CN112381922A - Method, system and terminal for obtaining skeleton model of human skeleton missing part - Google Patents
Method, system and terminal for obtaining skeleton model of human skeleton missing part Download PDFInfo
- Publication number
- CN112381922A CN112381922A CN202011166009.1A CN202011166009A CN112381922A CN 112381922 A CN112381922 A CN 112381922A CN 202011166009 A CN202011166009 A CN 202011166009A CN 112381922 A CN112381922 A CN 112381922A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud model
- model
- bone
- feature points
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 210000000988 bone and bone Anatomy 0.000 claims abstract description 169
- 230000001131 transforming effect Effects 0.000 claims abstract description 8
- 230000007547 defect Effects 0.000 claims description 9
- 206010065687 Bone loss Diseases 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000012217 deletion Methods 0.000 claims description 5
- 230000037430 deletion Effects 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 4
- 238000013519 translation Methods 0.000 claims description 3
- 230000002950 deficient Effects 0.000 abstract description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 16
- 238000006073 displacement reaction Methods 0.000 description 6
- 239000011159 matrix material Substances 0.000 description 5
- 210000000689 upper leg Anatomy 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 210000003414 extremity Anatomy 0.000 description 3
- 210000002082 fibula Anatomy 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000008439 repair process Effects 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000006467 substitution reaction Methods 0.000 description 3
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 2
- 101001022170 Homo sapiens FYN-binding protein 2 Proteins 0.000 description 2
- 101000990990 Homo sapiens Midkine Proteins 0.000 description 2
- 102100030335 Midkine Human genes 0.000 description 2
- 238000011960 computer-aided design Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 210000002414 leg Anatomy 0.000 description 2
- 239000012567 medical material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000399 orthopedic effect Effects 0.000 description 2
- 238000010146 3D printing Methods 0.000 description 1
- 206010061363 Skeletal injury Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005266 casting Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000007688 edging Methods 0.000 description 1
- 210000001930 leg bone Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000037314 wound repair Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
- G06F17/13—Differential equations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Operations Research (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention provides a method, a system and a terminal for obtaining a skeleton model of a human skeleton missing part, which can be used for: reconstructing CT images of the target bone and the reference bone to obtain a first bone model S1 and a second bone model S2; converting S1, S2 into a first point cloud model C1 and a second point cloud model C2; symmetrically transforming the second point cloud model C2 to obtain a third point cloud model C2'; carrying out deformation registration on the third point cloud model C2' to obtain a registration point cloud model C3; removing the points overlapped with the first point cloud model C1 in the registered point cloud model C3 to obtain a fourth point cloud model C4; judging whether the fourth point cloud model C4 has outliers, and correspondingly acquiring a target point cloud model; and converting the obtained target point cloud model into a three-dimensional skeleton model to obtain a skeleton model corresponding to the target skeleton. The invention is used for obtaining a skeleton model of a bone missing part to be repaired on a human skeleton, and is used for assisting in manufacturing a defective bone which is inosculated with the bone missing part to be repaired of a patient.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a method, a system and a terminal for obtaining a skeleton model of a human skeleton missing part, which are mainly suitable for a class of patients with symmetrical skeletons of the bone missing part to be repaired and used for providing data for establishing a defective bone of a limb skeleton of a patient made of medical materials.
Background
Computer-aided orthopedics has become a popular research, such as extracting information about a fractured bone of a patient from a CT image of the fractured bone to perform three-dimensional reconstruction to obtain a fractured bone model of the patient, and then registering the fractured bone model of the patient to obtain a complete bone model corresponding to the fractured bone.
However, in recent years, patients have been subjected to bone defects in their limbs in more and more cases, and it is often necessary to repair the bone defect sites in their limbs.
But computer-assisted orthopedic techniques have relatively few applications in bone defect repair. The currently common computer aided design bone wound repair method is that mimics software is adopted to register standard data to a three-dimensional bone model of a bone injury part for simulation repair in a computer, specifically, a personalized bone entity is manufactured by embedding casting, and subsequent processing including cleaning, edging, rebounding, passivating, modifying and the like is performed. The model adopted by the Mimics software is a standard human skeleton model (the software internally provides only one determined registration model/reference model), and is relatively single.
However, the bone size and local characteristics of each individual are different, and the standard human skeletal model of the prior art is adapted to patient limitations. Therefore, the invention provides a method, a system and a terminal for obtaining a bone model of a human bone missing part, which are used for solving the problems.
Disclosure of Invention
In view of the above-mentioned deficiencies of the prior art, the present invention provides a method, a system and a terminal for obtaining a bone model of a missing part of a human bone, which are used for obtaining a bone model of a missing part of a bone to be repaired on a human bone, so as to assist in manufacturing a defective bone matching with the missing part of the bone to be repaired of a patient.
In a first aspect, the present invention provides a method for obtaining a bone model of a human bone-missing part, comprising the steps of:
respectively reconstructing the CT image of the target bone and the CT image of the reference bone to obtain respectively corresponding 3D bone models which are sequentially recorded as a first bone model S1 and a second bone model S2; the target skeleton is a skeleton of a patient with a bone loss part to be repaired; the reference bone is a symmetrical bone of the target bone in the body of the patient; the first bone model S1 and the second bone model S2 are of equal size and are symmetrically placed;
converting the first bone model S1 into a first point cloud model C1, converting the second bone model S2 into a second point cloud model C2;
symmetrically transforming the second point cloud model C2 to obtain a third point cloud model C2' in the same state as the first point cloud model C1;
taking the first point cloud model C1 as a reference point cloud model, and performing deformation registration on the third point cloud model C2' relative to the reference point cloud model to obtain a registration point cloud model C3;
removing the points overlapped with the first point cloud model C1 in the registered point cloud model C3 to obtain a fourth point cloud model C4;
and (3) judging whether the fourth point cloud model C4 has outliers: if so, deleting outliers in the fourth point cloud model C4 to obtain a target point cloud model; if not, the fourth point cloud model C4 is the obtained target point cloud model;
and converting the obtained target point cloud model into a three-dimensional skeleton model to obtain a skeleton model of the to-be-repaired bone deletion part of the target skeleton.
Further, the first point cloud model C1 is used as a reference point cloud model, and the third point cloud model C2' is subjected to deformation registration relative to the reference point cloud model to obtain a registered point cloud model C3, and the specific implementation steps include:
selecting a corresponding number of feature points on the first point cloud model C1, and respectively recording the feature points as first feature points; selecting feature points of corresponding positions on the third point cloud model C2', and respectively recording the feature points as second feature points; each first characteristic point is not a point on the bone missing part to be repaired of the target bone;
based on the constraint of the feature points, the third point cloud model C2 ' is translated and rotated to align the second feature points selected on the third point cloud model C2 ' with the first feature points on the first point cloud model C1, and then the first point cloud model C1 is used as a reference point cloud model to perform deformation registration on the third point cloud model C2 ' to obtain a registered point cloud model C3.
Further, the number of the first feature points is at least 6.
In a second aspect, the present invention provides a bone model obtaining system for a human bone defect site, comprising:
a 3D bone model reconstruction unit, which reconstructs the CT image of the target bone and the CT image of the reference bone to obtain corresponding 3D bone models, and the 3D bone models are recorded as a first bone model S1 and a second bone model S2 in sequence; the target skeleton is a skeleton of a patient with a bone loss part to be repaired; the reference bone is a symmetrical bone of the target bone in the body of the patient; the first bone model S1 and the second bone model S2 are of equal size and are symmetrically placed;
a point cloud model conversion unit which converts the first skeleton model S1 into a first point cloud model C1, and converts the second skeleton model S2 into a second point cloud model C2;
the point cloud model transformation unit is used for symmetrically transforming the second point cloud model C2 to obtain a third point cloud model C2' in the same state as the first point cloud model C1;
the registration deformation unit is used for taking the first point cloud model C1 as a reference point cloud model, and performing deformation registration on the third point cloud model C2' relative to the reference point cloud model to obtain a registration point cloud model C3;
an overlapping point removing unit, which is used for removing the point overlapping with the first point cloud model C1 in the registration point cloud model C3 to obtain a fourth point cloud model C4;
an outlier deleting unit that determines whether or not an outlier exists in the fourth point cloud model C4: if so, deleting outliers in the fourth point cloud model C4 to obtain a target point cloud model; if not, obtaining a fourth point cloud model C4 as the target point cloud model;
and the target point cloud conversion unit is used for converting the obtained target point cloud model into a three-dimensional skeleton model, and the converted three-dimensional skeleton model is the skeleton model of the missing part of the bone to be repaired of the obtained target skeleton.
Further, the registration deformation unit includes:
a feature point selecting unit, configured to select a corresponding number of feature points on the first point cloud model C1, and record the feature points as first feature points respectively; the feature points corresponding to the positions are selected from the third point cloud model C2' and are respectively marked as second feature points; each first characteristic point is not a point on the bone missing part to be repaired of the target bone;
and the registration model obtaining unit is used for carrying out translational and rotational deformation on the third point cloud model C2 ' based on the constraint of the feature points to align the second feature points selected on the third point cloud model C2 ' with the first feature points on the first point cloud model C1, and then carrying out deformation registration on the third point cloud model C2 ' by taking the first point cloud model C1 as a reference point cloud model to obtain a registration point cloud model C3.
Further, the number of the first feature points is at least 6.
In a third aspect, the present invention provides a terminal, including:
a processor;
a memory for storing instructions for execution by the processor;
wherein the processor is configured to perform the method as described in the above aspects.
The invention has the beneficial effects that:
(1) the method, the system and the terminal for obtaining the skeleton model of the human skeleton missing part can finally obtain the skeleton model of the bone missing part to be repaired of the patient based on the CT image of the target skeleton of the patient and the CT image of the reference skeleton of the patient through computer aided design.
(2) The invention provides a method, a system and a terminal for obtaining a skeleton model of a human skeleton missing part, wherein a reference skeleton is a symmetrical skeleton of a target skeleton in a patient body, when the method is used, a skeleton model of the target skeleton and the reference skeleton is converted into a point cloud model, then the point cloud model of the reference skeleton is symmetrically converted to obtain a third point cloud model in the same state as the point cloud model of the target skeleton, then the point cloud model of the reference skeleton is used as the reference point cloud model, the third point cloud model C2' is aligned and registered to obtain a registered point cloud model C3, then a point overlapped with the first point cloud model C1 in the registered point cloud model C3 is removed to obtain a fourth point cloud model C4, then whether an outlier exists in the fourth point cloud model C4 or not is judged to obtain a corresponding target point cloud model, and then the obtained target point cloud model is converted into a three-dimensional skeleton model, the bone model of the bone missing part of the target bone of the patient is obtained, the realization is easy, and when the bone models of the bone missing parts of different patients are manufactured, the reference bones related in the invention are respectively symmetrical bones of the patient, so that the use of the standard human body bone model in the background technology is avoided, and the application degree of the model to the patient is increased to a certain extent.
In addition, the invention has reliable design principle, simple structure and very wide application prospect.
Drawings
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a schematic flow diagram of a method of one embodiment of the invention.
FIG. 2 is a schematic view of an embodiment of the first bone model S1 and the second bone model S2 of the present invention.
FIG. 3 is a schematic diagram of an embodiment of the first point cloud model C1 and the second point cloud model C2 according to the invention.
FIG. 4 is a schematic diagram of an embodiment of the third point cloud model C2' and the second point cloud model C2 in the present invention.
Fig. 5 is a schematic diagram of an embodiment of the first point cloud model C1 and the third point cloud model C2' in the present invention.
Fig. 6 is a schematic diagram of an embodiment of performing deformed registration on the third point cloud model C2' in the present invention.
FIG. 7 is a schematic view of the first point cloud model C1 and its registered point cloud model C3 superimposed thereon.
Fig. 8 is a schematic point cloud structure diagram of a target point cloud model obtained in the present invention.
FIG. 9 is a schematic block diagram of the three-dimensional bone model into which the target point cloud model of FIG. 8 is converted.
FIG. 10 is a schematic block diagram of a system of one embodiment of the present invention.
Fig. 11 is a schematic structural diagram of a terminal according to an embodiment of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present invention, the technical solution in the embodiment of the present invention will be clearly and completely described below with reference to the drawings in the embodiment of the present invention, and it is obvious that the described embodiment is only a part of the embodiment of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1:
FIG. 1 is a schematic flow diagram of a method of one embodiment of the invention.
As shown in fig. 1, the method 100 includes:
step 110, reconstructing a CT image of a target bone and a CT image of a reference bone respectively to obtain corresponding 3D bone models, and recording the 3D bone models as a first bone model S1 and a second bone model S2 in sequence; the target skeleton is a skeleton of a patient with a bone loss part to be repaired; the reference bone is a symmetrical bone of the target bone in the body of the patient; the first bone model S1 and the second bone model S2 are of equal size and are symmetrically placed;
step 120, converting the first bone model S1 into a first point cloud model C1, and converting the second bone model S2 into a second point cloud model C2;
step 130, symmetrically transforming the second point cloud model C2 to obtain a third point cloud model C2' in the same state as the first point cloud model C1;
step 140, taking the first point cloud model C1 as a reference point cloud model, and performing deformation registration on the third point cloud model C2' relative to the reference point cloud model to obtain a registration point cloud model C3;
150, removing the points overlapped with the first point cloud model C1 in the registered point cloud model C3 to obtain a fourth point cloud model C4;
step 160, determining whether the fourth point cloud model C4 has outliers: if so, deleting outliers in the fourth point cloud model C4 to obtain a target point cloud model; if not, the fourth point cloud model C4 is the obtained target point cloud model;
step 170, converting the obtained target point cloud model into a three-dimensional skeleton model, namely obtaining a skeleton model of the to-be-repaired bone deletion part of the target skeleton.
Optionally, in step 140, the first point cloud model C1 is used as a reference point cloud model, and the third point cloud model C2' is subjected to deformation registration with respect to the reference point cloud model to obtain a registered point cloud model C3, and the specific implementation steps include:
selecting a corresponding number of feature points on the first point cloud model C1, and respectively recording the feature points as first feature points; selecting feature points of corresponding positions on the third point cloud model C2', and respectively recording the feature points as second feature points; each first characteristic point is not a point on the bone missing part to be repaired of the target bone;
based on the constraint of the feature points, the third point cloud model C2 ' is subjected to translational and rotational deformation to enable the second feature points selected on the third point cloud model C2 ' to be aligned with the first feature points on the first point cloud model C1, and then the first point cloud model C1 is used as a reference point cloud model to perform deformation registration on the third point cloud model C2 ' to obtain a registration point cloud model C3.
Optionally, the number of first feature points is at least 6.
In order to facilitate understanding of the present invention, the method for obtaining a bone model of a human bone-missing part according to the present invention will be further described below with reference to the principle of the method for obtaining a bone model of a human bone-missing part according to the present invention, taking the injured femur of the left leg of patient a (with the bone-missing part 300 to be repaired) as an example. The method comprises the following concrete steps:
specifically, the method for obtaining the bone model of the human bone deletion site comprises the following steps:
step P1: a CT image of the injured left leg femur (i.e., the target bone) of patient a and a CT image of the healthy right leg femur (i.e., the reference bone) of patient a are reconstructed to obtain respective 3D bone models, which are sequentially denoted as a first bone model S1 and a second bone model S2, as shown in fig. 2.
In fig. 2: s1 is the bone model located on the left side of the figure, and S2 is the bone model located on the right side of the figure. The bone model S1 on the left side of fig. 2 is the same size and symmetrically placed as the bone model S2 on the right side of fig. 2.
Before step P1 is performed, a CT image of the injured left leg femur of patient a and a CT image of the healthy right leg femur (i.e., the reference bone) of patient a are first acquired.
Step P2 is then performed.
Step P2: the first bone model S1 is converted into a first point cloud model C1 and the second bone model S2 is converted into a second point cloud model C2. As shown in fig. 3.
In fig. 3: the point cloud model on the left side of the diagram is C1, and the point cloud model on the right side of the diagram is C2.
Step P3 is then performed.
Step P3: the second point cloud model C2 is subjected to symmetric transformation to obtain a third point cloud model C2' in the same state as the first point cloud model C1.
Specifically, in the present embodiment: and (3) symmetrically transforming the point cloud of the second point cloud model C2 of the leg bone of the right leg in a three-dimensional coordinate system (X, Y, Z) of the C2 relative to an XOZ (O is the origin of the three-dimensional coordinate system (X, Y, Z)) plane to obtain the third point cloud model C2'. Based on this, the point cloud of the right leg calf bone becomes a point cloud having the same distribution state as the point cloud of the left leg calf bone. As shown in fig. 4: the point cloud model on the left side of the diagram is the third point cloud model C2', and the point cloud model on the right side of the diagram is the second point cloud model C2.
Step P4: and taking the first point cloud model C1 as a reference point cloud model, and performing deformation registration on the third point cloud model C2' relative to the reference point cloud model to obtain a registration point cloud model C3.
Specifically, the step P4 may include:
step P41: selecting a corresponding number of feature points on the first point cloud model C1, and respectively recording the feature points as first feature points; and selecting the feature points corresponding to the first feature points on the third point cloud model C2', and marking as second feature points.
When selecting a selected feature point, for example, selecting a first feature point on the first point cloud model C1, and marking as a first feature point 1, selecting a second feature point on the third point cloud model C2' at a position corresponding to the first feature point 1, and marking as a second feature point 1; the other feature points on the first point cloud model C1 and the third point cloud model C2' may be selected by referring to the selection manner of the first feature point 1 and the second feature point 1.
In the present embodiment, 9 first feature points are clicked on the first point cloud model C1, and 9 second feature points are clicked on corresponding positions of the third point cloud model C2'. The 9 first landmark points clicked on the first point cloud model C1 are opposite to the 9 second landmark points clicked on the third point cloud model C2'. All the first feature points selected are not points on the bone loss part 300 to be repaired of the target bone. As shown in fig. 5, in fig. 5: the point cloud model on the left side is the first point cloud model C1, and the point cloud model on the right side is the third point cloud model C2'. In fig. 5, 9 points on the first point cloud model C1, which are identified by black dots, are all first feature points, such as the first feature point 400; the 9 points identified by black dots on the third point cloud model C2' are all the second feature points, such as the second feature point 500.
Step P42: based on the constraint of the feature points, the third point cloud model C2 ' is subjected to translational and rotational deformation to enable the second feature points selected on the third point cloud model C2 ' to be aligned with the first feature points on the first point cloud model C1, and then the first point cloud model C1 is used as a reference point cloud model to perform deformation registration on the third point cloud model C2 ' to obtain a registration point cloud model C3. Specifically, the method comprises the following steps:
(I) pretreatment
First, a preprocessing is performed on the third point cloud model C2' (source data): converting the triangular mesh on the point cloud model C2' into tetrahedral meshes, wherein each tetrahedron is formed by four triangular faces;
next, feature point constraints (corresponding to the connecting lines between the first feature points and the second feature points) are established for the 9 first feature points clicked on the first point cloud model C1 and the 9 second feature points clicked on the third point cloud model C2', as shown in fig. 6.
In fig. 6: the point cloud model on the left side in the figure is the first point cloud model C1, and the point cloud model on the right side is the third point cloud model C2'.
(II) deformed registration
The registered point cloud model C3 obtained in this step P42 is used to overlap the first point cloud model C1. A schematic diagram of the overlapping point cloud structure of the registered point cloud model C3 and the first point cloud model C1 is shown in fig. 7. Specifically, in step P42, the method steps of deformed registration are initial registration and volume registration:
1) the global displacement (initial registration) is performed using the ARAP deformation method.
The specific method comprises the following steps:
defining a formula:
where N is the number of feature points (in this embodiment, N is 9), and piDenotes the coordinates, q, of the ith first feature point on C1iThe coordinates of the ith second feature point on C2' are shown,
is the rotation and translation matrix of the ith second feature point on C2 ', and E represents the deformation energy E for ARAP deformation of the third point cloud model C2'.
Using alternating iterationsSolving the deformation energy E by a miniaturization strategy to obtain a translation rotation matrix when the energy is minimum
The whole of C2' is based on
Translational rotation is performed such that C1 is aligned with C2' (being the initial alignment).
2) Volume registration
The purpose of volume registration is to iteratively deform the model C2' to C1 and calculate a more accurate registration result based on the deformed model. Since the initial registration has been globally shifted, in order to fully use the voxel information and the mesh information of the model C2 ', and to make C2' have large deformation and deviate from the original model structure during deformation as little as possible, the local neighborhood information is used in the step of volume registration. To this end, the volume registration algorithm defines the following target equation:
Edef(S0)＝aEdis(S0)+bEdeform(S0)+Ecorr(S0)+Eaffine(S0), (2.1)
wherein E isdis(S0),Edeform(S0),Ecorr(S0) And Eaffine(S0) Is an energy term for model registration. a and b are coefficients. In this objective equation Edis(S0) And Ecorr(S0) The euclidean structural information of model C2' is captured. Edef(S0) Riemann information Structure, E, capturing model C2affine(S0) The topology information of model C2' is captured. S0Is solved, and the final result is C2' after the registration deformation.
EdisThe term is an energy term related to the displacement position, and the term is mainly used for limiting larger displacement generated in the current iteration compared with the last displacement result, so that the position of the tetrahedral mesh is not deviated from the position of the last iteration as much as possible. In accordance withConstructing a first term according to a Gihonov regularization term, wherein a specific formula is as follows:
Edis(S0)＝‖S0-Sprev||2, (2.2)
Edeformthe term is an energy term for the deformed position, and is mainly used to limit the large displacement of the new tetrahedral mesh position generated in each iteration compared with the initial tetrahedral mesh position, so that the position of the tetrahedral mesh is constrained to be as far as possible not to deviate from the initially aligned position, so as not to reduce the registration accuracy. The specific formula is as follows:
Edeform(S0)＝(S0-S′)TL(S0-S′), (2.3)
wherein S 'represents the point cloud model after C2' has been initially aligned, and L represents the laplacian of the tetrahedral model, and L is calculated by the following formula:
where S refers to the model C2',iis the deformation gradient, v, with respect to the ith tetrahedral mesh in model C2iRepresenting the volume of the ith tetrahedral mesh in the model.
EcorrThe term is an energy term related to the registration relationship, and is mainly used for improving the registration accuracy of the tetrahedral mesh and the voxel data model and enabling the tetrahedral mesh to deform towards the voxel data as much as possible. The specific formula is as follows:
Ecorr(S0)＝||S0-P×T||2， (2.5)
wherein
Record S0And C1, which is an n × m matrix. T representsIs a point set of C1, when a point i in C2 is the closest point to a point j in C1 or the characteristic constraint point is 1, for example, siAnd cjIs the nearest point or a feature constraint point, then
E is an m x m identity matrix,
is the kronecker product.
EaffineThe term is an energy term about the neighborhood structure, which is mainly based on the spatial structure of the points to guarantee the accuracy of the registration relationship. Since the spatial position relationship of the registration points in the two models with respect to the other points in the local region should differ very little, the geometry of the points and the surrounding neighborhood points is relatively stable. The vertices of the tetrahedral mesh are thus represented by neighborhood points, the formula being:
si＝∑j∈N(i)wijsj, (2.6)
where w is a weighted affine matrix and N (i) represents siThe neighborhood vertex index set, affine w can be derived by the above equation. W has two main properties: 1) for a value other than siNear point of (a), wij0; 2) the row of w is 1. Finally w-based energy term E on neighborhood structureaffineThe definition is as follows:
wherein
Refer to
Row j.
In order to solve the point where the objective function yields the deformation, the formula of 2.1 is differentiated to yield the following formula:
setting the derivative after differentiation to be 0 to obtain a final linear equation, wherein the formula is as follows:
[bL+(a+2)E]S0＝PT+PLS′+aSprev+Q, (2.9)
solve to S0And obtaining a result model of the volume registration. Then, the body registration is carried out in an iteration mode, and a final result S is obtained0S finally solved0The point cloud model is C3. In this embodiment, only ten iterations are required.
For the selection problem of coefficients, since the initial correspondence relationship cannot guarantee the accuracy at the beginning of the volume registration iteration, the algorithm should increase the constraints of the energy terms of displacement and deformation as much as possible, and in this embodiment, the algorithm sets the initial values of a and b to 0.8 and 1, respectively. As the iteration increases, when S0Gradually approaches SprevWe need to reduce the values of a and b to reduce the regularization strength. In the present embodiment, the values of a and b are halved 5 times per iteration.
Step P5 is then performed.
It should be noted that, in specific implementation, step P4 can be implemented by a person skilled in the art by using any other relevant deformation-based registration method in the prior art.
Step P5: and removing the overlapped points of the registration point cloud model and the first point cloud model C1 to obtain a fourth point cloud model C4.
In the present embodiment, a point in the registered point cloud model which is less than 0.04cm away from the point in the first point cloud model C1 is regarded as an overlapping point.
Step P6 is then performed.
Step P6: and (3) judging whether the fourth point cloud model C4 has outliers: if so, deleting outliers in the fourth point cloud model C4 to obtain a target point cloud model; if not, the fourth point cloud model C4 is the obtained target point cloud model.
Specifically, the variance can be calculated by taking 50 adjacent points through a statistical outlierremove method in a PCL point cloud library, and points with a distance greater than one time of a standard variance are regarded as outliers and deleted.
Step P7 is then performed.
Step P7: and converting the currently obtained target point cloud model (as shown in fig. 8) into a three-dimensional skeleton model (as shown in fig. 9), wherein the converted three-dimensional skeleton model is the obtained skeleton model of the missing part of the left leg calf bone.
When the method is used specifically, the model is printed out from the bone model of the missing part obtained by the method 100 through a 3D printing technology, so that data is provided for a medical institution to establish the missing part skeleton made of special medical materials, a doctor can make up the missing bone on the skeleton of a patient, and the method is helpful for assisting in repairing the missing part skeleton.
Example 2:
fig. 10 is an embodiment of the bone model acquiring system for a human bone defect site according to the present invention.
Referring to fig. 10, the system 200, comprises:
a 3D bone model reconstruction unit 201, which reconstructs a CT image of a target bone and a CT image of a reference bone to obtain 3D bone models respectively corresponding to the CT images, and records the 3D bone models as a first bone model S1 and a second bone model S2 in sequence; the target skeleton is the skeleton of the patient with the bone loss part 300 to be repaired; the reference bone is a symmetrical bone of the target bone in the body of the patient; the first bone model S1 and the second bone model S2 are of equal size and are symmetrically placed;
a point cloud model conversion unit 202 converting the first bone model S1 into a first point cloud model C1, and converting the second bone model S2 into a second point cloud model C2;
the point cloud model transformation unit 203 is used for symmetrically transforming the second point cloud model C2 to obtain a third point cloud model C2' in the same state as the first point cloud model C1;
a registration deformation unit 204, which takes the first point cloud model C1 as a reference point cloud model, and performs deformation registration on the third point cloud model C2' relative to the reference point cloud model to obtain a registration point cloud model C3;
an overlapping point removing unit 205, which removes the point overlapping with the first point cloud model C1 in the registered point cloud model C3, to obtain a fourth point cloud model C4;
outlier removing section 206 determines whether or not an outlier exists in fourth cloud model C4: if so, deleting outliers in the fourth point cloud model C4 to obtain a target point cloud model; if not, obtaining a fourth point cloud model C4 as the target point cloud model;
the target point cloud conversion unit 207 converts the obtained target point cloud model into a three-dimensional skeleton model, which is the skeleton model of the missing bone part 300 to be repaired of the obtained target skeleton.
Optionally, as an embodiment of the present invention, the registration deformation unit 204 includes:
a feature point selecting unit 2041, configured to select a corresponding number of feature points from the first point cloud model C1, and record the feature points as first feature points respectively; the feature points corresponding to the positions are selected from the third point cloud model C2' and are respectively marked as second feature points; each first characteristic point is not a point on the missing part 300 of the target bone to be repaired;
the registration model obtaining unit 2042 performs translational and rotational deformation on the third point cloud model C2 'based on the constraint of the feature points to align the second feature points selected thereon with the first feature points on the first point cloud model C1, and then performs deformation registration on the third point cloud model C2' with the first point cloud model C1 as a reference point cloud model to obtain a registration point cloud model C3.
Optionally, the number of first feature points is at least 6.
Example 3:
fig. 11 is a schematic structural diagram of a terminal 600 according to an embodiment of the present invention, where the terminal 600 may be used to execute the method 100 according to embodiment 1 of the present invention.
The terminal 600 may include: a processor 610, a memory 620, and a communication unit 630. The components communicate via one or more buses, and those skilled in the art will appreciate that the architecture of the servers shown in the figures is not intended to be limiting, and may be a bus architecture, a star architecture, a combination of more or less components than those shown, or a different arrangement of components.
The memory 620 may be used for storing instructions executed by the processor 610, and the memory 620 may be implemented by any type of volatile or non-volatile storage terminal or combination thereof, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The executable instructions in memory 620, when executed by processor 610, enable terminal 600 to perform some or all of the steps in the method embodiments described below.
The processor 610 is a control center of the storage terminal, connects various parts of the entire electronic terminal using various interfaces and lines, and performs various functions of the electronic terminal and/or processes data by operating or executing software programs and/or modules stored in the memory 620 and calling data stored in the memory. The processor may be composed of an Integrated Circuit (IC), for example, a single packaged IC, or a plurality of packaged ICs connected with the same or different functions. For example, the processor 610 may include only a Central Processing Unit (CPU). In the embodiment of the present invention, the CPU may be a single operation core, or may include multiple operation cores.
A communication unit 630, configured to establish a communication channel so that the storage terminal can communicate with other terminals. And receiving user data sent by other terminals or sending the user data to other terminals.
The same and similar parts in the various embodiments in this specification may be referred to each other. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and the relevant points can be referred to the description in the method embodiment.
Although the present invention has been described in detail by referring to the drawings in connection with the preferred embodiments, the present invention is not limited thereto. Various equivalent modifications or substitutions can be made on the embodiments of the present invention by those skilled in the art without departing from the spirit and scope of the present invention, and these modifications or substitutions are within the scope of the present invention/any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (7)
1. A method for obtaining a bone model of a human bone deletion part is characterized by comprising the following steps:
respectively reconstructing the CT image of the target bone and the CT image of the reference bone to obtain respectively corresponding 3D bone models which are sequentially recorded as a first bone model S1 and a second bone model S2; the target skeleton is a skeleton of a patient with a bone loss part to be repaired; the reference bone is a symmetrical bone of the target bone in the body of the patient; the first bone model S1 and the second bone model S2 are of equal size and are symmetrically placed;
converting the first bone model S1 into a first point cloud model C1, converting the second bone model S2 into a second point cloud model C2;
symmetrically transforming the second point cloud model C2 to obtain a third point cloud model C2' in the same state as the first point cloud model C1;
taking the first point cloud model C1 as a reference point cloud model, and performing deformation registration on the third point cloud model C2' relative to the reference point cloud model to obtain a registration point cloud model C3;
removing the points overlapped with the first point cloud model C1 in the registered point cloud model C3 to obtain a fourth point cloud model C4;
and (3) judging whether the fourth point cloud model C4 has outliers: if so, deleting outliers in the fourth point cloud model C4 to obtain a target point cloud model; if not, the fourth point cloud model C4 is the obtained target point cloud model;
and converting the obtained target point cloud model into a three-dimensional skeleton model to obtain a skeleton model of the to-be-repaired bone deletion part of the target skeleton.
2. The method for obtaining the bone model of the human bone defect site according to claim 1, wherein the first point cloud model C1 is used as a reference point cloud model, and the third point cloud model C2' is transformed and registered with respect to the reference point cloud model to obtain a registered point cloud model C3, and the specific implementation steps include:
selecting a corresponding number of feature points on the first point cloud model C1, and respectively recording the feature points as first feature points; selecting feature points of corresponding positions on the third point cloud model C2', and respectively recording the feature points as second feature points; each first characteristic point is not a point on the bone missing part to be repaired of the target bone;
based on the constraint of the feature points, the third point cloud model C2 ' is subjected to translational and rotational deformation to enable the second feature points selected on the third point cloud model C2 ' to be aligned with the first feature points on the first point cloud model C1, and then the first point cloud model C1 is used as a reference point cloud model to perform deformation registration on the third point cloud model C2 ' to obtain a registration point cloud model C3.
3. The method of claim 2, wherein the number of the first feature points is at least 6.
4. A bone model acquisition system for a human bone defect site, comprising:
a 3D bone model reconstruction unit, which reconstructs the CT image of the target bone and the CT image of the reference bone to obtain corresponding 3D bone models, and the 3D bone models are recorded as a first bone model S1 and a second bone model S2 in sequence; the target skeleton is a skeleton of a patient with a bone loss part to be repaired; the reference bone is a symmetrical bone of the target bone in the body of the patient; the first bone model S1 and the second bone model S2 are of equal size and are symmetrically placed;
a point cloud model conversion unit which converts the first skeleton model S1 into a first point cloud model C1, and converts the second skeleton model S2 into a second point cloud model C2;
the point cloud model transformation unit is used for symmetrically transforming the second point cloud model C2 to obtain a third point cloud model C2' in the same state as the first point cloud model C1;
the registration deformation unit is used for taking the first point cloud model C1 as a reference point cloud model, and performing deformation registration on the third point cloud model C2' relative to the reference point cloud model to obtain a registration point cloud model C3;
an overlapping point removing unit, which is used for removing the point overlapping with the first point cloud model C1 in the registration point cloud model C3 to obtain a fourth point cloud model C4;
an outlier deleting unit that determines whether or not an outlier exists in the fourth point cloud model C4: if so, deleting outliers in the fourth point cloud model C4 to obtain a target point cloud model; if not, obtaining a fourth point cloud model C4 as the target point cloud model;
and the target point cloud conversion unit is used for converting the obtained target point cloud model into a three-dimensional skeleton model, and the converted three-dimensional skeleton model is the skeleton model of the missing part of the bone to be repaired of the obtained target skeleton.
5. The system for obtaining a bone model of a human bone defect site according to claim 4, wherein the registration deformation unit comprises:
a feature point selecting unit, configured to select a corresponding number of feature points on the first point cloud model C1, and record the feature points as first feature points respectively; the feature points corresponding to the positions are selected from the third point cloud model C2' and are respectively marked as second feature points; each first characteristic point is not a point on the bone missing part to be repaired of the target bone;
and the registration model obtaining unit is used for carrying out translation rotation on the third point cloud model C2 ' based on the constraint of the feature points to align the second feature points selected on the third point cloud model C2 ' with the first feature points on the first point cloud model C1, and then carrying out deformation registration on the third point cloud model C2 ' by taking the first point cloud model C1 as a reference point cloud model to obtain a registration point cloud model C3.
6. The system for obtaining a bone model of a human bone defect site according to claim 5, wherein the number of the first feature points is at least 6.
7. A terminal, comprising:
a processor;
a memory for storing instructions for execution by the processor;
wherein the processor is configured to perform the method of any one of claims 1-3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011166009.1A CN112381922A (en) | 2020-10-27 | 2020-10-27 | Method, system and terminal for obtaining skeleton model of human skeleton missing part |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011166009.1A CN112381922A (en) | 2020-10-27 | 2020-10-27 | Method, system and terminal for obtaining skeleton model of human skeleton missing part |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112381922A true CN112381922A (en) | 2021-02-19 |
Family
ID=74577779
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011166009.1A Pending CN112381922A (en) | 2020-10-27 | 2020-10-27 | Method, system and terminal for obtaining skeleton model of human skeleton missing part |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112381922A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114332378A (en) * | 2021-12-31 | 2022-04-12 | 西安交通大学 | Human skeleton three-dimensional model obtaining method and system based on two-dimensional medical image |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070198022A1 (en) * | 2001-05-25 | 2007-08-23 | Conformis, Inc. | Patient Selectable Joint Arthroplasty Devices and Surgical Tools |
| CN105046698A (en) * | 2015-07-06 | 2015-11-11 | 嘉恒医疗科技（上海）有限公司 | Shoulder joint defect parameter statistics method and system based on left and right symmetry information |
| CN105303604A (en) * | 2015-10-19 | 2016-02-03 | 中国科学院软件研究所 | Measuring method and system for single-side osteal damage of human body |
| CN110288638A (en) * | 2019-06-18 | 2019-09-27 | 济南大学 | A kind of knochenbruch model rough registration method, system and knochenbruch Model registration method |
| CN111297478A (en) * | 2020-03-10 | 2020-06-19 | 南京市第一医院 | Preoperative planning method for knee joint revision surgery |
-
2020
- 2020-10-27 CN CN202011166009.1A patent/CN112381922A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070198022A1 (en) * | 2001-05-25 | 2007-08-23 | Conformis, Inc. | Patient Selectable Joint Arthroplasty Devices and Surgical Tools |
| CN105046698A (en) * | 2015-07-06 | 2015-11-11 | 嘉恒医疗科技（上海）有限公司 | Shoulder joint defect parameter statistics method and system based on left and right symmetry information |
| CN105303604A (en) * | 2015-10-19 | 2016-02-03 | 中国科学院软件研究所 | Measuring method and system for single-side osteal damage of human body |
| CN110288638A (en) * | 2019-06-18 | 2019-09-27 | 济南大学 | A kind of knochenbruch model rough registration method, system and knochenbruch Model registration method |
| CN111297478A (en) * | 2020-03-10 | 2020-06-19 | 南京市第一医院 | Preoperative planning method for knee joint revision surgery |
Non-Patent Citations (2)
| Title |
|---|
| 张铭宣: "基于结构信息约束非刚体形变的三维物体配准方法研究", 《济南大学硕士学位论文》 * |
| 陈中中 等: "《快速成型技术与生物医学导论》", 31 October 2013 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114332378A (en) * | 2021-12-31 | 2022-04-12 | 西安交通大学 | Human skeleton three-dimensional model obtaining method and system based on two-dimensional medical image |
| CN114332378B (en) * | 2021-12-31 | 2024-01-16 | 西安交通大学 | Human skeleton three-dimensional model acquisition method and system based on two-dimensional medical image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7493464B2 (en) | Automated canonical pose determination for 3D objects and 3D object registration using deep learning | |
| WO2022037696A1 (en) | Bone segmentation method and system based on deep learning | |
| Fu et al. | Automatic and hierarchical segmentation of the human skeleton in CT images | |
| US20210012492A1 (en) | Systems and methods for obtaining 3-d images from x-ray information for deformed elongate bones | |
| Zheng | Statistical shape model‐based reconstruction of a scaled, patient‐specific surface model of the pelvis from a single standard AP x‐ray radiograph | |
| US20080205719A1 (en) | Method of Model-Based Elastic Image Registration For Comparing a First and a Second Image | |
| Stoll et al. | Template Deformation for Point Cloud Fitting. | |
| Bijar et al. | Atlas-based automatic generation of subject-specific finite element tongue meshes | |
| CN114792326A (en) | Surgical navigation point cloud segmentation and registration method based on structured light | |
| WO2019180746A1 (en) | A method for obtaining 3-d deformity correction for bones | |
| CN116797519A (en) | Brain glioma segmentation and three-dimensional visualization model training method and system | |
| Sarmah et al. | Survey of methods and principles in three-dimensional reconstruction from two-dimensional medical images | |
| Weiherer et al. | Learning the shape of female breasts: an open-access 3D statistical shape model of the female breast built from 110 breast scans | |
| CN112381922A (en) | Method, system and terminal for obtaining skeleton model of human skeleton missing part | |
| CN114004940B (en) | Non-rigid generation method, device and equipment of face defect reference data | |
| CN110473241A (en) | Method for registering images, storage medium and computer equipment | |
| WO2022229816A1 (en) | 3d reconstruction of anatomical images | |
| CN115953529A (en) | System and method for obtaining three-dimensional mannequin | |
| CN114881930A (en) | 3D target detection method, device, equipment and storage medium based on dimension reduction positioning | |
| CN115049764A (en) | Training method, device, equipment and medium for SMPL parameter prediction model | |
| CN102138159A (en) | Simultaneous model-based segmentation of objects satisfying pre-defined spatial relationships | |
| WO2020137677A1 (en) | Image processing device, image processing method, and program | |
| Borotikar et al. | Augmented statistical shape modeling for orthopedic surgery and rehabilitation | |
| US20240127440A1 (en) | Manufacturing method of learning model, learning model, estimation method, image processing system, and program | |
| Starzynski et al. | Morphing algorithm for building individualized 3D skeleton model from CT data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210219 |
|
| RJ01 | Rejection of invention patent application after publication |