CN109035311A - A kind of curved bone fracture autoregistration and internal fixation steel plate pre-bending modeling method - Google Patents
A kind of curved bone fracture autoregistration and internal fixation steel plate pre-bending modeling method Download PDFInfo
- Publication number
- CN109035311A CN109035311A CN201810754511.0A CN201810754511A CN109035311A CN 109035311 A CN109035311 A CN 109035311A CN 201810754511 A CN201810754511 A CN 201810754511A CN 109035311 A CN109035311 A CN 109035311A
- Authority
- CN
- China
- Prior art keywords
- point
- model
- mapping
- point set
- knochenbruch
- 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.)
- Granted
Links
Classifications
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of curved bone fracture autoregistration and internal fixation steel plate pre-bending modeling methods, the following steps are included: S1: extracting the principal direction of knochenbruch model point set using PCA algorithm and choose cluster seed point along principal direction；S2: the mapping model of capsule shape is constructed respectively at the both ends of knochenbruch model；S3: clustering processing is carried out using feature point set of the OPTICS algorithm to mapping model to the mapping model constructed in S2；S4: the point set of knochenbruch model section is extracted and is divided；S5: accuracy registration is carried out to knochenbruch model according to two section point sets that segmentation obtains；S6: control point is chosen in the fracture location of knochenbruch model, the tri patch controlled in point range is subjected to thickening and handles to obtain steel plate model.This method has used many algorithms such as curve matching, cluster, characteristic point mapping, efficiently and accurately realizes the automatic Mosaic of curved bone, full-automatic registration and the steel plate pre-bending of curved bone model is realized, without carrying out manual operations.
Description
Technical field
The present invention relates to the segmentation of the section of curved bone and registration technique field more particularly to a kind of curved bone fracture autoregistration and
Internal fixation steel plate pre-bending modeling method.
Background technique
Fracture operation generallys use manual reset and hurts the method that stationary phase combines in limb at present, asks existing for this method
Topic is that wound is big, bleeding is more, is easy to cause the complication such as neural blood vessel damage.Therefore, we can use computerized algorithm pair
Knochenbruch model is virtually spliced, to obtain the various geometric parameters of steel plate in the preoperative.However current knochenbruch virtually splices
Method is applicable only to grow straight bone, and in the section of curved bone, there are no more effective full-automatic registration sides for segmentation and registration field
Method.And in the rib cage method for registering being registrated in advance based on crestal line, need to pick up point of the seed point to carry out section point set by hand
It cuts, is not carried out full automatic section point set segmentation and curved bone registration.
Summary of the invention
According to problem of the existing technology, the invention discloses a kind of curved bone fracture autoregistration and internal fixation steel plate are pre-
Curved modeling method, specifically includes the following steps:
S1: the principal direction of knochenbruch model point set is extracted using PCA algorithm and chooses cluster seed point along principal direction, using K-
The cluster seed point set of model is carried out the central point of each cluster of cluster calculation by Means clustering method, utilizes the center of cluster
It puts to be fitted the tendency calibration curve information that Bezier curve obtains knochenbruch model；
S2: the mapping model of capsule shape is constructed respectively at the both ends of knochenbruch model, by the point in knochenbruch model along its normal direction
Amount direction is mapped on mapping model；
S3: the mapping model constructed in S2 carries out at cluster the feature point set of mapping model using OPTICS algorithm
Reason carries out the rough registration that characteristic matching realizes two mapping models to the feature point set of mapping model；
S4: the point set of knochenbruch model section is extracted and is divided: the characteristic point in rough registration is subjected to region
Spontaneous length is expanded point set, concentrates in extension point and carries out matching search and obtain section point set；
S5: accuracy registration is carried out to knochenbruch model according to two section point sets that segmentation obtains: being extracted and is cut using PCA algorithm
The principal direction of millet cake collection is simultaneously registrated according to principal direction in advance, carries out smart registration using ICP algorithm；
S6: control point is chosen in the fracture location of knochenbruch model, the tri patch controlled in point range is carried out at thickening
Reason obtains steel plate model.
Further, in S2 it is specific in the following way:
According to the tangent line of the knochenbruch model that fitting obtains walked power curve and calculate the curve both ends, it is with four tangent lines respectively
Four mapping models of axis configurations capsule shape, since four mapping models are of the same size, respectively by mapping model
Point set in range is mapped on mapping model along its normal vector direction, obtains the mapping model of four containment mapping point sets.
Further, the rough registration process of two mapping models is specifically in the following way in the S3:
One of two mapping models are subjected to mirror surface transformation, match two mapping models in a coordinate system, mapping
The surface portion for penetrating cylinder in model carries out grid dividing, judges that mapping point concentrates the grid ownership of column part point set, will justify
The point set of post part is all divided into grid, counts the number of mapping point in each grid, is greater than number in the grid of threshold value
All points extract the characteristic point as mapping point set, carry out clustering processing extraction to feature point set using OPTICS algorithm
Cluster centre traverses cluster centre point, calculates the matching degree between cluster, extracts the maximum two clusters point of matching degree
Collect and obtain rotation angle, rotated knochenbruch model to reach rough registration according to rotation angle.
Further, the point set that the S 4 interrupts bone model section extracts and cutting procedure is specifically used such as lower section
By extracted in rough registration two cluster point set back mappings, corresponding point set is found in knochenbruch model surface and is made likes:
For section feature point, the section feature point set progress spontaneous length in region is obtained into two extension point sets, according to the coordinate between two o'clock
Relationship and normal direction magnitude relation carry out the search of match point, obtain the point set in two sections.
By adopting the above-described technical solution, a kind of curved bone fracture autoregistration provided by the invention and internal fixation steel plate are pre-
The autoregistration of curved bone fracture may be implemented in curved modeling method, this method, and is fitted steel plate according to the curved bone model after registration,
Algorithm proposed by the present invention can fully automatically carry out curved bone registration, and robustness is high, is adapted to the section of various complexity,
With very high precision, it can be fitted to obtain required steel plate model of performing the operation in the preoperative.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The some embodiments recorded in application, for those of ordinary skill in the art, without creative efforts,
It is also possible to obtain other drawings based on these drawings.
Fig. 1 is the flow chart of method disclosed by the invention；
Fig. 2 is that the present invention interrupts bone model curve matching effect picture；
Fig. 3 is to carry out lines detection schematic diagram to knochenbruch model in the present invention；
Fig. 4 is K-Means Clustering Effect schematic diagram in the present invention；
Fig. 5 is that mapping model schematic diagram is constructed in the present invention；
Fig. 6 is rough registration effect diagram in the present invention；
Fig. 7 is the schematic diagram of positive angle θ in the present invention；
Fig. 8 is the effect diagram of feature point extraction in the present invention；
Fig. 9 is OPTICS Clustering Effect schematic diagram in the present invention；
Figure 10 is the central point schematic diagram that cluster is extracted in the present invention；
Figure 11 is the section point set schematic diagram divided in the present invention；
Figure 12 is effect diagram of the feature point set region from after growing in the present invention；
Figure 13 is to be registrated effect diagram in advance in the present invention；
Figure 14 is the result schematic diagram of point set essence in middle section of the present invention registration；
Figure 15 is the effect diagram that the present invention interrupts bone model essence registration；
Figure 16 is to choose control point schematic diagram in model surface in the present invention；
Figure 17 is that fitting obtains steel plate model schematic in the present invention.
Specific embodiment
To keep technical solution of the present invention and advantage clearer, with reference to the attached drawing in the embodiment of the present invention, to this
Technical solution in inventive embodiments carries out clear and complete description:
A kind of curved bone fracture autoregistration as shown in Figure 1 and internal fixation steel plate pre-bending modeling method, specifically include following
Step:
It is as shown in Figure 2: S1: to extract the principal direction of knochenbruch model point set using PCA algorithm and choose cluster kind along principal direction
It is sub-, the cluster seed point set of model is carried out to the central point of each cluster of cluster calculation using K-Means clustering method, is utilized
The central point of cluster is fitted the tendency calibration curve information that Bezier curve obtains curved bone model.Concrete mode are as follows: utilize PCA algorithm
Extract the principal direction of knochenbruch model point set.Extraction effect is as shown in figure 3, PCA is the molding algorithm for extracting point set principal direction, tool
Body step are as follows:
It calculates point set center point coordinate: calculating separately the average value of the X, Y, Z coordinate of all the points in knochenbruch model, broken
The coordinate p of bone point set central point0(x0,y0,z0)。
Eigencenter: the space coordinate of knochenbruch point set is configured to 3 × n matrixN is sum a little
Amount calculates separately the coordinate of each point and center point coordinate p in matrix0Difference, obtain updated matrix A, i.e.,
It calculates covariance matrix: matrix A being multiplied with its transposed matrix A ', obtains covariance matrix M, i.e. M=AA '
Ask the characteristic value and feature vector of covariance matrix M: the corresponding vector of maximum characteristic value is the model point set
Principal direction, that is, the rectilinear direction extracted.
5 equidistant seed points are chosen along straight line, then point carries out K-Means cluster centered on this 5 points.Cluster
Effect is as shown in Figure 4.K-Means is classical clustering algorithm, the specific steps are that:
Respectively centered on 5 seed points point creation 5 cluster, to concentrate an all the points, be calculated from it away from
From nearest seed point, and the point is added to the cluster where the seed point.
After all the points are all sorted out, the central point of 5 classes is recalculated, and carry out down in this, as new seed point
One wheel iteration.When reaching the number of iterations of setting, stop iteration.
Bezier curve is fitted using the central point of cluster.The parameter of calculated curve expression formula, n rank Bezier curve is public
Formula are as follows:Wherein t be parameter, value range be [0,1), n be Bezier curve order, PiGeneration
The coordinate of i-th of central point of table.
Interpolation is carried out, in the enterprising row interpolation of curve, i.e., by giving different t, multiple points on curve are calculated
Coordinate, so that approximate fits obtain complete curve.Final effect is as shown in Figure 2.
S2: the mapping model of capsule shape is constructed respectively at the both ends of knochenbruch model, by the point in knochenbruch model along its normal direction
Amount direction is mapped on mapping model.Concrete mode are as follows: the tangent line at calculated curve both ends carries out for convenience of to each mapping model
Building, knochenbruch model in each model construction, will be moved, the endpoint curve is made to be in coordinate origin, tangent line side
To for positive direction of the x-axis.
The mapping model shape of construction be similar to capsule, centre be it is cylindric, both ends be hemispherical, the geometric center of model
For the endpoint of knochenbruch model curve, the axis of model is the above-mentioned tangent line being calculated.The cross of mapping model column part is set
Circle of contact radius is r, and the length of column part is l, then the equation in coordinates of mapping model is
Point set is mapped.To the point set in model scope, carried out respectively along the direction of its position normal vector
Mapping, i.e., the intersection point of straight line and mapping model repeats the above steps as the mapping point of the point where calculating normal vector, altogether structure
It builds to obtain 4 mapping models, such as Fig. 5.
S3: the mapping model constructed in S2 carries out at cluster the feature point set of mapping model using OPTICS algorithm
Reason carries out the rough registration that characteristic matching realizes two mapping models to the feature point set of mapping model.Detailed process is: by two
One of mapping model carries out mirror surface transformation, matches two mapping models in a coordinate system, to cylinder in mapping model
Surface portion carries out grid dividing, judges that mapping point concentrates the grid of column part point set belong to, the point set of column part is complete
Portion is divided into grid, counts the number of mapping point in each grid, and number is greater than point all in the grid of threshold value and is extracted
As the characteristic point of mapping point set, clustering processing is carried out to feature point set using OPTICS algorithm, extracts cluster centre, to poly-
Class central point is traversed, calculate cluster between matching degree, extract matching degree it is maximum two cluster and obtain rotation angle,
Knochenbruch model is rotated to reach rough registration according to rotation angle.Specific algorithm includes 4 steps:
S31: mirror image processing.Because two models to match are mirror symmetries in initial construction, for side
Just subsequent match needs first to do mirror image processing.In the method, since the geometric center of mapping model is all in origin, so mirror
As that only mapping point need to be concentrated the coordinate of the X-axis of each point become opposite number when processing.
S32: feature point extraction.Grid dividing is carried out to the surface portion of cylinder in mapping model, the height along cylinder is (parallel
X-direction) it is longitudinally divided for nxA section, along side surface, circumference is divided into nyA section, so total grid number is Ngrid=nx
×ny.Then judge that mapping point concentrates the grid ownership of column part, if the space coordinate of mapping point is (x, y, z), then its institute
Belong to grid number beL is the length of cylinder, and θ is using Y-axis positive direction as initial line, with the point and cross
Section circle center line connecting is the positive angle of end edge, such as Fig. 7.After the point set of column part is all divided into grid, each net is counted
Number is greater than point all in the grid of threshold value (being set as 3) and extracted by the number of mapping point in lattice, as mapping point set
Characteristic point, extraction effect are as shown in Figure 8.
S33:OPTICS cluster.Feature point set is clustered using OPTICS algorithm.Specific steps
(1) feature point set is traversed, finds the point of proximity (point for being less than threshold epsilon with point distance) of each point, statistics is each
The number of the corresponding point of proximity of point, if it is greater than threshold value M, then the point is core point, and calculates the core distance of the point, p point
Core distance definition is
Wherein, Nε(p) indicate that point p's closes on point set,Indicate Nε(p) in the point of p the i-th neighbour of point.
(2) two queues are established, are ordered into queue and result queue respectively.Traverse all the points in point set set, selection
One not processed core point is added in result queue, and all point of proximity of the point are added in ordered queue.Meter
Calculate in ordered queue that each point is from the reach distance of core point, from p0To p1The definition of reach distance be
Point in ordered queue is arranged according to reach distance ascending order, if ordered queue is not sky, to point therein
It is traversed, takes out a point every time, if the point is core point and not in result queue, which is added to result team
Column, and its point of proximity are added in ordered queue, if its point of proximity in ordered queue and reach distance is smaller,
Reach distance is updated, is finally resequenced to ordered queue.It repeats the above process, is completed until all characteristic points traverse.
(3) cluster result is exported.Setting threshold radius ε ' first, the sequence off-take point from result queue, if the point
Reach distance is not more than ε ', then the point belongs to current class；If the reach distance of the point is greater than ε ' and the point is core point,
Open up a new cluster；If the reach distance of the point is greater than ε ' and is not core point, which is noise point.The knot of cluster
Fruit is as shown in Figure 9.
S34: cluster centre is extracted.OPTICS cluster result is merged by processing using the method for hierarchical clustering first,
The closer classification of combined distance, and delete the few classification of element number.Specific method is: it is poly- to traverse OPTICS in numerical order
The central point of class traverses the central point of its following categories to the central point of each classification, if the distance of two central points is less than threshold
Value (is set as 4.0), then merges the two classes.After the completion of cluster, traverses and remove element number and (be set as less than threshold value
20) classification finally renumbers the cluster for updating completion.After hierarchical clustering is completed, in the coordinate that calculates each classification
Heart point, the results are shown in Figure 10.
S35: characteristic matching is carried out to mapping point collection.Specific steps are as follows:
(1) vector of the center of circle to the central point of the crosscutting circle where cluster centre point isIt willIt is counterclockwise around X-axis
Rotating the angle passed through to Z axis positive direction is α, then calculates its corresponding α to the central point of each cluster, by all α values into
Row sorts and removes coupling, is added in angle queue Q.
(2) when being compared to the feature point set of two of them mapping model, first according to mapping model head (endpoint spherical surface
Part) point set number be judged as which kind of section type.If the head point set number of two mapping models is all larger than threshold value and (sets
For 300) when, execute following method: the head point set in Ergodic Maps model two, to each point pi, in the head of mapping model one
Portion's point, which is concentrated, to be found its closest point and calculates minimum distance si, siFor the shortest distance on spherical surface, and non-space most short distance
From.Matching distance S is introduced to characterize the matching degrees of two mapping point sets, S is smaller, and to represent matching degree higher.On research head
In the method for point set,Wherein nhFor the number for mapping two midpoint of point set.If the head point of two mapping models
When collection is not both greater than threshold value, then the column part of mapping model is analyzed, the method is as follows: the cluster centre point angle of two mapping models
Spending queue is respectively Q1And Q2, to Q1Q2In all angle combinations traversed, it is assumed that the angle combinations currently traversed are
(αi,βj), wherein αi∈Q1, βj∈Q2, then by the rotation alpha counterclockwise of mapping model onei, by the rotation β counterclockwise of mapping model twoj, and
By αi, βjC is clustered where corresponding central point, C ' is set as currently clustering.If the point set number in C and C ' is more than that threshold value (is set as
500), the then column part point set in Ergodic Maps model two, to each point pi', in the column part point set of mapping model one
Middle its neighborhood point of searching, the grid model established in step before can use improve recall precision.To every kind of angle combinations,
Calculate separately its corresponding matching distance S.The case where judging a kind of C and C ' matched is first had to, calculation formula isWherein a1Represent the number of point of the match point in C ' in C, b1Represent all point sets for having match point in C
Number, a2b2Similarly.If S ' < 0.1, S=S '.Otherwise
m1And m2Respectively indicate the total number at the midpoint C and C '.
(3) all angle combinations are traversed and are calculated with matching distance S, the corresponding angle combinations of the smallest S are
The optimal registration angle of two mapping models.The matching analysis two-by-two is carried out to totally 4 mapping point sets of two knochenbruch models, altogether
4 groups of the matching analysis are carried out, wherein the smallest two mappings point set of matching distance is the mapping point set of two sections, corresponding
Registration angle is the angle of rough registration.
(4) knochenbruch model is rotated according to the angle acquired in step (3) around X-axis counterclockwise, then appropriate translation
Keep rough registration result more accurate.The distance of translation by two mapping models head point set number nh1And nh2To calculate.
As t≤0, without being moved；
As t > 0, translation vector v is calculatedt=[0 t*0.008 of t*0.016]T。
Knochenbruch model two is pressed into vtIt is translated, rough registration can be realized, the result of rough registration is as shown in Figure 6.
S4: as point set of Fig. 7-Figure 13 to knochenbruch model section extracts and divides: by the feature in rough registration
Point carries out the spontaneous length in region and is expanded point set, concentrates in extension point and carries out matching search and obtain section point set.This step will
On the basis of rough registration, the point set of knochenbruch model section is accurately extracted and divided, segmentation result such as Figure 11
It is shown.Specific step is as follows:
S41: section feature point carries out region from growth.During rough registration, two of available Optimum Matching are poly-
The two cluster point set back mappings are returned knochenbruch model (such as Fig. 6) in this step by class point set C and C ', and available two are cut
The set of characteristic points F in face1And F1', the two set of characteristic points are then subjected to more wheel spontaneous length in (being set as 5) region and are expanded
Point set F afterwards2And F2', such as Figure 12.
S42: to F2And F2' carry out matching search.The strategy of search is: traversal F2In all data points, calculate at the point
Normal vector and section at tangent line angle β, directly the point is extracted to section point set if β > 0.35 cos, is otherwise arranged one
A coordinate threshold value r1=20.0 and vector threshold r2=1.0, for the data point p of current research, in F2' middle traversal, if F2' in
Point p ' meet following three conditions simultaneously, then p and p ' is match point:
1. the squared-distance of coordinate is less than threshold value, i.e. (x1-x2)2+(y1-y2)2+(z1-z2)2<r1
②(xp+xp′)2+(yp+yp′)2+(zp+zp′)2<r2,
(xp,yp,zp)(xp′,yp′,zp') be respectively the place point p and p ' normal vector
3. the normal vector angle at the place point p and p ' is γ, γ > 0 cos.
If can find at least five p ' to point p and meet above-mentioned condition, then p be added to section point set.
S43: make F respectively2And F2' it is used as target point set to carry out traversal search twice altogether, final extract obtains two section points
Collect P1And P2, it is as shown in figure 11 to extract result.
S5: accuracy registration is carried out to knochenbruch model according to two section point sets that segmentation obtains: being extracted and is cut using PCA algorithm
The principal direction of millet cake collection is simultaneously registrated according to principal direction in advance, carries out smart registration using ICP algorithm.
Pre- registration process: the purpose being registrated in advance is that section is made substantially to coincide, and convenient for subsequent essence registration, method is: first
The principal direction that two section point sets are extracted using PCA algorithm, the direction increased using X-axis as positive direction, by two principal directions to
Amount is adjusted to positive direction.Then projection vector v of two principal direction vectors in y-o-z plane is extracted1v2, fixation of bone fragments 2,
Knochenbruch 1 is rotated by axis of main shaft, until v1v2In parallel, two sections can be accomplished greatly on main shaft direction at this time
It causes to coincide.The result being registrated in advance is as shown in figure 13.
Smart registration process: on the basis of pre- registration, carrying out accuracy registration using ICP algorithm, and detailed process is (1) by two
Part point set is denoted as U and P respectively.(2) calculate closest approach, i.e., for each of set U point, all found out in set P away from
The nearest corresponding points of the point, if the new point set being made of in set P these corresponding points is Q={ qi, i=0,1,2 ..., n }.
(3) lowest mean square root method is used, being registrated between point set U and Q is calculated, makes to obtain registration transformation matrix R, T, wherein R is 3 × 3
Spin matrix, T is 3 × 1 translation matrix.(4) calculating coordinate change, i.e., for set U, with registration transformation matrix R, T into
Row coordinate transform obtains new point set U1, i.e. U1=RU+T.(5) U is calculated1Root-mean-square error between Q, it is such as less than preset
Limiting value e, then terminate, otherwise, with point set U1U is replaced, is repeated the above steps.Figure 14 illustrates the effect of point set essence registration, figure
15 illustrate the result that two knochenbruch models are finally registrated.
S6: control point is chosen in the fracture location of knochenbruch model, the tri patch controlled in point range is carried out at thickening
Reason obtains steel plate model.User is clicked near the knochenbruch section spliced such as Figure 16, determines the general shape of steel plate model
Shape and size, record the planar delta value clicked, and all surface tri patch in selected scope calculates each triangular facet
The normal direction magnitude of piece, and record.Each plane is subjected to a degree of thickening according to its normal vector direction, and fills it
The position in gap.Obtained reinforcement is the steel plate model three-dimensional data simulated, and can export and export as result.
Figure 17 illustrates the effect of steel plate fitting.
A kind of curved bone fracture autoregistration disclosed by the invention and internal fixation steel plate pre-bending modeling method, this method are used
The many algorithms such as curve matching, cluster, characteristic point mapping, can efficiently and accurately realize the automatic Mosaic of curved bone, realize curved
The full-automatic registration of bone model and steel plate pre-bending, without carrying out manual operations.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (4)
1. a kind of curved bone fracture autoregistration and internal fixation steel plate pre-bending modeling method, it is characterised in that the following steps are included:
S1: the principal direction of knochenbruch model point set is extracted using PCA algorithm and chooses cluster seed point along principal direction, using K-
The cluster seed point set of model is carried out the central point of each cluster of cluster calculation by Means clustering method, utilizes the center of cluster
It puts to be fitted the tendency calibration curve information that Bezier curve obtains knochenbruch model；
S2: the mapping model of capsule shape is constructed respectively at the both ends of knochenbruch model, by the point in knochenbruch model along its normal vector side
To being mapped on mapping model；
S3: carrying out clustering processing using feature point set of the OPTICS algorithm to mapping model to the mapping model constructed in S2, right
The feature point set of mapping model carries out the rough registration that characteristic matching realizes two mapping models；
S4: the point set of knochenbruch model section is extracted and divided: it is spontaneous that the characteristic point in rough registration is carried out region
The long point set that is expanded carries out matching search in extension point concentration and obtains section point set；
S5: accuracy registration is carried out to knochenbruch model according to two section point sets that segmentation obtains: extracting Section Point using PCA algorithm
The principal direction of collection is simultaneously registrated according to principal direction in advance, carries out smart registration using ICP algorithm；
S6: control point is chosen in the fracture location of knochenbruch model, the tri patch controlled in point range is subjected to thickening and is handled
To steel plate model.
2. a kind of curved bone fracture autoregistration according to claim 1 and internal fixation steel plate pre-bending modeling method, feature
It also resides in: in S2 specifically in the following way:
According to the tangent line of the knochenbruch model that fitting obtains walked power curve and calculate the curve both ends, respectively using four tangent lines as axis
Four mapping models for constructing capsule shape, since four mapping models are of the same size, respectively by mapping model range
Interior point set is mapped on mapping model along its normal vector direction, obtains the mapping model of four containment mapping point sets.
3. a kind of curved bone fracture autoregistration according to claim 1 and internal fixation steel plate pre-bending modeling method, feature
Also reside in: the rough registration process of two mapping models is specifically in the following way in the S3:
One of two mapping models are subjected to mirror surface transformation, match two mapping models in a coordinate system, to mapping mould
The surface portion of cylinder carries out grid dividing, judges that mapping point concentrates the grid ownership of column part point set in type, by cylindrical portion
The point set divided is all divided into grid, counts the number of mapping point in each grid, number is greater than in the grid of threshold value and is owned
Point extract as mapping point set characteristic point, using OPTICS algorithm to feature point set carry out clustering processing extract cluster
Center traverses cluster centre point, calculates the matching degree between cluster, extracts the maximum two clusters point set of matching degree simultaneously
It obtains rotation angle, rotated knochenbruch model to reach rough registration according to rotation angle.
4. a kind of curved bone fracture autoregistration according to claim 1 and internal fixation steel plate pre-bending modeling method, feature
Also reside in: the point set that the S4 interrupts bone model section extract and cutting procedure specifically in the following way: by rough registration
In two cluster point set back mappings extracting, find corresponding point set as section feature in knochenbruch model surface
The section feature point set progress spontaneous length in region is obtained two extension point sets, according to the coordinate relationship and normal direction between two o'clock by point
Magnitude relation carries out the search of match point, obtains the point set in two sections.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810754511.0A CN109035311B (en) | 2018-07-11 | 2018-07-11 | Automatic registration and internal fixation steel plate pre-bending modeling method for curved bone fracture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810754511.0A CN109035311B (en) | 2018-07-11 | 2018-07-11 | Automatic registration and internal fixation steel plate pre-bending modeling method for curved bone fracture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109035311A true CN109035311A (en) | 2018-12-18 |
| CN109035311B CN109035311B (en) | 2021-10-01 |
Family
ID=64641927
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810754511.0A Active CN109035311B (en) | 2018-07-11 | 2018-07-11 | Automatic registration and internal fixation steel plate pre-bending modeling method for curved bone fracture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109035311B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111383353A (en) * | 2020-04-01 | 2020-07-07 | 大连理工大学 | Fractured bone model registration method based on Gaussian mixture model and contour descriptor |
| CN112785591A (en) * | 2021-03-05 | 2021-05-11 | 杭州健培科技有限公司 | Method and device for detecting and segmenting costal fracture in CT image |
| CN113624219A (en) * | 2021-07-27 | 2021-11-09 | 北京理工大学 | Magnetic compass ellipse fitting error compensation method based on OPTICS algorithm |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8078255B2 (en) * | 2006-03-29 | 2011-12-13 | University Of Georgia Research Foundation, Inc. | Virtual surgical systems and methods |
| CN104462720A (en) * | 2014-12-25 | 2015-03-25 | 河海大学常州校区 | Feature-based quick bone plate design method |
| CN105869149A (en) * | 2016-03-24 | 2016-08-17 | 大连理工大学 | Principal vector analysis based broken bone section segmentation and broken bone model registration method |
| CN106204720A (en) * | 2016-06-30 | 2016-12-07 | 深圳市智汇十方科技有限公司 | A kind of method simulating fracture steel plate pre-bend |
| CN107330281A (en) * | 2017-07-05 | 2017-11-07 | 大连理工大学 | The full-automatic personalized reconstructing method of fracture steel plate model |
| CN108154525A (en) * | 2017-11-21 | 2018-06-12 | 四川大学 | A kind of matched bone fragments joining method of feature based |
-
2018
- 2018-07-11 CN CN201810754511.0A patent/CN109035311B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8078255B2 (en) * | 2006-03-29 | 2011-12-13 | University Of Georgia Research Foundation, Inc. | Virtual surgical systems and methods |
| CN104462720A (en) * | 2014-12-25 | 2015-03-25 | 河海大学常州校区 | Feature-based quick bone plate design method |
| CN105869149A (en) * | 2016-03-24 | 2016-08-17 | 大连理工大学 | Principal vector analysis based broken bone section segmentation and broken bone model registration method |
| CN106204720A (en) * | 2016-06-30 | 2016-12-07 | 深圳市智汇十方科技有限公司 | A kind of method simulating fracture steel plate pre-bend |
| CN107330281A (en) * | 2017-07-05 | 2017-11-07 | 大连理工大学 | The full-automatic personalized reconstructing method of fracture steel plate model |
| CN108154525A (en) * | 2017-11-21 | 2018-06-12 | 四川大学 | A kind of matched bone fragments joining method of feature based |
Non-Patent Citations (3)
| Title |
|---|
| ACHIA KRONMAN ET AL: "AUTOMATIC BONE FRACTURE REDUCTION BY FRACTURE CONTACT SURFACE IDENTIFICATION AND REGISTRATION", 《2013 IEEE 10TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING》 * |
| 刘斌等: "基于断骨轴线预配准的长骨骨折钢板预弯", 《高技术通讯》 * |
| 陈宣煌: "胫骨近端骨折三维重建复位及数字钢板设计", 《中国组织工程研究》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111383353A (en) * | 2020-04-01 | 2020-07-07 | 大连理工大学 | Fractured bone model registration method based on Gaussian mixture model and contour descriptor |
| CN111383353B (en) * | 2020-04-01 | 2023-05-23 | 大连理工大学 | Fractured bone model registration method based on Gaussian mixture model and contour descriptor |
| CN112785591A (en) * | 2021-03-05 | 2021-05-11 | 杭州健培科技有限公司 | Method and device for detecting and segmenting costal fracture in CT image |
| CN112785591B (en) * | 2021-03-05 | 2023-06-13 | 杭州健培科技有限公司 | Method and device for detecting and segmenting rib fracture in CT image |
| CN113624219A (en) * | 2021-07-27 | 2021-11-09 | 北京理工大学 | Magnetic compass ellipse fitting error compensation method based on OPTICS algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109035311B (en) | 2021-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109887015B (en) | Point cloud automatic registration method based on local curved surface feature histogram | |
| CN107369161A (en) | A kind of workpiece point cloud segmentation method at random based on the European cluster of improvement | |
| CN111080684B (en) | Point cloud registration method for point neighborhood scale difference description | |
| CN103310481B (en) | A kind of point cloud compressing method based on fuzzy entropy iteration | |
| CN110111430B (en) | Method for extracting quadric surface from three-dimensional point cloud | |
| Li et al. | A leaf segmentation and phenotypic feature extraction framework for multiview stereo plant point clouds | |
| CN110599506B (en) | Point cloud segmentation method for three-dimensional measurement of complex special-shaped curved surface robot | |
| CN109903319B (en) | Multi-resolution-based fast iteration closest point registration algorithm | |
| Gélard et al. | Model-based segmentation of 3D point clouds for phenotyping sunflower plants | |
| CN107330901B (en) | Skeleton-based object component decomposition method | |
| CN108022262A (en) | A kind of point cloud registration method based on neighborhood of a point center of gravity vector characteristics | |
| CN114972459B (en) | Point cloud registration method based on low-dimensional point cloud local feature descriptor | |
| CN106504331A (en) | Tooth modeling method based on three-dimensional model search | |
| CN105678747B (en) | A kind of tooth mesh model automatic division method based on principal curvatures | |
| CN109035311A (en) | A kind of curved bone fracture autoregistration and internal fixation steel plate pre-bending modeling method | |
| CN113628263A (en) | Point cloud registration method based on local curvature and neighbor characteristics thereof | |
| CN111145129A (en) | Point cloud denoising method based on hyper-voxels | |
| CN112837309A (en) | Fruit tree canopy target recognition device and method, computing equipment and storage medium | |
| CN111652855A (en) | Point cloud simplification method based on survival probability | |
| CN117274339A (en) | Point cloud registration method based on improved ISS-3DSC characteristics combined with ICP | |
| CN115861397A (en) | Point cloud registration method based on improved FPFH-ICP | |
| CN116258857A (en) | Outdoor tree-oriented laser point cloud segmentation and extraction method | |
| CN114463396B (en) | Point cloud registration method utilizing plane shape and topological graph voting | |
| CN113223063B (en) | Tooth registration method based on ICP algorithm and point cloud elimination algorithm | |
| CN117095145B (en) | Training method and terminal of tooth grid segmentation model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |