the physiological anatomical feature points selected on the surface of the object to be registered
  • the second registration point set is the physiological anatomical feature points selected on the surface of the object to be registered. points outside;
  • Step 104 Register the first sub-registration point set on the surface of the first image model with the first registration point set on the surface of the object to be registered to obtain a first transformation matrix
  • Step 105 Adjust the first image model based on the first transformation matrix to obtain the second image model
  • Step 106 Project the first registration point set on the surface of the object to be registered onto the second image model surface to obtain the first projection point set;
  • Step 107 Register the first registration point set and the first projection point set to obtain the second transformation matrix
  • the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the first projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity and substituted into the first objective function to solve the first objective function.
  • Step 108 Determine whether the number of registrations for the current registration reaches the first iteration number threshold
  • Step 109 If the number of registrations for the current registration does not reach the first iteration threshold, use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to step 105 to step 109. ;
  • Step 110 If the number of registrations for the current registration reaches the first iteration threshold, perform the next registration:
  • Step 111 For the next registration, adjust the first image model based on the second transformation matrix to obtain the third image model; project the second registration point set on the surface of the object to be registered to the surface of the third image model to obtain the third image model. Two projection point sets;
  • Step 112 Obtain the registration weight corresponding to each registration point in the second registration point set; wherein, for each registration point in the second registration point set, the registration weight is based on the number of registrations for the next registration and the number of first iterations.
  • the difference between the thresholds, the preset number of registrations, and the pose difference between the registration point and the second projection point corresponding to the registration point are obtained; where the preset number of registrations is equal to the sum of the second iteration number threshold and the first iteration number threshold. The difference between;
  • Step 113 Take the registration weight corresponding to each registration point in the second registration point set, the pose coordinate of each registration point in the second registration point set, and the pose coordinate of each projection point in the second projection point set as known quantities, and use the The third transformation matrix is used as an unknown quantity and is substituted into the second objective function to solve the third transformation matrix.
  • Step 114 Determine whether the number of registrations for the next registration reaches the second iteration number threshold
  • Step 115 If not, use the third image model as the first image model and the third transformation matrix as the second transformation moment. Array, return to step 111-step 115.
  • steps in the flowchart in the figure are shown in sequence as indicated by arrows, these steps are not necessarily executed in the order indicated by arrows. Unless explicitly stated in this article, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the figure may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but may be executed at different times. The execution order of these sub-steps or stages is It does not necessarily need to be performed sequentially, but may be performed in turn or alternately with other steps or sub-steps of other steps or at least part of the stages.
  • embodiments of the present application also provide a registration device for implementing the above-mentioned registration method.
  • the solution to the problem provided by this device is similar to the solution described in the above method. Therefore, for the specific limitations in one or more registration device embodiments provided below, please refer to the above limitations on the registration method. I won’t go into details here.
  • a registration device 10 which includes an acquisition module 11 , an adjustment module 12 , a projection module 13 and a registration module 14 .
  • the acquisition module 11 is used to obtain the first image model of the object to be registered used in the last registration and the first transformation matrix obtained from the last registration
  • the adjustment module 12 is used for the current registration, based on the first The conversion matrix adjusts the first image model to obtain a second image model
  • the projection module 13 is used to project the first registration point set on the surface of the object to be registered to the second image model surface to obtain a projection point set
  • the registration module 14 Used to register the first registration point set and the projection point set to obtain the second transformation matrix.
  • each projection point in the projection point set is the closest projection point to each corresponding registration point in the first registration point set.
  • the registration module 14 is specifically configured to use the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set as known quantities, and use the second transformation matrix as an unknown quantity. , substituted into the first objective function to solve the second transformation matrix; where, the optimization goal of the first objective function is: minimizing the pose gap between the first transformation point set and the first registration point set, the first transformation point The set is obtained by converting the projection point set based on the second transformation matrix.
  • the registration module 14 is specifically configured to obtain the registration weight corresponding to each registration point in the first registration point set; wherein, for each registration point in the first registration point set, the registration weight is based on the current The number of registrations, the number of preset registrations, and the pose difference between the registration point and the projection point corresponding to the registration point are obtained; the registration weight corresponding to each registration point in the first registration point set, and the registration weight corresponding to each registration point in the first registration point set are obtained.
  • the pose coordinates of the registration point and the pose coordinates of each projection point in the projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity and substituted into the second objective function to solve the second transformation matrix.
  • each registration point in the first registration point set is a physiological anatomical feature point on the surface of the object to be registered
  • the registration device 10 further includes a judgment module.
  • the judgment module is used to determine whether the number of registrations for the current registration has reached the first iteration threshold; if the number of registrations for the current registration has not reached the first iteration threshold, the second image model is used as the first image model, Using the second transformation matrix as the first transformation matrix, return to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the judgment module is also configured to use the second image model as the first image model and the second transformation matrix as the first transformation matrix if the number of registrations for the current registration reaches the first iteration threshold.
  • the second set of registration points on the surface of the object to be registered is used as the first set of registration points, and the step of adjusting the first image model based on the first transformation matrix to obtain the second image model is returned to the current registration.
  • the number of times reaches the second iteration threshold; wherein, each registration point in the second registration point set is a point other than the physiological anatomical feature point on the surface of the object to be registered.
  • each registration point in the first registration point set is outside the physiological and anatomical feature points on the surface of the object to be registered.
  • the judgment module is also used to determine whether the number of registrations for the current registration has reached the third iteration threshold; if the number of registrations for the current registration has not reached the third iteration threshold, the second image model is used as the first
  • the image model uses the second transformation matrix as the first transformation matrix and returns to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the acquisition module 11 is specifically configured to register the registration point set on the first image model surface with the first registration point set to obtain the first transformation matrix.
  • the acquisition module 11 is specifically configured to acquire a template image model, which has a template point set on its surface; and match the template image model with the first image model to map the template point set to the first image model. surface to obtain the registration point set of the first image model surface.
  • Each module in the above-mentioned registration device can be implemented in whole or in part by software, hardware and combinations thereof.
  • Each of the above modules may be embedded in or independent of the processor of the computer device in the form of hardware, or may be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
  • a computer device is provided.
  • the computer device may be a terminal, and its internal structure diagram may be as shown in FIG. 12 .
  • the computer device includes a processor, memory, communication interface, display screen and input device connected through a system bus.
  • the processor of the computer device is used to provide computing and control capabilities.
  • the memory of the computer device includes non-volatile storage media and internal memory.
  • the non-volatile storage medium stores operating systems and computer programs. This internal memory provides an environment for the execution of operating systems and computer programs in non-volatile storage media.
  • the communication interface of the computer device is used for wired or wireless communication with external terminals.
  • the wireless mode can be implemented through WIFI, mobile cellular network, NFC (Near Field Communication) or other technologies.
  • the computer program when executed by the processor implements a registration method.
  • the display screen of the computer device may be a liquid crystal display or an electronic ink display.
  • the input device of the computer device may be a touch layer covered on the display screen, or may be a button, trackball or touch pad provided on the computer device shell. , it can also be an external keyboard, trackpad or mouse, etc.
  • Figure 12 is only a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer equipment to which the solution of the present application is applied.
  • Specific computer equipment can May include more or fewer parts than shown, or combine certain parts, or have a different arrangement of parts.
  • a computer device including a memory and a processor.
  • a computer program is stored in the memory.
  • the processor executes the computer program, it implements the following steps: Obtaining the third value of the object to be registered used in the last registration. An image model and the first transformation matrix obtained from the last registration; for the current registration, the first image model is adjusted based on the first transformation matrix to obtain the second image model; the first registration of the surface of the object to be registered is The quasi-point set is projected onto the surface of the second image model to obtain a projection point set; the first registration point set and the projection point set are registered to obtain the second transformation matrix.
  • each projection point in the projection point set is the closest projection point to each corresponding registration point in the first registration point set.
  • the processor when the processor executes the computer program, the processor also implements the following steps: taking the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set as known quantities, and converting the second transformation As an unknown quantity, the matrix is substituted into the first objective function to solve the second transformation matrix; where, the optimization goal of the first objective function is: minimizing the pose gap between the first transformation point set and the first registration point set, The first conversion point set is obtained by converting the projection point set based on the second conversion matrix.
  • the processor also implements the following steps when executing the computer program: obtaining the registration weight corresponding to each registration point in the first registration point set; wherein, for each registration point in the first registration point set, the registration weight is Based on current allocation
  • the number of accurate registrations, the number of preset registrations and the pose difference between the registration points and the corresponding projection points are obtained; the registration weight corresponding to each registration point in the first registration point set, the registration weight of each registration point in the first registration point set are obtained
  • the pose coordinates of the accurate point and the pose coordinates of each projection point in the projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity and substituted into the second objective function to solve the second transformation matrix.
  • the processor executes the computer program, the following steps are also implemented: determine whether the number of registrations for the current registration has reached the first iteration threshold; if the number of registrations for the current registration has not reached the first iteration threshold, , then use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the processor executes the computer program, the following steps are also implemented: if the number of registrations for the current registration reaches the first iteration number threshold, the second image model is used as the first image model, and the second image model is used as the first image model.
  • the transformation matrix is used as the first transformation matrix
  • the second registration point set on the surface of the object to be registered is used as the first registration point set
  • the first image model is adjusted based on the first transformation matrix to obtain the second image model. Steps, until the number of registrations for the current registration reaches the second iteration number threshold; wherein each registration point in the second registration point set is a point other than the physiological and anatomical feature points on the surface of the object to be registered.
  • the processor when the processor executes the computer program, the processor also implements the following steps: determine whether the number of registrations for the current registration has reached the third iteration threshold; if the number of registrations for the current registration has not reached the third iteration threshold, , then use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the processor when the processor executes the computer program, the processor further implements the following steps: registering the registration point set on the first image model surface with the first registration point set to obtain the first transformation matrix.
  • the processor also implements the following steps when executing the computer program: acquiring a template image model with a template point set on its surface; matching the template image model with the first image model to map the template point set to The first image model surface is used to obtain a registration point set of the first image model surface.
  • a computer-readable storage medium is provided, with a computer program stored thereon.
  • the computer program When the computer program is executed by a processor, the following steps are implemented: Obtain the first image of the object to be registered used in the last registration. model and the first transformation matrix obtained from the last registration; for the current registration, the first image model is adjusted based on the first transformation matrix to obtain the second image model; the first registration point set on the surface of the object to be registered is Project to the surface of the second image model to obtain a projection point set; register the first registration point set and the projection point set to obtain a second transformation matrix.
  • each projection point in the projection point set is the closest projection point to each corresponding registration point in the first registration point set.
  • the following steps are also implemented: taking the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set as known quantities, and taking the second
  • the transformation matrix is used as an unknown quantity and is substituted into the first objective function to solve the second transformation matrix.
  • the optimization goal of the first objective function is to minimize the pose gap between the first transformation point set and the first registration point set.
  • the first conversion point set is obtained by converting the projection point set based on the second conversion matrix.
  • the following steps are also implemented: obtaining the registration weight corresponding to each registration point in the first registration point set; wherein, for each registration point in the first registration point set, the registration weight It is obtained based on the number of registrations for the current registration, the number of preset registrations, and the pose difference between the registration point and the projection point corresponding to the registration point; the first registration point is gathered into the registration weight corresponding to each registration point, and the first The pose coordinates and projection points of each registration point in the registration point set are The pose coordinates of each projection point are used as known quantities, and the second transformation matrix is used as an unknown quantity, and substituted into the second objective function to solve the second transformation matrix.
  • the following steps are also implemented: determine whether the number of registrations for the current registration has reached a first iteration number threshold; if the number of registrations for the current registration has not reached the first iteration number, threshold, then use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the following steps are also implemented: if the number of registrations for the current registration reaches a first iteration threshold, the second image model is used as the first image model, and the second image model is used as the first image model.
  • the second transformation matrix is used as the first transformation matrix
  • the second registration point set on the surface of the object to be registered is used as the first registration point set
  • the first image model is adjusted based on the first transformation matrix to obtain the second image model.
  • the following steps are also implemented: determine whether the number of registrations for the current registration has reached the third iteration number threshold; if the number of registrations for the current registration has not reached the third iteration number, threshold, then use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the following steps are also implemented: registering the registration point set on the surface of the first image model with the first registration point set to obtain the first transformation matrix.
  • the following steps are also implemented: obtaining a template image model, which has a template point set on its surface; matching the template image model with the first image model to map the template point set to the first image model surface, and obtain a registration point set of the first image model surface.
  • a computer program product including a computer program.
  • the computer program When executed by a processor, the computer program implements the following steps: obtaining the first image model of the object to be registered used in the last registration and the last registration. The first transformation matrix obtained accurately; for the current registration, the first image model is adjusted based on the first transformation matrix to obtain the second image model; the first registration point set on the surface of the object to be registered is projected to the second image model surface to obtain a projection point set; register the first registration point set and the projection point set to obtain the second transformation matrix.
  • each projection point in the projection point set is the closest projection point to each corresponding registration point in the first registration point set.
  • the following steps are also implemented: taking the pose coordinates of each registration point in the first registration point set and the pose coordinates of each projection point in the projection point set as known quantities, and taking the second
  • the transformation matrix is used as an unknown quantity and is substituted into the first objective function to solve the second transformation matrix.
  • the optimization goal of the first objective function is to minimize the pose gap between the first transformation point set and the first registration point set.
  • the first conversion point set is obtained by converting the projection point set based on the second conversion matrix.
  • the following steps are also implemented: obtaining the registration weight corresponding to each registration point in the first registration point set; wherein, for each registration point in the first registration point set, the registration weight It is obtained based on the number of registrations for the current registration, the number of preset registrations, and the pose difference between the registration point and the projection point corresponding to the registration point; the first registration point is gathered into the registration weight corresponding to each registration point, and the first The pose coordinates of each registration point in the registration point set and the pose coordinates of each projection point in the projection point set are used as known quantities, and the second transformation matrix is used as an unknown quantity and substituted into the second objective function to solve the second transformation matrix.
  • the following steps are also implemented: determine whether the number of registrations for the current registration has reached a first iteration number threshold; if the number of registrations for the current registration has not reached the first iteration number, threshold, then use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the following steps are also implemented: if the number of registrations for the current registration reaches a first iteration threshold, the second image model is used as the first image model, and the second image model is used as the first image model.
  • the second transformation matrix is used as the first transformation matrix
  • the second registration point set on the surface of the object to be registered is used as the first registration point set
  • the first image model is adjusted based on the first transformation matrix to obtain the second image model.
  • the following steps are also implemented: determine whether the number of registrations for the current registration has reached the third iteration number threshold; if the number of registrations for the current registration has not reached the third iteration number, threshold, then use the second image model as the first image model, use the second transformation matrix as the first transformation matrix, and return to the step of adjusting the first image model based on the first transformation matrix to obtain the second image model.
  • the following steps are also implemented: registering the registration point set on the surface of the first image model with the first registration point set to obtain the first transformation matrix.
  • the following steps are also implemented: obtaining a template image model, which has a template point set on its surface; matching the template image model with the first image model to map the template point set to the first image model surface, and obtain a registration point set of the first image model surface.
  • the computer program can be stored in a non-volatile computer-readable storage.
  • the computer program when executed, may include the processes of the above method embodiments.
  • Any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory.
  • Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetic variable memory (Magnetoresistive Random Access Memory (MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc.
  • Volatile memory may include random access memory (Random Access Memory, RAM) or external cache memory, etc.
  • RAM Random Access Memory
  • RAM random access memory
  • RAM Random Access Memory
  • the databases involved in the various embodiments provided in this application may include at least one of a relational database and a non-relational database.
  • Non-relational databases may include blockchain-based distributed databases, etc., but are not limited thereto.
  • the processors involved in the various embodiments provided in this application may be general-purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to this.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Multimedia (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Optical Recording Or Reproduction (AREA)
  • Testing Of Coins (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)
  • Image Analysis (AREA)

Abstract

本申请涉及一种配准方法、装置、计算机设备和可读存储介质，该方法通过获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵；对于当前次配准，基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；将待配准对象表面的第一配准点集投影至第二影像模型表面得到投影点集；将第一配准点集与投影点集进行配准，得到第二转换矩阵。

Description

配准方法、装置、计算机设备和可读存储介质
本申请要求于2022年6月30日提交的申请号为202210758534.5、名称为“配准方法、装置、计算机设备和可读存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及医学图像处理技术领域，特别是涉及一种配准方法、装置、计算机设备和可读存储介质。
背景技术
随着人均寿命的增长，人口老龄化进程的加快，骨科病的发病率也在不断增长，骨科手术需求量逐年上升。在手术过程中，需要保证骨头的三维模型与待配准对象的真实骨头相对应。
发明内容
本申请提供一种配准方法、装置、计算机设备和可读存储介质。
第一方面，本申请一个实施例提供一种配准方法，该方法包括：获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵；执行当前次配准，包括：基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；将第一配准点集与投影点集进行配准，得到第二转换矩阵。
在其中一个实施例中，投影点集中各投影点分别为距离第一配准点集中各对应配准点最近的投影点。
在其中一个实施例中，将所述第一配准点集与所述投影点集进行配准，得到第二转换矩阵，包括：基于第一配准点集中各配准点的位姿坐标和所述投影点集中各投影点的位姿坐标确定所述第二转换矩阵。在其中一个实施例中，基于第一配准点集中各配准点的位姿坐标和所述投影点集中各投影点的位姿坐标确定所述第二转换矩阵，包括：将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；其中，第一目标函数配置为最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进行转换后得到的。
在其中一个实施例中，将第一配准点集与投影点集进行配准，得到第二转换矩阵，包括：获取第一配准点集中各配准点对应的配准权重；其中，针对第一配准点集中的每个配准点，配准权重是基于当前次配准的配准次数、预设配准次数以及配准点与配准点对应投影点之间的位姿中的至少一个差距得到；将第一配准点集中各配准点对应的配准权重、第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第二目标函数中，以求解第二转换矩阵，所述第二目标函数配置为最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进 行转换后得到的。
在其中一个实施例中，第一配准点集中各配准点为待配准对象表面的生理解剖特征点，在将所述第一配准点集与投影点集进行配准，得到第二转换矩阵之后，配准方法还包括：判断当前次配准的配准次数是否达到第一迭代次数阈值；若当前次配准的配准次数未到达第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在其中一个实施例中，该配准方法还包括：若当前次配准的配准次数达到第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第一配准点集，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤，直至当前次配准的配准次数达到第二迭代次数阈值；其中，第二配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点。
在其中一个实施例中，第一配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点，在将所述第一配准点集与投影点集进行配准，得到第二转换矩阵之后，该配准方法还包括：判断当前次配准的配准次数是否达到第三迭代次数阈值；若当前次配准的配准次数未达到第三迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在其中一个实施例中，获取上一次配准得到的第一转换矩阵，包括：将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵。
在其中一个实施例中，获取第一影像模型表面的配准点集的方法包括：获取模板影像模型，模板影像模型表面具有模板点集；将模板影像模型与第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集。
在其中一个实施例中，该配准方法还包括：判断所述当前次配准得到的第二转换矩阵是否达到预设转换矩阵阈值，若所述第二转换矩阵未达到预设转换矩阵阈值，则将所述第二影像模型作为第一影像模型，将所述第二转换矩阵作为第一转换矩阵，返回执行所述基于所述第一转换矩阵对所述第一影像模型进行调整，得到第二影像模型的步骤，直至所述当前次配准得到的第二转换矩阵达到预设转换矩阵阈值。
在其中一个实施例中，基于所述第二转换矩阵中位移大小和旋转角度对所述第二影像模型进行移动和旋转，得到匹配影像模型。
第二方面，本申请一个实施例提供一种配准装置，包括：获取模块，用于获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵；调整模块，用于对于当前次配准，基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；投影模块，用于将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；配准模块，用于将第一配准点集与投影点集进行配准，得到第二转换矩阵。
第三方面，本申请一个实施例提供一种计算机设备，包括存储器和处理器，存储器存储有计算机程序，处理器执行计算机程序时实现如上述第一方面提供的配准方法的步骤。
第四方面，本申请一个实施例还提供一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现如上述第一方面提供的配准方法的步骤。
第五方面，本申请一个实施例还提供一种计算机程序产品，包括计算机程序，计算机程序被处理器执行时实现如上述第一方面提供的方法的步骤。
本发明的各个实施例的细节将在下面的附图和描述中进行说明。根据说明书、附图以及权利要求书的记载，本领域技术人员将容易理解本发明的其它特征、解决的问题以及有益效果。
附图说明
为了更清楚地说明本申请实施例或传统技术中的技术方案，下面将对实施例或传统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，对于本领域不同技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为一个实施例提供的配准方法的应用环境图；
图2为一个实施例提供的配准方法的步骤流程示意图；
图3为另一个实施例提供的配准方法的步骤流程示意图；
图4为另一个实施例提供的配准方法的步骤流程示意图；
图5为另一个实施例提供的配准方法的步骤流程示意图；
图6为另一个实施例提供的配准方法的步骤流程示意图；
图7为一个实施例提供的粗配准点的分布示意图；
图8为一个实施例提供的精配准点的分布示意图；
图9为另一个实施例提供的配准方法的步骤流程示意图；
图10为另一个实施例提供的配准方法的步骤流程示意图；
图11为一个实施例提供的配准装置的结构示意图；
图12为本申请一个实施例提供的计算机设备的结构示意图。
具体实施方式
为使本申请的上述目的、特征和优点能够更加明显易懂，下面结合附图对本申请的具体实施方式做详细的说明。在下面的描述中阐述了很多具体细节以便于充分理解本申请。但是本申请能够以很多不同于在此描述的其它方式来实施，本领域技术人员可以在不违背本申请内涵的情况下做类似改进，因此本申请不受下面公开的具体实施例的限制。
本文中为部件所编序号本身，例如“第一”、“第二”等，仅用于区分所描述的对象，不具有任何顺序或技术含义。
本申请实施例提供的配准方法，可以应用于如图1所示的应用环境中，该应用环境包括终端100和医学扫描设备200。其中，终端100可以通过网络与医学扫描设备200进行通信。终端100可以但不限于是各种个人计算机、笔记本电脑机和平板电脑。医学扫描设备200可以但不限于是CT(Computed Tomography，即电子计算机断层扫描)设备和PET(Positron Emission Computed Tomography，正电子发射型计算机断层显像)。
下面以具体的实施例对本申请的技术方案以及本申请的技术方案如何解决技术问题进行详细说明。下面这几个具体的实施例可以相互结合，对于相同或相似的概念或过程可能在某些实施例中不再赘述。下面将结合附图，对本申请的实施例进行描述。
在一个实施例中，如图2所示，提供了一种配准方法，以该方法应用于图1中的终端为例进行说明，包括以下步骤：
步骤200、获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵。
上一次配准所使用的待配准对象的第一影像模型是指在上一次配准过程中与待配准对象进行配准的影像模型。第一转换矩阵用于表征上一次配准确定的待配准对象的影像空间和物理空间之间的转换关系。在终端进行当前次配准前，先获取上一次配准所使用的待配准对象的第一影像模型，以及上一次配准得到的第一转换矩阵。本实施例对上一次配准所使用的具体方法不作限制，只要能够实现其功能即可。
执行当前次配准，包括步骤210-230。
步骤210、基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型。
终端在进行当前次配准时，根据上一次配准得到的第一转换矩阵对第一影像模型进行调整，能够得到第二影像模型。第一转换矩阵中包括需要移动的位移大小和需要转动的旋转角度，也就是说，终端按照第一转换矩阵中位移大小和旋转角度将第一影像模型进行移动和旋转，就能够得到第二影像模型。
步骤220、将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集。
待配准对象表面的第一配准点集中的各配准点可以是工作人员预先在待配准对象表面设置好的配准点。本实施例对第一配准点集中各配准点的具体数量，以及各配准点的具体位置不作限制，只要能够实现其功能即可。
终端在得到第二影像模型后，将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集。换句话说，针对待配准对象表面的第一配准点集中的每个配准点，将该配准点投影至第二影像模型表面，能够得到该配准点在第二影像模型表面上对应的投影点，从而能够得到第一配准点集中各配准点在第二影像模型表面上对应的各投影点，即投影点集。
步骤230、将第一配准点集与投影点集进行配准，得到第二转换矩阵。
终端在得到投影点集后，将待配准对象的第一配准点集与投影点集进行配准，得到第二转换矩阵。第二转换矩阵用于表征当前次配准确定的待配准对象的影像空间与物理空间之间的转换关系。第二转换矩阵包括需要移动的位移大小和需要转动的旋转角度。本实施例对将第一配准点集与投影点集进行配准的具体过程不作限制，只要能够实现其功能即可。
在一个可选的实施例中，终端在得到第二转换矩阵后，按照第二转换矩阵中位移大小和旋转角度将第二影像模型进行移动和旋转，就能够得到与待配准对象匹配的影像模型。
本申请实施例提供的配准方法通过获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的转换矩阵；对于当前次配准，基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；将第一配准点集与投影点集进行配准，得到第二转换矩阵。本实施例使用第一配准点集投影至第二影像模型表面的投影点集与第一配准点集进行配准，与将第二影像模型上的配准点集直接与第一配准点集进行配准相比，能够避免第二影像模型上的配准点不准确，导致最终确定的第二转换矩阵不准确，即能够减少对第二影像模型上的配准点的依赖，从而能够提高配准的准确性。
另外，使用本申请提供的配准方法利用少量的配准点就能够达到高精度的配准结果。
在一个实施例中，投影点集中各投影点分别为距离第一配准点集中各对应配准点最近的投影点。也就是说，第一配准点集中每个配准点投影至第二影像模型表面的有多个初始投影 点，投影点集中与该配准点对应的投影点是多个初始投影点中与配准点距离最近的投影点。
在本实施例中，投影点集中各投影点设置为距离第一配准点集中各对应的配准点最近的投影点，使用该投影点集与第一配准点集进行配准，能够提高配准的准确性，从而能够得到更加准确的第二转换矩阵。
在一个实施例中，涉及将第一配准点集与投影点集进行配准，得到第二转换矩阵的一种可能的实现方式包括：
将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；其中，第一目标函数的优化目标为：最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进行转换后得到的。
假设，第一配准点集中第i个配准点的位姿坐标表示为qi，投影点集中与第i个配准点对应的投影点的位姿坐标表示为pi’，第二转换矩阵中包括位移t2和旋转角度R2。第一转换点集可以表示为：(R2*p'i+t2)，则第一目标函数可以表示为：其中，N1为第一配准点集中配准点的数量。终端将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标代入第一目标函数中，能够得到第二转换矩阵中的位移和旋转角度，则能够得到第二转换矩阵。
在本实施例中，使用第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标对第一目标函数求解，就能够得到第一转换矩阵。这样计算第二转换矩阵的方法逻辑简单，容易实现，且计算效率高。
在一个实施例中，如图3所示，涉及第一配准点集与投影点集进行配准，得到第二转换矩阵的一种可能的实现方式，包括：
步骤300、获取第一配准点集中各配准点对应的配准权重；其中，针对第一配准点集中的每个配准点，配准权重是基于当前次配准的配准次数、预设配准次数以及配准点与配准点对应投影点之间的位姿差距得到。
在其中一个实施例中，针对第一配准点集中的每个配准点，配准权重可以是基于当前次配准的配准次数、预设配准次数以及配准点与配准点对应投影点之间的位姿差距中的至少一个得到。例如，基于配准点与配准点对应投影点之间的位姿差距获得该配准点的配准权重。
在配准过程中，配准次数不同，各配准点对应的配准权重不同。第一配准点集中各配准点对应的配准权重与当前次配准的配准次数、工作人员预先设置的预设配准次数，以及该配准点与该配准点对应投影点之间的位姿差距相关。本实施例对确定配准权重的具体方法不作限制，只要能够实现其功能即可。终端获取第一配准点集中各配准点对应的配准权重。
在一个可选的实施例中，配准权重可以使用如下公式计算：
其中，l表示当前次配准的配准次数，M为预设配准次数，wi为第i个配准点对应的配准权重，||p'i-qi||为第i个配准点对应的投影点pi’与第i个配准 点qi之间的位姿差距。
步骤310、将第一配准点集中各配准点对应的配准权重、第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第二目标函数中，以求解第二转换矩阵。其中，第二目标函数的优化目标为：最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进行转换后得到的。
第二目标函数可以表示为：终端将已知的第一配准点集中各配准点对应的配准权重、第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标代入第二目标函数，能够求解得到第二转换矩阵中的位移和旋转角度，则能够得到第二转换矩阵。
在本实施例中，为了减少配准过程中携带噪声的各配准点对配准结果的影响，通过在配准过程中增加配准权重，随着配准次数的增加，噪声越大的配准点对应的权重越小，这样能够降低噪声对配准结果的影响，从而能够提高配准的准确性。并且，本实施例中第二转换矩阵的计算方法逻辑简单，容易实现，且计算效率高。
在一个实施例中，在第一配准点集中各配准点为待配准对象表面的生理解剖特征点，即粗配准点。粗配准点的设置方法可以采用探针的形式，探针的定位形式可以分为视觉追踪或者磁导航定位等方式。本实施例对此不作限制。基于此，如图4所示，在将第一配准点集和投影点集进行配准，得到第二转换矩阵后，该配准方法的步骤还包括：
步骤400、判断当前次配准的配准次数是否达到第一迭代次数阈值。
第一迭代次数阈值可以是工作人员预先设置并存储在终端中的。终端在得到第二转换矩阵后，获取得到第二转换矩阵的当前次配准的配准次数，以及第一迭代次数阈值，并对比当前次配准的配准次数和第一迭代次数阈值，判断当前次配准的配准次数是否达到第一迭代次数阈值，即，当前次配准的配准次数是否等于第一迭代次数阈值。
步骤410、若当前次配准的配准次数未到达第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
若终端确定当前次配准的配准次数未到达第一迭代次数阈值，即当前次配准的配准次数小于第一迭代次数阈值，则将当前次配准使用的模型(第二影像模型)作为第一影像模型，将当前次配准得到的第二转换矩阵作为第一转换矩阵，返回执行步骤210、步骤220、步骤230、步骤400和步骤410。
在本实施例中，在当前次配准的配准次数未达到第一迭代次数阈值时，将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行步骤210、步骤220、步骤230、步骤400和步骤410，基于粗配准点实现多次配准，这样能够提高配准的准确度。
在一个可选的实施例中，当前次配准停止的条件还包括当前次配准得到的第二转换矩阵是否达到预设转换矩阵阈值。也就是说，若当前次配准得到的第二转换矩阵未达到预设转换矩阵阈值；则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行步骤210、步骤220、步骤230、步骤400和步骤410，直至当前次配准得到的第二转换 矩阵达到预设转换阈值。预设转换阈值可以包括，例如，第二转换矩阵中的位移和旋转角度的阈值。
在一个可选的实施例中，当前次配准停止的条件可以包括当前次配准的配准次数是否达到第一迭代次数阈值或当前次配准得到的第二转换矩阵是否达到预设转换矩阵阈值。也就是说，只要这两个当前次配准停止的条件中有一个满足，配准就停止。
在另一个可选的实施例中，当前次配准停止的条件可以同时包括当前次配准的配准次数是否达到第一迭代次数阈值以及当前次配准得到的第二转换矩阵是否达到预设转换矩阵阈值。仅当两项条件同时满足时，停止配准。
在一个实施例中，请继续参见图4，在一个实施例中，该配准方法的步骤还包括：
步骤420、若当前次配准的配准次数达到第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第一配准点集，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤，直至当前次配准的配准次数达到第二迭代次数阈值；其中，第二配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点。
待配准对象表面的第二配准点集是待配准对象表面的生理解剖特征点之外的点，即，第二配准点集中的各配准点与第一配准点集中的各配准点不同。第二配准点集中的各配准点也可以称为精配准点，精配准点可以采用探针的形式，探针的定位形式可以分为视觉追踪或者磁导航定位等方式。本实施例对此不作限制。
终端在确定当前次配准的配准次数达到第一迭代次数阈值，即当前次配准次数等于第一迭代次数阈值，则将当前次配准使用的模型(第二影像模型)作为第一影像模型，将当前次配准得到的第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第一配准点集，返回执行步骤210、步骤220、步骤230、步骤400和步骤410，直到确定当前次配准次数达到第二迭代次数阈值为止。
在本实施例中，在当前次配准的配准次数到达第一迭代次数，即使用第一配准点集进行了多次配准(即使用粗配准点进行粗配准)后，再使用第二配准点集进行多次配准(即使用精配准点进行精配准)，能够提高配准的准确度。
在一个实施例中，在第一配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点，即，精配准点。对于精配准点的具体描述可以参考上述实施例中的描述，在此不再赘述。基于此，如图5所示，在将第一配准点集与投影点集进行配准，得到第二转换矩阵之后，配准方法的步骤还包括：
步骤500、判断当前次配准的配准次数是否达到第三迭代次数阈值。
第三迭代次数阈值可以是由工作人员预先设置并存储在终端的存储器中的。第三迭代次数阈值与第一迭代次数阈值可以相同，也可以不同。终端在得到第二转换矩阵后，获取得到第二转换矩阵的当前次配准的配准次数，以及第三迭代次数阈值，并对比当前次配准的配准次数和第三迭代次数阈值，判断当前次配准的配准次数是否达到第三迭代次数阈值，即，当前次配准的配准次数是否等于第三迭代次数阈值。
步骤510、若当前次配准的配准次数未达到第三迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
若终端确定当前次配准的配准次数未达到第三迭代次数阈值，则将当前次使用的模型(第二影像模型)作为第一影像模型，将当前次配准得到的第二转换矩阵作为第一转换矩阵，返回执行步骤210、步骤220、步骤230、步骤500和步骤510。
在本实施例中，在当前次配准的配准次数未达到第三迭代次数阈值时，将第二影像模型作为第一影像模型，第二转换矩阵作为第一转换矩阵，返回执行步骤210、步骤220、步骤230、步骤500和步骤510，基于精配准点实现多次配准，这样能够提高配准的准确度。
在一个实施例中，在上一次配准为初始配准，即第一次配准使用的第一影像模型是通过CT设备对待配准对象进行扫描得到的影像数据，并将该影像数据发送至终端，终端利用三维表面重建算法，对影像数据进行三维重建，得到待配准对象的影像模型。基于此，获取上一次配准得到的第一转换矩阵的一种可能的实现方式包括：
将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵。
终端在获取第一影像模型后，通过第一影像模型表面的配准点集在待配准对象表面选取对应的点组成第一配准点集，将第一影像模型表面的配准点集与第一配准点集进行配准，能够得到第一转换矩阵。本实施例对将第一影像模型表面的配准点集与第一配准点集进行配准的具体方法不作限制，只要能够得到第一转换矩阵即可。
在一个可选的实施例中，终端可以采用点对配准算法将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵。
在一个可选的实施例中，终端可以将第一影像模型表面的配准点集中各配准点的位姿坐标与第一配准点集中各配准点的位姿坐标作为已知量，将第一转换矩阵作为未知量，代入第三目标函数，以求解第一转换矩阵。
假设，第一影像模型表面的配准点集中的第i个配准点的位姿坐标为pi，第一配准点集中的第i个配准点的位姿坐标为qi，第一转换矩阵中包括位移t1和旋转角度R1。第三目标函数可以表示为：终端将第一影像模型表面的配准点集中各配准点的位姿坐标与第一配准点集中各配准点的位姿坐标代入第三目标函数中，能够得到第一转换矩阵中的位移和旋转角度，则能够得到第一转换矩阵。
在本实施例中，提供了一种通过将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵的配准方法，该方法简单易懂，且容易实现。并且，在使用本实施例提供的配准方法先进行一次配准，再进行当前次配准，能够提高当前次配准的效率和准确性。
在一个实施例中，请参见图6，涉及获取第一影像模型表面的配准点集的一种可能的实现方式，该实现方式包括：
步骤600、获取模板影像模型，模板影像模型表面具有模板点集。
模板影像模型是指进行配准时的通用模型。例如，通用的骨盆三维数字模型。模板影像模型可以是工作人员预先存储在终端的存储器中的。终端直接从存储器中获取模板影像模型。模板影像模型表面具有模板点集，模板点集可以包括生理解剖特征点(粗配准点)，以及生理解剖特征点之外的点(精配准点)。生理解剖特征点的设置便于工作人员在待配准对象上找到与其对应的点。通常情况下，模板点集中的生理解剖特征点的数量为至少3个不共线的 点。
具体的，对于骨盆影像，粗配准点为骨盆生理解剖特征点，精配准点通设置在待处理的区域的骨骼表面，精配准点的数量至少为3个，最优为10个以上。
步骤610、将模板影像模型与第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集。
终端在获取模板影像模型后，将该模板影像模型与第一影像模型进行匹配，能够将模板影像模型上的模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集。
在本实施例中，通过将模型影像模型与第一影像模型匹配，将模板点集映射至第一影像模型表面，得到第一影像模型的配准点集，即能够得到待检测对象的个体化配准点，从而能够更加直观的引导工作人员确定待检测对象表面与第一影像模型表面的配准点集对应的配准点集。
在一个具体的实施例中，第一影像模型表面的粗配准点的分布如图7所示，第一影像模型表面的精配准点的分布如图8所示。
请参见图9，在一个实施例中，本申请提供一种配准方法，该方法的步骤包括：
步骤900、获取模板影像模型；模板影像模型表面具有模板点集；其中，模板点集包括第一模板点集和第二模板点集，第一模板点集中各模板点为通用的模板对象的影像模型表面的生理解剖特征点，第二模板点集中各模板点为通用的模板对象的影像模型表面的生理解剖特征点之外的点；
步骤910、将模板影像模型与待配准对象的第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集；其中，第一影像模型表面的配准点集包括与第一模板点集对应的第一子配准点集，以及与第二模板点集对应的第二子配准点集，第一子配准点集中各子配准点为待配准对象的影像模型表面的生理解剖特征点，第二子配准点集中各子配准点为待配准对象的影像模型表面的生理解剖特征点之外的点；
步骤920、通过第一影像模型表面的第一子配准点集在待配准对象表面选取对应的点组成第一配准点集，通过第一影像模型表面的第二子配准点集在待配准对象表面选取对应的点组成第二配准点集，第一配准点集为在待配准对象表面选取的生理解剖特征点，第二配准点集为在待配准对象表面选取的生理解剖特征点之外的点；
步骤930、将第一影像模型表面的第一子配准点集与待配准对象表面的第一配准点集进行配准，得到第一转换矩阵；
步骤940、基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；
步骤950、将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；
步骤960、将第一配准点集和投影点集进行配准，得到第二转换矩阵；
将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；
步骤970、判断当前次配准的配准次数是否达到第一迭代次数阈值；
步骤980、若当前次配准的配准次数未达到第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行步骤940-步骤980；
步骤990、若当前次配准的配准次数达到第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第 一配准点集，返回执行步骤940-步骤990，直至当前次配准的配准次数达到第二迭代次数阈值；其中，第二配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点。
请参见图10，在一个实施例中，本申请提供一种配准方法，该方法的步骤包括：
步骤101、获取模板影像模型；模板影像模型表面具有模板点集；其中，模板点集包括第一模板点集和第二模板点集，第一模板点集中各模板点为通用的模板对象的影像模型表面的生理解剖特征点，第二模板点集中各模板点为通用的模板对象的影像模型表面的生理解剖特征点之外的点；
步骤102、将模板影像模型与待配准对象的第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集；其中，第一影像模型表面的配准点集包括与第一模板点集对应的第一子配准点集，以及与第二模板点集对应的第二子配准点集，第一子配准点集中各子配准点为待配准对象的影像模型表面的生理解剖特征点，第二子配准点集中各子配准点为待配准对象的影像模型表面的生理解剖特征点之外的点；
步骤103、通过第一影像模型表面的第一子配准点集在待配准对象表面选取对应的点组成第一配准点集，通过第一影像模型表面的第二子配准点集在待配准对象表面选取对应的点组成第二配准点集，第一配准点集为在待配准对象表面选取的生理解剖特征点，第二配准点集为在待配准对象表面选取的生理解剖特征点之外的点；
步骤104、将第一影像模型表面的第一子配准点集与待配准对象表面的第一配准点集进行配准，得到第一转换矩阵；
步骤105、基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；
步骤106、将待配准对象表面的第一配准点集投影至第二影像模型表面，得到第一投影点集；
步骤107、将第一配准点集和第一投影点集进行配准，得到第二转换矩阵；
将第一配准点集中各配准点的位姿坐标和第一投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；
步骤108、判断当前次配准的配准次数是否达到第一迭代次数阈值；
步骤109、若当前次配准的配准次数未达到第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行步骤105-步骤109；
步骤110、若当前次配准的配准次数达到第一迭代次数阈值，则进行下一次配准：
步骤111、对于下一次配准，基于第二转换矩阵对第一影像模型进行调整，得到第三影像模型；将待配准对象表面的第二配准点集投影至第三影像模型表面，得到第二投影点集；
步骤112、获取第二配准点集中各配准点对应的配准权重；其中，针对第二配准点集中的每个配准点，配准权重是基于下一次配准的配准次数与第一迭代次数阈值的差值、预设配准次数以及配准点与配准点对应的第二投影点之间的位姿差距得到；其中，预设配准次数等于第二迭代次数阈值与第一迭代次数阈值之间的差值；
步骤113、将第二配准点集中各配准点对应的配准权重、第二配准点集中各配准点的位姿坐标和第二投影点集中各投影点的位姿坐标作为已知量，将第三转换矩阵作为未知量，代入第二目标函数中，以求解第三转换矩阵。
步骤114、判断下一次配准的配准次数是否达到第二迭代次数阈值；
步骤115、若否，将第三影像模型作为第一影像模型，将第三转换矩阵作为第二转换矩 阵，返回执行步骤111-步骤115。
应该理解的是，虽然图中的流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，这些步骤可以以其它的顺序执行。而且，图中的至少一部分步骤可以包括多个子步骤或者多个阶段，这些子步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，这些子步骤或者阶段的执行顺序也不必然是依次进行，而是可以与其它步骤或者其它步骤的子步骤或者阶段的至少一部分轮流或者交替地执行。
基于同样的发明构思，本申请实施例还提供了一种用于实现上述所涉及的配准方法的配准装置。该装置所提供的解决问题的实现方案与上述方法中所记载的实现方案相似，故下面所提供的一个或多个配准装置实施例中的具体限定可以参见上文中对于配准方法的限定，在此不再赘述。
请参见图11，本申请一个实施例提供一种配准装置10，该装置包括获取模块11、调整模块12、投影模块13和配准模块14。其中，获取模块11用于获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵；调整模块12用于对于当前次配准，基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；投影模块13用于将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；配准模块14用于将第一配准点集与投影点集进行配准，得到第二转换矩阵。
在一个实施例中，投影点集中各投影点分别为距离第一配准点集中各对应配准点最近的投影点。
在一个实施例中，配准模块14具体用于将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；其中，第一目标函数的优化目标为：最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将所述投影点集进行转换后得到的。
在一个实施例中，配准模块14具体还用于获取第一配准点集中各配准点对应的配准权重；其中，针对第一配准点集中的每个配准点，配准权重是基于当前次配准的配准次数、预设配准次数以及配准点与配准点对应投影点之间的位姿差距得到；将第一配准点集中各配准点对应的配准权重、第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第二目标函数中，以求解第二转换矩阵。
在一个实施例中，第一配准点集中各配准点为待配准对象表面的生理解剖特征点，配准装置10还包括判断模块。判断模块用于判断当前次配准的配准次数是否达到第一迭代次数阈值；若当前次配准的配准次数未到达第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，判断模块还用于若当前次配准的配准次数达到第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第一配准点集，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤，直至当前次配准的配准次数达到第二迭代次数阈值；其中，第二配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点。
在一个实施例中，第一配准点集中各配准点为待配准对象表面的生理解剖特征点之外的 点，判断模块还用于判断当前次配准的配准次数是否达到第三迭代次数阈值；若当前次配准的配准次数未达到第三迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，获取模块11具体用于将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵。
在一个实施例中，获取模块11具体还用于获取模板影像模型，模板影像模型表面具有模板点集；将模板影像模型与第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集。
上述配准装置中的中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
在一个实施例中，提供了一种计算机设备，该计算机设备可以是终端，其内部结构图可以如图12所示。该计算机设备包括通过系统总线连接的处理器、存储器、通信接口、显示屏和输入装置。其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作系统和计算机程序。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该计算机设备的通信接口用于与外部的终端进行有线或无线方式的通信，无线方式可通过WIFI、移动蜂窝网络、NFC(近场通信)或其他技术实现。该计算机程序被处理器执行时以实现一种配准方法。该计算机设备的显示屏可以是液晶显示屏或者电子墨水显示屏，该计算机设备的输入装置可以是显示屏上覆盖的触摸层，也可以是计算机设备外壳上设置的按键、轨迹球或触控板，还可以是外接的键盘、触控板或鼠标等。
本领域技术人员可以理解，图12中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的计算机设备的限定，具体的计算机设备可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
在一个实施例中，提供了一种计算机设备，包括存储器和处理器，存储器中存储有计算机程序，处理器执行计算机程序时实现以下步骤：获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵；对于当前次配准，基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；将第一配准点集与投影点集进行配准，得到第二转换矩阵。
在一个实施例中，投影点集中各投影点分别为距离第一配准点集中各对应配准点最近的投影点。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；其中，第一目标函数的优化目标为：最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进行转换后得到的。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：获取第一配准点集中各配准点对应的配准权重；其中，针对第一配准点集中的每个配准点，配准权重是基于当前次配 准的配准次数、预设配准次数以及配准点与配准点对应投影点之间的位姿差距得到；将第一配准点集中各配准点对应的配准权重、第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第二目标函数中，以求解第二转换矩阵。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：判断当前次配准的配准次数是否达到第一迭代次数阈值；若当前次配准的配准次数未到达第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：若当前次配准的配准次数达到第一迭代次数阈值，则将第二影像模型作为所述第一影像模型，将第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第一配准点集，返回执行基于所述第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤，直至当前次配准的配准次数达到第二迭代次数阈值；其中，第二配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：判断当前次配准的配准次数是否达到第三迭代次数阈值；若当前次配准的配准次数未达到第三迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于所述第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：获取模板影像模型，模板影像模型表面具有模板点集；将模板影像模型与第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集。
在一个实施例中，提供了一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现以下步骤：获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵；对于当前次配准，基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；将第一配准点集与投影点集进行配准，得到第二转换矩阵。
在一个实施例中，投影点集中各投影点分别为距离第一配准点集中各对应配准点最近的投影点。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；其中，第一目标函数的优化目标为：最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进行转换后得到的。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：获取第一配准点集中各配准点对应的配准权重；其中，针对第一配准点集中的每个配准点，配准权重是基于当前次配准的配准次数、预设配准次数以及配准点与配准点对应投影点之间的位姿差距得到；将第一配准点集中各配准点对应的配准权重、第一配准点集中各配准点的位姿坐标和投影点集中 各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第二目标函数中，以求解第二转换矩阵。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：判断当前次配准的配准次数是否达到第一迭代次数阈值；若当前次配准的配准次数未到达第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：若当前次配准的配准次数达到第一迭代次数阈值，则将第二影像模型作为所述第一影像模型，将第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第一配准点集，返回执行基于所述第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤，直至当前次配准的配准次数达到第二迭代次数阈值；其中，第二配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：判断当前次配准的配准次数是否达到第三迭代次数阈值；若当前次配准的配准次数未达到第三迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于所述第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：获取模板影像模型，模板影像模型表面具有模板点集；将模板影像模型与第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集。
在一个实施例中，提供了一种计算机程序产品，包括计算机程序，计算机程序被处理器执行时实现以下步骤：获取上一次配准所使用的待配准对象的第一影像模型以及上一次配准得到的第一转换矩阵；对于当前次配准，基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型；将待配准对象表面的第一配准点集投影至第二影像模型表面，得到投影点集；将第一配准点集与投影点集进行配准，得到第二转换矩阵。
在一个实施例中，投影点集中各投影点分别为距离第一配准点集中各对应配准点最近的投影点。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：将第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第一目标函数中，以求解第二转换矩阵；其中，第一目标函数的优化目标为：最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进行转换后得到的。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：获取第一配准点集中各配准点对应的配准权重；其中，针对第一配准点集中的每个配准点，配准权重是基于当前次配准的配准次数、预设配准次数以及配准点与配准点对应投影点之间的位姿差距得到；将第一配准点集中各配准点对应的配准权重、第一配准点集中各配准点的位姿坐标和投影点集中各投影点的位姿坐标作为已知量，将第二转换矩阵作为未知量，代入第二目标函数中，以求解第二转换矩阵。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：判断当前次配准的配准次数是否达到第一迭代次数阈值；若当前次配准的配准次数未到达第一迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：若当前次配准的配准次数达到第一迭代次数阈值，则将第二影像模型作为所述第一影像模型，将第二转换矩阵作为第一转换矩阵，将待配准对象表面的第二配准点集作为第一配准点集，返回执行基于所述第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤，直至当前次配准的配准次数达到第二迭代次数阈值；其中，第二配准点集中各配准点为待配准对象表面的生理解剖特征点之外的点。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：判断当前次配准的配准次数是否达到第三迭代次数阈值；若当前次配准的配准次数未达到第三迭代次数阈值，则将第二影像模型作为第一影像模型，将第二转换矩阵作为第一转换矩阵，返回执行基于所述第一转换矩阵对第一影像模型进行调整，得到第二影像模型的步骤。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：将第一影像模型表面的配准点集与第一配准点集进行配准，得到第一转换矩阵。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：获取模板影像模型，模板影像模型表面具有模板点集；将模板影像模型与第一影像模型进行匹配，以将模板点集映射至第一影像模型表面，得到第一影像模型表面的配准点集。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是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非易失性计算机可读取存储介质中，该计算机程序在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、数据库或其它介质的任何引用，均可包括非易失性和易失性存储器中的至少一种。非易失性存储器可包括只读存储器(Read-Only Memory，ROM)、磁带、软盘、闪存、光存储器、高密度嵌入式非易失性存储器、阻变存储器(ReRAM)、磁变存储器(Magnetoresistive Random Access Memory，MRAM)、铁电存储器(Ferroelectric Random Access Memory，FRAM)、相变存储器(Phase Change Memory，PCM)、石墨烯存储器等。易失性存储器可包括随机存取存储器(Random Access Memory，RAM)或外部高速缓冲存储器等。作为说明而非局限，RAM可以是多种形式，比如静态随机存取存储器(Static Random Access Memory，SRAM)或动态随机存取存储器(Dynamic Random Access Memory，DRAM)等。本申请所提供的各实施例中所涉及的数据库可包括关系型数据库和非关系型数据库中至少一种。非关系型数据库可包括基于区块链的分布式数据库等，不限于此。本申请所提供的各实施例中所涉及的处理器可为通用处理器、中央处理器、图形处理器、数字信号处理器、可编程逻辑器、基于量子计算的数据处理逻辑器等，不限于此。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本申请专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在 不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请的保护范围应以所附权利要求为准。

Claims (16)

  1. 一种配准方法，包括：
    获取上一次配准所使用的待配准对象的第一影像模型以及所述上一次配准得到的第一转换矩阵；
    执行当前次配准，包括：
    基于所述第一转换矩阵对所述第一影像模型进行调整，得到第二影像模型；
    将所述待配准对象表面的第一配准点集投影至所述第二影像模型表面，得到投影点集；
    将所述第一配准点集与所述投影点集进行配准，得到第二转换矩阵。
  2. 根据权利要求1所述的配准方法，其中，所述投影点集中各投影点分别为距离所述第一配准点集中各对应配准点最近的投影点。
  3. 根据权利要求1所述的配准方法，其中，所述将所述第一配准点集与所述投影点集进行配准，得到第二转换矩阵，包括：
    基于第一配准点集中各配准点的位姿坐标和所述投影点集中各投影点的位姿坐标确定所述第二转换矩阵。
  4. 根据权利要求3所述的配准方法，其中，所述基于第一配准点集中各配准点的位姿坐标和所述投影点集中各投影点的位姿坐标确定所述第二转换矩阵，包括：
    将所述第一配准点集中各配准点的位姿坐标和所述投影点集中各投影点的位姿坐标作为已知量，将所述第二转换矩阵作为未知量，代入第一目标函数中，以求解所述第二转换矩阵；
    其中，所述第一目标函数配置为最小化第一转换点集与所述第一配准点集之间的位姿差距，所述第一转换点集是基于所述第二转换矩阵将所述投影点集进行转换后得到的。
  5. 根据权利要求1所述的配准方法，其中，所述将所述第一配准点集与所述投影点集进行配准，得到第二转换矩阵，包括：
    获取所述第一配准点集中各配准点对应的配准权重；其中，针对所述第一配准点集中的每个配准点，所述配准权重是基于当前次配准的配准次数、预设配准次数以及所述配准点与所述配准点对应的所述投影点之间的位姿差距中的至少一个得到；
    将所述第一配准点集中各配准点对应的配准权重、所述第一配准点集中各配准点的位姿坐标和所述投影点集中各投影点的位姿坐标作为已知量，将所述第二转换矩阵作为未知量，代入第二目标函数中，以求解所述第二转换矩阵，所述第二目标函数配置为最小化第一转换点集与第一配准点集之间的位姿差距，第一转换点集是基于第二转换矩阵将投影点集进行转换后得到的。
  6. 根据权利要求1所述的配准方法，其中，所述第一配准点集中各配准点为所述待配准对象表面的生理解剖特征点，在所述将所述第一配准点集与所述投影点集进行配准，得到第二转换矩阵之后，所述配准方法还包括：
    判断所述当前次配准的配准次数是否达到第一迭代次数阈值；
    若所述当前次配准的配准次数未到达所述第一迭代次数阈值，则将所述第二影像模型作 为所述第一影像模型，将所述第二转换矩阵作为所述第一转换矩阵，返回执行所述基于所述第一转换矩阵对所述第一影像模型进行调整，得到第二影像模型的步骤。
  7. 根据权利要求6所述的配准方法，其中，所述配准方法还包括：
    若所述当前次配准的配准次数达到所述第一迭代次数阈值，则将所述第二影像模型作为所述第一影像模型，将所述第二转换矩阵作为所述第一转换矩阵，将所述待配准对象表面的第二配准点集作为所述第一配准点集，返回执行所述基于所述第一转换矩阵对所述第一影像模型进行调整，得到第二影像模型的步骤，直至所述当前次配准的配准次数达到第二迭代次数阈值；其中，所述第二配准点集中各配准点为所述待配准对象表面的生理解剖特征点之外的点。
  8. 根据权利要求1所述的配准方法，其中，所述第一配准点集中各配准点为所述待配准对象表面的生理解剖特征点之外的点，在所述将所述第一配准点集与所述投影点集进行配准，得到第二转换矩阵之后，所述配准方法还包括：
    判断所述当前次配准的配准次数是否达到第三迭代次数阈值；
    若所述当前次配准的配准次数未达到所述第三迭代次数阈值，则将所述第二影像模型作为所述第一影像模型，将所述第二转换矩阵作为所述第一转换矩阵，返回执行所述基于所述第一转换矩阵对所述第一影像模型进行调整，得到第二影像模型的步骤。
  9. 根据权利要求1所述的配准方法，其中，获取所述上一次配准得到的第一转换矩阵，包括：
    将所述第一影像模型表面的配准点集与所述第一配准点集进行配准，得到所述第一转换矩阵。
  10. 根据权利要求9所述的配准方法，其中，获取所述第一影像模型表面的配准点集的方法包括：
    获取模板影像模型，所述模板影像模型表面具有模板点集；
    将所述模板影像模型与所述第一影像模型进行匹配，以将所述模板点集映射至所述第一影像模型表面，得到所述第一影像模型表面的配准点集。
  11. 根据权利要求1所述的配准方法，还包括：判断所述当前次配准得到的第二转换矩阵是否达到预设转换矩阵阈值，若所述第二转换矩阵未达到预设转换矩阵阈值，则将所述第二影像模型作为第一影像模型，将所述第二转换矩阵作为第一转换矩阵，返回执行所述基于所述第一转换矩阵对所述第一影像模型进行调整，得到第二影像模型的步骤，直至所述当前次配准得到的第二转换矩阵达到预设转换矩阵阈值。
  12. 根据权利要求1所述的配准方法，还包括：基于所述第二转换矩阵中位移大小和旋转角度对所述第二影像模型进行移动和旋转，得到匹配影像模型。
  13. 一种配准装置，包括：
    获取模块，用于获取上一次配准所使用的待配准对象的第一影像模型以及所述上一次配准得到的第一转换矩阵；
    调整模块，用于对于当前次配准，基于所述第一转换矩阵对所述第一影像模型进行调整，得到第二影像模型；
    投影模块，用于将所述待配准对象表面的第一配准点集投影至所述第二影像模型表面，得到投影点集；
    配准模块，用于将所述第一配准点集与所述投影点集进行配准，得到第二转换矩阵。
  14. 一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时实现权利要求1至12中任一项所述的配准方法的步骤。
  15. 一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现权利要求1至12中任一项所述的配准方法的步骤。
  16. 一种计算机程序产品，包括计算机程序，计算机程序被处理器执行时实现权利要求1-12中任一项所述的配准方法的步骤。
PCT/CN2023/105062 2022-06-30 2023-06-30 配准方法、装置、计算机设备和可读存储介质 WO2024002360A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210758534.5 2022-06-30
CN202210758534.5A CN117372317A (zh) 2022-06-30 2022-06-30 配准方法、装置、计算机设备和可读存储介质

Publications (1)

Publication Number Publication Date
WO2024002360A1 true WO2024002360A1 (zh) 2024-01-04

Family

ID=89383331

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/105062 WO2024002360A1 (zh) 2022-06-30 2023-06-30 配准方法、装置、计算机设备和可读存储介质

Country Status (2)

Country Link
CN (1) CN117372317A (zh)
WO (1) WO2024002360A1 (zh)

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080095422A1 (en) * 2006-10-18 2008-04-24 Suri Jasjit S Alignment method for registering medical images
CN104778688A (zh) * 2015-03-27 2015-07-15 华为技术有限公司 点云数据的配准方法及装置
US20160217576A1 (en) * 2013-10-18 2016-07-28 Koninklijke Philips N.V. Registration of medical images
CN105844586A (zh) * 2007-12-18 2016-08-10 皇家飞利浦电子股份有限公司 基于特征的2d/3d图像配准
CN109559338A (zh) * 2018-11-20 2019-04-02 西安交通大学 一种基于加权主成分分析法及m估计的三维点云配准方法
CN109785374A (zh) * 2019-01-23 2019-05-21 北京航空航天大学 一种牙科增强现实手术导航的自动实时无标记图像配准方法
CN110363800A (zh) * 2019-06-19 2019-10-22 西安交通大学 一种基于点集数据与特征信息相融合的精确刚体配准方法
CN110415281A (zh) * 2019-07-30 2019-11-05 西安交通大学深圳研究院 一种基于Loam曲率加权的点集刚体配准方法
CN111414798A (zh) * 2019-02-03 2020-07-14 沈阳工业大学 基于rgb-d图像的头部姿态检测方法及系统
CN113160290A (zh) * 2021-03-31 2021-07-23 上海联影医疗科技股份有限公司 2d-3d医学图像配准方法、装置及可读存储介质
CN113205547A (zh) * 2021-03-18 2021-08-03 北京长木谷医疗科技有限公司 点云配准方法、骨头配准方法、装置、设备及存储介质
CN114219717A (zh) * 2021-11-26 2022-03-22 杭州三坛医疗科技有限公司 点云配准方法、装置、电子设备及存储介质
CN114529594A (zh) * 2022-02-17 2022-05-24 武汉联影智融医疗科技有限公司 配准方法、系统、计算机设备和存储介质

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080095422A1 (en) * 2006-10-18 2008-04-24 Suri Jasjit S Alignment method for registering medical images
CN105844586A (zh) * 2007-12-18 2016-08-10 皇家飞利浦电子股份有限公司 基于特征的2d/3d图像配准
US20160217576A1 (en) * 2013-10-18 2016-07-28 Koninklijke Philips N.V. Registration of medical images
CN104778688A (zh) * 2015-03-27 2015-07-15 华为技术有限公司 点云数据的配准方法及装置
CN109559338A (zh) * 2018-11-20 2019-04-02 西安交通大学 一种基于加权主成分分析法及m估计的三维点云配准方法
CN109785374A (zh) * 2019-01-23 2019-05-21 北京航空航天大学 一种牙科增强现实手术导航的自动实时无标记图像配准方法
CN111414798A (zh) * 2019-02-03 2020-07-14 沈阳工业大学 基于rgb-d图像的头部姿态检测方法及系统
CN110363800A (zh) * 2019-06-19 2019-10-22 西安交通大学 一种基于点集数据与特征信息相融合的精确刚体配准方法
CN110415281A (zh) * 2019-07-30 2019-11-05 西安交通大学深圳研究院 一种基于Loam曲率加权的点集刚体配准方法
CN113205547A (zh) * 2021-03-18 2021-08-03 北京长木谷医疗科技有限公司 点云配准方法、骨头配准方法、装置、设备及存储介质
CN113160290A (zh) * 2021-03-31 2021-07-23 上海联影医疗科技股份有限公司 2d-3d医学图像配准方法、装置及可读存储介质
CN114219717A (zh) * 2021-11-26 2022-03-22 杭州三坛医疗科技有限公司 点云配准方法、装置、电子设备及存储介质
CN114529594A (zh) * 2022-02-17 2022-05-24 武汉联影智融医疗科技有限公司 配准方法、系统、计算机设备和存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
WANG YONG, TANG JING; RAO QIN-FEI; YUAN CHAO-YAN: "Point Cloud Data Registration Based on Boundary Detection by the Center of Mass Distance Feature", JOURNAL OF CHINESE COMPUTER SYSTEMS, GAI-KAN BIANJIBU , SHENYANG, CN, vol. 36, no. 9, 30 September 2015 (2015-09-30), CN , pages 2096 - 2101, XP093124620, ISSN: 1000-1220 *

Also Published As

Publication number Publication date
CN117372317A (zh) 2024-01-09

Similar Documents

Publication Publication Date Title
CN107123137B (zh) 医学图像处理方法及设备
WO2021189843A1 (zh) Ct图像椎骨定位方法、装置、设备及介质
WO2022041598A1 (zh) 一种遥感影像分割方法、系统、终端以及存储介质
WO2023078309A1 (zh) 目标特征点提取方法、装置、计算机设备和存储介质
CN107492120B (zh) 点云配准方法
CN109584327B (zh) 人脸老化模拟方法、装置以及设备
CN111553985B (zh) 邻图配对式的欧式三维重建方法及装置
WO2023010565A1 (zh) 单目散斑结构光系统的标定方法、装置及终端
WO2024002360A1 (zh) 配准方法、装置、计算机设备和可读存储介质
CN114529594A (zh) 配准方法、系统、计算机设备和存储介质
CN116467896B (zh) 一种口腔正畸疗效模拟系统及方法
CN114469151A (zh) 数据校正方法、装置、计算机设备、存储介质和程序产品
CN116350958A (zh) 放疗计划参数的设定方法、装置、设备、介质和程序产品
WO2023174333A1 (zh) 磁共振梯度校正补偿因子的确定方法、磁共振梯度校正方法和装置
CN116385575A (zh) 图像重构方法、装置、计算机设备和存储介质
CN116630239A (zh) 影像分析方法、装置和计算机设备
CN116049330A (zh) 独立坐标系建立方法、装置、计算机设备、存储介质
CN111080734B (zh) 一种处理正电子发射断层扫描pet数据的方法及终端
CN108629798A (zh) 基于gpu的图像快速配准方法
CN113378929A (zh) 一种肺结节生长预测方法和计算机设备
WO2021102614A1 (zh) 一种处理正电子发射断层扫描pet数据的方法及终端
WO2023178527A1 (zh) 肿瘤放射治疗区域的生成方法及生成装置
WO2022141531A1 (zh) 图像配准评估方法、装置、电子设备及可读存储介质
CN115601451B (zh) 外参数据标定方法、装置、计算机设备和存储介质
TWI840833B (zh) 多尺度自編碼器生成方法、電腦設備及儲存介質

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: 23830522

Country of ref document: EP

Kind code of ref document: A1