CN110009562A - A method of comminuted fracture threedimensional model is spliced using template - Google Patents
A method of comminuted fracture threedimensional model is spliced using template Download PDFInfo
- Publication number
- CN110009562A CN110009562A CN201910068604.2A CN201910068604A CN110009562A CN 110009562 A CN110009562 A CN 110009562A CN 201910068604 A CN201910068604 A CN 201910068604A CN 110009562 A CN110009562 A CN 110009562A
- Authority
- CN
- China
- Prior art keywords
- point
- template
- fragment
- model
- matching
- 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 42
- 208000024779 Comminuted Fractures Diseases 0.000 title claims abstract description 23
- 239000012634 fragment Substances 0.000 claims abstract description 39
- 239000000284 extract Substances 0.000 claims abstract description 21
- 208000010392 Bone Fractures Diseases 0.000 claims abstract description 11
- 210000000988 bone and bone Anatomy 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 13
- 230000008859 change Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- 230000009466 transformation Effects 0.000 claims description 8
- 238000002591 computed tomography Methods 0.000 claims description 4
- 239000013598 vector Substances 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 2
- 238000000926 separation method Methods 0.000 claims description 2
- 210000001519 tissue Anatomy 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 6
- 238000011160 research Methods 0.000 description 4
- 208000027418 Wounds and injury Diseases 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 210000002303 tibia Anatomy 0.000 description 2
- 210000000689 upper leg Anatomy 0.000 description 2
- 230000008033 biological extinction Effects 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000002980 postoperative effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000002432 robotic surgery Methods 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/41—Medical
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Geometry (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Computer Graphics (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of method spliced using template to comminuted fracture threedimensional model, include four steps: the CT data of comminuted fracture patient being rebuild and pre-processed first, different tissues is split, and hollow-out part is filled up, reconstruct the threedimensional model of fragment and template;Then it is analyzed by the geometrical characteristic to fragment, extracts characteristic point on every piece of fragment and template;Then description is calculated for each characteristic point;Last matching characteristic realizes the automatic Mosaic of fragment, to restore broken bone.The present invention can the threedimensional model fragment that reconstructs of automatic Mosaic comminuted fracture CT, and accuracy with higher and clinical reference value.
Description
Technical field
The present invention relates to a kind of methods spliced using template to comminuted fracture threedimensional model, belong to virtual operation
Technical field also can be applied to relevant field, for example the measurements of the chest, waist and hips of other objects are rebuild and restored.
Background technique
Traditional splicing operation is completed by manually.People detect by an unaided eye the feature of the plane of disruption, thus it is speculated that fragment
Between script positional relationship, or even directly pick up fragment with hand to attempt to piece together them, see whether they kiss
It closes.For such method for firm not easy to wear, geometry is simple, is fairly simple for the less object of number of tiles
Easily.However, when complexity increases, fully relying on when object degree of crushing increases and manually splicing fragment, just become very
It is difficult.On the one hand, the personnel of need of work complicated in this way splicing have enough specialized capabilities, splicing experience abundant,
The structure of object to be spliced together is very familiar to.On the other hand, it even if having above-mentioned quality, to complete with guaranteeing both quality and quantity
Complicated splicing operation, still can expend a large amount of time and efforts of splicing personnel.And for the article frangible for rapid wear, make
It is just bigger using the difficulty manually spliced with traditional method.
But in practical applications, and many problems are faced with, need to splice the fragment of such complexity, reduction object is original
Looks.Such as the evidence of crime that suspect damages is restored in court, the preciousness of natural calamity damage is restored in archaeological research
Artifact, the skeleton remains etc. of the splicing extinction biology in extinct plants and animal research.Therefore, how research is by the powerful of computer
Operational capability assists even to substitute the complete fragmented splicing operation of people, is just particularly important.
And for the operation of comminuted bone splicing clinically, not only have difficulty above-mentioned, but also
It is the image that doctor combines CT scan synthesis that bigger limitation --- traditional comminuted fracture, which restores operation, by manually into
Row operation, operation is carried out directly on bone and means that and is attempted, and on the one hand increases operating time, another party
Face is observed for convenience, is needed more to expose broken bone, is caused bigger wound.
It is pre-splicing in a computer to connect if going out the threedimensional model of broken bone using CT data reconstruction, so as in surgical procedure
Reference is provided to doctor.It is considerably reduced the time-consuming of operation in this way, reduces wound.And the scheme spliced can be preoperative can
Depending on discussing verifying repeatedly in the system of change, the goodness of fit of bone is improved, post-operative recovery is more advantageous to.
The problem of three dimensional fragment automatic Mosaic there are many research, but these methods be mostly apply historical relic restore
Splice aspect with material evidence.And the clinically application of comminuted fracture operation, it is also more rare.And by the elaboration of front as it can be seen that
The solution of this problem has very big clinical value, it may have wide application prospect.
Summary of the invention
Technical problem solved by the present invention is the present invention propose it is a kind of using template to comminuted fracture threedimensional model carry out
The method of splicing can be improved accuracy as reference by existing template in splicing, while introducing after template and have more
More reference informations can reduce splicing difficulty.
A kind of the technical solution adopted by the present invention are as follows: side that comminuted fracture threedimensional model is spliced using template
Method is rebuild and is pre-processed to the CT data of comminuted fracture patient first, obtains the threedimensional model of fragment.Then by pair
The geometrical characteristic of fragment is analyzed, and characteristic point is extracted, and calculates description;Last matching characteristic realizes the automatic Mosaic of fragment,
To restore broken bone, specifically comprise the following steps:
The first step, the reconstruction and pretreatment of model are carried out the data of CT scan pre- using medical treatment software MIMICS
Processing, the pretreatment include: that Threshold segmentation carries out background separation, and region-growing method extracts, denoises and simplify;
Second step extracts characteristic point, before matching first using ISS (Intrinsic Shape Signatures, inside
Shape feature description) the apparent point of local feature is extracted to each broken bone model that the first step reconstructs respectively, with this
As characteristic point, matched complexity can be reduced in this way, the process for extracting characteristic point includes: definition local coordinate system,
The weight on each vertex of computation model constructs the covariance matrix of each point, calculates characteristic value, extracts characteristic point;
Third step, calculates description, calculates SHOT to each characteristic point on each model extracted in second step
(Signature of Histograms of OrienTations, direction histogram description) is used as Feature Descriptor, embodies
With the partial structurtes around characteristic point, the process for calculating Feature Descriptor includes: to establish local frame of reference, statistic histogram
Feature;
4th step, template matching respectively describe the feature of each comminuted fracture fragment model reconstructed and template
Son is matched, and each broken bone model is snapped in template, to complete to splice;The process of the template matching includes:
Building matching set, screens matching set, calculates and using transformation matrix.
The first step, the reconstruction and pretreatment of model are implemented as follows:
(1) skeleton density value is provided in MIMICS, selecting system to separate bone from background as threshold value;
(2) region-growing method then is executed to the bone that step (1) extracts, to separate fragment;
(3) finally the model separated in step (2) is imported in Materialise 3-matic software and is denoised
It is handled with simplifying, can finally reconstruct required bone threedimensional model.
The second step extracts characteristic point, is implemented as follows:
The data on the vertex that each threedimensional model reconstructed in the first step is included constitute cloud, if a point
Cloud data have n point (xi,yi,zi), i=0,1 ..., n-1 remember pi=(xi,yi,zi), then extract the specific steps of feature such as
Under:
(1) to point p each of on cloudiA local coordinate system is defined, and gives one radius r of each pointframe;
Inquire the p of each point in point cloud dataiRadius rframeAll the points in surrounding, and calculate the power that each pair of point is answered
Value wij, wij=1/ | | pi-pj| |, | pi|<rframe, wherein piAnd pjIt is a three-dimensional vector, represents the coordinate of the point;
It (2) is each point piConstruct covariance matrix cov (pi):
(3) each point p is calculatediCovariance matrix cov (pi) characteristic valueAnd by arranging from big to small;
(4) threshold epsilon is set1With ε2, will meet simultaneouslyWithPoint be considered as characteristic point, thus complete mention
It takes.
The third step calculates description and is implemented as follows:
(1) information for being primarily based on feature vertex neighborhood establishes LRF (local reference frame, local frame of reference),
The spherical surface neighborhood of characteristic point is divided according to radial direction, longitude and latitude, is radially divided into two parts, and longitude is divided into eight parts, latitude
Degree is divided into two parts, amounts to 2*2*8=32 region;
(2) histogram feature in each region is then counted using bilinear interpolation, is divided into 11 according to cosine value
Section, the length of SHOT descriptor are 32*11=352.
4th step, template matching are implemented as follows:
(1) for template and each fragment FiWith template T, point p ∈ F is takeni, point q ∈ T, if met simultaneously:
A.Dis (p, q)=| S (p)-S (q) | < δD, wherein S (p) and S (q) respectively represents SHOT description of p and q;
B. for each point p, there is q ∈ T, so that Dis (p, q) reaches minimum value;
Then this pair of p and q is added in the set M of qualified characteristic point pair, matching set is constructed with this;
(2) three groups of point (p are randomly selected every time1,q1),(p2,q2),(p3,q3) ∈ M, it is then screened, is become based on rigidity
It changes, if Euler's distance difference between two on fragment point, two points corresponding in template is receiving in range, depending on
For matching, i.e., meet simultaneously:
|||p1p2||-||q1q2|||<δp, | | | p2p3||-||q2q3|||<δp, | | | p3p1||-||q3q1|||<δp, wherein
δpFor the allowable error range of setting, | | p1p2| | indicate p1To p2Euclidean distance, then by matched point to retaining, i.e.,
It completes to screen matching set;
(3) each fragment F is utilizediCharacteristic point searches out corresponding point in template in matching set, then calculates change
Change matrix Ti, then to each fragment FiUsing transformation TiCan fragment and template alignment can be completed entirely spliced in this way
Journey.
The advantages of the present invention over the prior art are that: it is that doctor sweeps in conjunction with CT that traditional comminuted fracture, which restores operation,
The image for retouching synthesis is carried out operation directly on bone and is meant that and attempted by what is operated manually, a side
Face increases operating time, on the other hand observes for convenience, needs more to expose broken bone, cause bigger wound.This hair
It is bright that comminuted fracture threedimensional model is spliced using template, it can be improved by existing template as reference in splicing
Accuracy, while introducing after template and have more reference informations, it can reduce and splice difficulty.
Detailed description of the invention
Fig. 1 is the process flow diagram for the method that the present invention splices comminuted fracture threedimensional model using template;
Fig. 2 is the broken bone model sample reconstructed in Materialise 3-matic of the invention;
Fig. 3 is the characteristic point schematic diagram that the present invention extracts;
Fig. 4 is LRF (local reference frame) local referential schematic diagram of the invention;
Fig. 5 is the erroneous matching schematic diagram that present invention needs are rejected;
Fig. 6 is the women shin bone schematic diagram spliced with the method for the invention;
Fig. 7 is the women femur schematic diagram spliced with the method for the invention.
Specific embodiment
The following describes the present invention in detail with reference to the accompanying drawings and embodiments.
As shown in Figure 1, a kind of processing for the method spliced using template to comminuted fracture threedimensional model of the present invention
Flow chart is divided into four steps, the reconstruction of model and preprocessing process, extracts feature point process, calculates description subprocess, template
Matching process.Four steps execute in the order described above, extract characteristic point respectively to the model rebuild and pre-processed after completing,
It then is each characteristic point calculating description extracted, one by one by description on each fragment and description in template finally
It is corresponding, to complete to splice.
The present invention specifically includes following four step.
The reconstruction and pretreatment of model: the first step utilizes medical treatment software MIMICS, the data reconstruction of CT scan is gone out
Broken bone model.Since the model that CT data reconstruction comes out excessively includes details, matching can be impacted, it is unnecessary also to will increase
Operation.So also need model to carry out certain processing, including region growth method is split, remove impurity, filling cavity with
And model simplification.
Second step, extract characteristic point: using CT reconstruct come model accuracy it is relatively high, comprising largely put and details.
If directly matched to these points, operand is still very huge.Therefore, it is first extracted before matching using ISS algorithm
The apparent point of local feature out, to reduce matched complexity.
Third step calculates description: SHOT description is calculated to the characteristic point that extracts, during reflection with characteristic point is
The heart, the position of the point within the scope of specified radius and the distribution of normal vector, to embody with the partial structurtes around characteristic point.
4th step, template matching: the matching of the plane of disruption needs a preferable relative position, directly to plane of disruption progress
That matches is ineffective, is difficult to converge to ideal error, therefore to consider to find a preferable initial position with template matching.
Template matching is substantially exactly to match respectively to feature point description of fragment and template.
It is specifically described below.
Step 1, the reconstruction of model and preprocessing process:
The CT data got need just to can be carried out subsequent splicing operation through over-segmentation and pretreatment.Materialise
Mimics (Materialise's interactive medical image control system) is the doctor of a profession
Image processing software is learned, CT data are split using the software, is denoised, the processing such as filling-up hole.
(1) skeleton density value is provided in MIMICS, in selecting system to separate bone from background as threshold value;
(2) region-growing method then is executed to the bone that (1) extracts, to separate fragment;
(3) finally the model separated in (2) is imported in Materialise 3-matic software and carries out denoising and letter
Change processing, can finally reconstruct required bone threedimensional model, as shown in Fig. 2, being gone out in software using CT data reconstruction
The threedimensional model of one complete bone.
Step 2 extracts feature point process:
Characteristic point detection is widely used in object matching, the scenes such as target following and three-dimensional reconstruction.Common feature includes color,
Angle point, characteristic point, profile and texture.In the detection of characteristic point, need to consider scale invariability, rotational invariance and antinoise
Effect, these contribute to the whether stable important indicator of evaluating characteristic point.
If point cloud data has n point (xi,yi,zi), i=0,1 ..., n-1.Remember pi=(xi,yi,zi), then extract feature
Specific step is as follows:
(1) to point p each of on cloudiA local coordinate system is defined, and gives each one search radius of point
rframe;
(2) each point p in point cloud data is inquirediRadius rframeAll the points in surrounding, and its weight is calculated,
wij=1/ | | pi-pj| |, | pi|<rframe;
(3) each point p is calculatediCovariance matrix:
(4) each point p is calculatediCovariance matrix cov (pi) characteristic valueAnd by arranging from big to small;
(5) threshold epsilon is set1With ε2, meetWithPoint be then considered as characteristic point.
The characteristic point sample that extracts as shown in figure 3, these characteristic points are distributed in the apparent region of model local feature,
It is representative.
Step 3 calculates description subprocess:
Description is calculated to each characteristic point.In order to adapt to the difference between fragment and template, using SHOT describe son come
Indicate the local feature of each point, it can combine geometry distributed intelligence with statistics with histogram.
(1) information for being primarily based on feature vertex neighborhood establishes LRF (local reference frame), as shown in figure 4,
The spherical surface neighborhood of characteristic point is divided according to radial direction, longitude and latitude.Radial direction is divided into two parts, and longitude is divided into eight parts, latitude
Degree is divided into two parts, amounts to 32 regions;
(2) histogram feature in each region is then counted using bilinear interpolation.Cosine value is divided into 11 sections,
Therefore the length of SHOT descriptor is 32*11=352.
Step 4, template matching process:
Template matching needs to obtain one in advance with the complete bone of the bone same area fractured as template, general
Fragment is aligned with the template, so that bone substantially be restored, while also providing ideal for subsequent fine splicing
Initial position.
(1) there is p ∈ F for template and each fragmenti, q ∈ T, if met:
A.Dis (p, q)=| D (p)-D (q) | < δD.
B. for each p, there is q ∈ T that Dis (p, q) is made to reach minimum value;
Then they are added in the set M of qualified characteristic point pair;
(2) still there can be undesirable point pair by processing in this way, because p and q are substantially 352 dimensional vectors,
It is to be judged according to the mould of two description son differences, as shown in figure 5, by the processing of (1), most point is to can complete
It mixes, but still has the point of a small number of matching errors.Therefore, it is necessary to matched feature to advanced optimizing.It is of the invention herein
The thinking for having used for reference RANSAC (Random Sample Consensus) algorithm, randomly selects three groups of point (p every time1,q1),(p2,
q2),(p3,q3) ∈ M, then screened.Because it is contemplated that be all rigid transformation, so if two points and mould on fragment
Euler's distance difference on plate between corresponding two points is within an acceptable range, so that it may be considered as matching.Meet:
|||p1p2||-||q1q2|||<δp
|||p2p3||-||q2q3|||<δp
|||p3p1||-||q3q1|||<δp
(3) each fragment F is utilizediCharacteristic point searches out corresponding point in template in matching set, then calculates change
Change matrix Ti, then to each fragment FiUsing transformation TiCan fragment and template alignment can be completed entirely spliced in this way
Journey.
As shown in fig. 6, being the women shin bone experiment effect figure spliced with the method for the invention.As shown in fig. 7, being with this
The women femur experiment effect figure of invention the method splicing.In figure 6 and figure 7, the left side respectively be input splicing before it is broken
Piece, the right are the model for completing fragments mosaicing of output respectively.
The present invention can the threedimensional model fragment that reconstructs of automatic Mosaic comminuted fracture CT, and accuracy with higher and
Clinical reference value.The method of proposition of the invention is also applied in virtual operation and robotic surgery.
The technology contents that the present invention does not elaborate belong to the well-known technique of those skilled in the art.
Although the illustrative specific embodiment of the present invention is described above, in order to the technology people of this technology neck
Member understands the present invention, it should be apparent that the present invention is not limited to the range of specific embodiment, to the ordinary skill of the art
For personnel, as long as various change is in the spirit and scope of the present invention that the attached claims limit and determine, these become
Change is it will be apparent that all utilize the innovation and creation of present inventive concept in the column of protection.
Claims (5)
1. a kind of method spliced using template to comminuted fracture threedimensional model, which comprises the steps of:
The first step, the reconstruction and pretreatment of model are located the data of CT scan using medical treatment software MIMICS in advance
Reason, the pretreatment include: that Threshold segmentation carries out background separation, and region-growing method extracts, denoises and simplify;
Second step extracts characteristic point, first uses ISS (Intrinsic Shape Signatures, interior shape before matching
Feature Descriptor) the apparent point of local feature is extracted to each broken bone model that the first step reconstructs respectively, in this, as
Characteristic point can reduce matched complexity in this way, and the process for extracting characteristic point includes: definition local coordinate system, is calculated
The weight on each vertex of model constructs the covariance matrix of each point, calculates characteristic value, extracts characteristic point;
Third step, calculates description, calculates SHOT to each characteristic point on each model extracted in second step
(Signature of Histograms of OrienTations, direction histogram description) is used as Feature Descriptor, embodies
With the partial structurtes around characteristic point, the process for calculating Feature Descriptor includes: to establish local frame of reference, statistic histogram
Feature;
4th step, template matching, respectively to the Feature Descriptor of each comminuted fracture fragment model reconstructed and template into
Row matching, each broken bone model is snapped in template, to complete to splice;The process of the template matching includes: building
Matching set screens matching set, calculates and using transformation matrix.
2. a kind of method spliced using template to comminuted fracture threedimensional model according to claim 1, special
Sign is: the first step, and the reconstruction and pretreatment of model are implemented as follows:
(1) skeleton density value is provided in MIMICS, selecting system to separate bone from background as threshold value;
(2) region-growing method then is executed to the bone that step (1) extracts, to separate fragment;
(3) finally the model separated in step (2) is imported in Materialise 3-matic software and carries out denoising and letter
Change processing, can finally reconstruct required bone threedimensional model.
3. a kind of method spliced using template to comminuted fracture threedimensional model according to claim 1, special
Sign is: the second step, extracts characteristic point, is implemented as follows:
The data on the vertex that each threedimensional model reconstructed in the first step is included constitute cloud, if a cloud number
According to there is n point (xi,yi,zi), i=0,1 ..., n-1 remember pi=(xi,yi,zi), then extracting feature, specific step is as follows:
(1) to point p each of on cloudiA local coordinate system is defined, and gives one radius r of each pointframe;
Inquire the p of each point in point cloud dataiRadius rframeAll the points in surrounding, and calculate the weight that each pair of point is answered
wij, wij=1/ | | pi-pj| |, | pi|<rframe, wherein piAnd pjIt is a three-dimensional vector, represents the coordinate of the point;
It (2) is each point piConstruct covariance matrix cov (pi):
(3) each point p is calculatediCovariance matrix cov (pi) characteristic valueAnd by arranging from big to small;
(4) threshold epsilon is set1With ε2, will meet simultaneouslyWithPoint be considered as characteristic point, to complete to extract.
4. a kind of method spliced using template to comminuted fracture threedimensional model according to claim 1, special
Sign is: the third step, calculates description and is implemented as follows:
(1) information for being primarily based on feature vertex neighborhood establishes LRF (local reference frame, local frame of reference), feature
The spherical surface neighborhood of point is divided according to radial direction, longitude and latitude, is radially divided into two parts, longitude is divided into eight parts, minute of latitude
For two parts, amount to 2*2*8=32 region;
(2) histogram feature in each region is then counted using bilinear interpolation, 11 sections is divided into according to cosine value,
The length of SHOT descriptor is 32*11=352.
5. a kind of method spliced using template to comminuted fracture threedimensional model according to claim 1, special
Sign is: the 4th step, template matching are implemented as follows:
(1) for template and each fragment FiWith template T, point p ∈ F is takeni, point q ∈ T, if met simultaneously:
A.Dis (p, q)=| S (p)-S (q) | < δD, wherein S (p) and S (q) respectively represents SHOT description of p and q;
B. for each point p, there is q ∈ T, so that Dis (p, q) reaches minimum value;
Then this pair of p and q is added in the set M of qualified characteristic point pair, matching set is constructed with this;
(2) three groups of point (p are randomly selected every time1,q1),(p2,q2),(p3,q3) ∈ M, it is then screened, is based on rigid transformation,
If Euler's distance difference between two points, two points corresponding in template on fragment is receiving in range, it is considered as
Match, i.e., meet simultaneously:
|||p1p2||-||q1q2|||<δp, | | | p2p3||-||q2q3|||<δp, | | | p3p1||-||q3q1|||<δp, wherein δpTo set
Fixed allowable error range, | | p1p2| | indicate p1To p2Euclidean distance, then matched point is completed to retaining
Matching set is screened;
(3) each fragment F is utilizediCharacteristic point searches out corresponding point in template in matching set, then calculates transformation matrix
Ti, then to each fragment FiUsing transformation TiEntire splicing can be can be completed in fragment and template alignment in this way.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910068604.2A CN110009562A (en) | 2019-01-24 | 2019-01-24 | A method of comminuted fracture threedimensional model is spliced using template |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910068604.2A CN110009562A (en) | 2019-01-24 | 2019-01-24 | A method of comminuted fracture threedimensional model is spliced using template |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110009562A true CN110009562A (en) | 2019-07-12 |
Family
ID=67165545
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910068604.2A Pending CN110009562A (en) | 2019-01-24 | 2019-01-24 | A method of comminuted fracture threedimensional model is spliced using template |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110009562A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113298784A (en) * | 2021-05-25 | 2021-08-24 | 中南大学 | Bone fracture line extraction and mapping method based on visual angle |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103778662A (en) * | 2014-01-07 | 2014-05-07 | 北京师范大学 | Virtual restoration method for interactive broken relics |
CN105809118A (en) * | 2016-03-03 | 2016-07-27 | 重庆中科云丛科技有限公司 | Three-dimensional object identifying method and apparatus |
WO2016161311A1 (en) * | 2015-04-02 | 2016-10-06 | The New York Stem Cell Foundation | In vitro methods for assessing tissue compatibility of a material |
CN106780460A (en) * | 2016-12-13 | 2017-05-31 | 杭州健培科技有限公司 | A kind of Lung neoplasm automatic checkout system for chest CT image |
CN106934228A (en) * | 2017-03-06 | 2017-07-07 | 杭州健培科技有限公司 | Lung's pneumothorax CT image classification diagnostic methods based on machine learning |
CN107843261A (en) * | 2017-10-31 | 2018-03-27 | 国网黑龙江省电力有限公司检修公司 | A kind of method and system based on laser scanning data positioning robot position |
CN108154525A (en) * | 2017-11-21 | 2018-06-12 | 四川大学 | A kind of matched bone fragments joining method of feature based |
CN108470358A (en) * | 2018-02-13 | 2018-08-31 | 南京旅游职业学院 | A kind of point cloud registration method for shaking artificial bee colony algorithm based on second order |
-
2019
- 2019-01-24 CN CN201910068604.2A patent/CN110009562A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103778662A (en) * | 2014-01-07 | 2014-05-07 | 北京师范大学 | Virtual restoration method for interactive broken relics |
WO2016161311A1 (en) * | 2015-04-02 | 2016-10-06 | The New York Stem Cell Foundation | In vitro methods for assessing tissue compatibility of a material |
CN105809118A (en) * | 2016-03-03 | 2016-07-27 | 重庆中科云丛科技有限公司 | Three-dimensional object identifying method and apparatus |
CN106780460A (en) * | 2016-12-13 | 2017-05-31 | 杭州健培科技有限公司 | A kind of Lung neoplasm automatic checkout system for chest CT image |
CN106934228A (en) * | 2017-03-06 | 2017-07-07 | 杭州健培科技有限公司 | Lung's pneumothorax CT image classification diagnostic methods based on machine learning |
CN107843261A (en) * | 2017-10-31 | 2018-03-27 | 国网黑龙江省电力有限公司检修公司 | A kind of method and system based on laser scanning data positioning robot position |
CN108154525A (en) * | 2017-11-21 | 2018-06-12 | 四川大学 | A kind of matched bone fragments joining method of feature based |
CN108470358A (en) * | 2018-02-13 | 2018-08-31 | 南京旅游职业学院 | A kind of point cloud registration method for shaking artificial bee colony algorithm based on second order |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113298784A (en) * | 2021-05-25 | 2021-08-24 | 中南大学 | Bone fracture line extraction and mapping method based on visual angle |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110599508B (en) | Artificial intelligence-based spine image processing method and related equipment | |
US20230360313A1 (en) | Autonomous level identification of anatomical bony structures on 3d medical imagery | |
Fürnstahl et al. | Computer assisted reconstruction of complex proximal humerus fractures for preoperative planning | |
CN105228548B (en) | The designing system and method for operative correction for abnormal bone | |
CN110459301B (en) | Brain neurosurgery navigation registration method based on thermodynamic diagram and facial key points | |
US8743119B2 (en) | Model-based face image super-resolution | |
EP2402907B1 (en) | Information processing apparatus, three-dimensional position calculation method, and program | |
CN112132166B (en) | Intelligent analysis method, system and device for digital cell pathology image | |
CN111932492B (en) | Medical image processing method and device and computer readable storage medium | |
US11935213B2 (en) | Laparoscopic image smoke removal method based on generative adversarial network | |
CN110188694A (en) | It is a kind of that shoeprints recognition sequence method is worn based on pressure characteristic | |
Wei et al. | Skull assembly and completion using template-based surface matching | |
Pei et al. | Improved brain tumor growth prediction and segmentation in longitudinal brain MRI | |
Piras et al. | Current options for visualization of local deformation in modern shape analysis applied to paleobiological case studies | |
CN110009562A (en) | A method of comminuted fracture threedimensional model is spliced using template | |
CN113706514A (en) | Focus positioning method, device and equipment based on template image and storage medium | |
CN112102327B (en) | Image processing method, device and computer readable storage medium | |
CN110084270A (en) | Pathological section image-recognizing method and equipment | |
CN114782454A (en) | Image recognition system for preoperative navigation of pelvic tumor images | |
Wang et al. | Automated labeling of the airway tree in terms of lobes based on deep learning of bifurcation point detection | |
Bin et al. | Rapid multimodal medical image registration and fusion in 3D conformal radiotherapy treatment planning | |
WO2023143625A1 (en) | Process and system for three-dimensional modelling of tissue of a subject, and surgical planning process and system | |
CN117710711B (en) | Optical and SAR image matching method based on lightweight depth convolution network | |
Hu et al. | A study on automatic assembly method of fractured bone fragments based on measured point clouds | |
CN110047151B (en) | Fragment fracture surface splicing method and system based on thermonuclear characteristics |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190712 |