CN108898627A - A kind of Model registration method and apparatus based on characteristic point - Google Patents
A kind of Model registration method and apparatus based on characteristic point Download PDFInfo
- Publication number
- CN108898627A CN108898627A CN201810794495.8A CN201810794495A CN108898627A CN 108898627 A CN108898627 A CN 108898627A CN 201810794495 A CN201810794495 A CN 201810794495A CN 108898627 A CN108898627 A CN 108898627A
- Authority
- CN
- China
- Prior art keywords
- region
- point
- source
- coordinate
- divides
- 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
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
-
- 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
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
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 purpose of the application is to provide a kind of Model registration method based on characteristic point, and the application identifies the characteristic point and its coordinate of preset quantity on source model;Region division is carried out to all characteristic points using preset regions partitioning, and at least one location point in region is divided to obtained each source and carries out topological coding, obtains the topological relation of each location point;Obtain object module to be registered;From the target feature point and its coordinate for identifying, preset quantity corresponding with the characteristic point in source model in object module, region division is carried out to all target feature points based on preset regions partitioning, each target is obtained and divides the corresponding target feature point in region and its coordinate;And combine the topological relation of the location point in source model, all location points and its coordinate being registrated out in object module in each target division region corresponding with source model, being registrated one by one to source model and the location point in object module is realized, and improves the accuracy of registration.
Description
Technical field
This application involves computer field more particularly to a kind of Model registration method and apparatus based on characteristic point.
Background technique
As shown in Figure 1, being transformed to for the face in the prior art by the face and standard header of 3D scanning with geometric coordinate
Then the point point correspondence between scan model and standard header is established in similar position by orthographic projection, realize that standard is thrown
In face be registrated with the projection between the face of actual scanning.It will definitely be according to feature although projection mapping in the prior art is matched
The corresponding relationship of the position of point calculates the transformation matrix of projection, but cannot be guaranteed that each characteristic point is after projection in source model
The individual features point that object module can be corresponded to, can not handle the mapping one by one of non-characteristic point, cause the accuracy of registration very low.
Summary of the invention
The purpose of the application is to provide a kind of Model registration method and apparatus based on characteristic point, to solve existing skill
The low problem of accuracy caused by being registrated in art based on characteristic point.
According to the one aspect of the application, a kind of Model registration method based on characteristic point is provided, wherein the method
Including:
The characteristic point and its coordinate of preset quantity are identified on source model;Based on preset regions partitioning and the characteristic point
And its all characteristic points of coordinate pair carry out region division, and divide at least one in region to obtained each source respectively
Location point carries out topological coding, obtains the topological relation at least one location point that each source divides in region;
Obtain object module to be registered;
From the target for identifying preset quantity corresponding with the characteristic point in the source model, described in the object module
Characteristic point and its coordinate, and region division is carried out to all target feature points based on the preset regions partitioning, it obtains
Each target in object module divides the corresponding target feature point in region and its coordinate;
Based on the topological relation and the corresponding target feature point and its coordinate, it is registrated out corresponding with the source model
, each target in the object module divide at least one location point and its coordinate in region.
Further, in the above method, each characteristic point in the source model has one in the object module
A corresponding target feature point.
Further, in the above method, the characteristic point and its coordinate that preset quantity is identified on source model;Based on pre-
It sets region division method and the characteristic point and its all characteristic points of coordinate pair carries out region division, and is every to what is obtained respectively
A source divides at least one location point in region and carries out topological coding, obtains at least one position in each source division region
The topological relation of point, including:
Preset source model is obtained, the characteristic point of identification and the preset preset quantity in the source model, and obtain
The position of each characteristic point;
Based on the preset regions partitioning and the characteristic point and its coordinate, the characteristic point of the preset quantity is carried out
Region division obtains each source in the source model and divides the corresponding characteristic point in region and its coordinate;
Each source in source model divides in region respectively, corresponding based on the corresponding characteristic point and its coordinate pair
The source divides at least one location point in region and carries out topological coding, obtains at least one position in each source division region
Set topological relation a little.
Further, in the above method, each source in source model respectively is divided in region, is based on the correspondence
Characteristic point and its corresponding source of coordinate pair divide at least one location point in region and carry out topological coding, obtain each
Source divides the topological relation of at least one location point in region, including:
Obtain topological coding method corresponding with the preset regions partitioning;
Each source in the source model divides in region respectively, is based on the topological coding method and the corresponding spy
Sign point and its coordinate divide at least one location point in region to the corresponding source and carry out topological coding, obtain each source
Divide the topological relation of at least one location point in region.
Further, described from being identified in the object module and the characteristic point in the source model in the above method
The target feature point and its coordinate of corresponding, the described preset quantity, and based on the preset regions partitioning to all mesh
It marks characteristic point and carries out region division, obtain each target in object module and divide the corresponding target feature point in region and its seat
Mark, including:
From the target for identifying preset quantity corresponding with the characteristic point in the source model, described in the object module
Characteristic point, and obtain from the object module position of the target feature point;
It is special to the target of the preset quantity based on the predeterminable area partitioning and the target feature point and its coordinate
Sign point carries out region division, obtains each target in the object module and divides the corresponding target feature point in region and its seat
Mark,
Wherein, each source in the source model divides region in the object module in the presence of a corresponding target
Divide region.
Further, described based on the topological relation and the corresponding target feature point and its seat in the above method
Mark is registrated out at least one position in each target division region in object module corresponding with the source model, described
Point and its coordinate, including:
Each target in the object module divides in region,
It is based respectively on the topological relation and the corresponding target feature point and its coordinate, the target is registrated out and divides
At least one location point and its coordinate in region,
Wherein, it is corresponding in the presence of one in target division region to divide each location point in region for the source
Location point.
Further, in the above method, the preset regions partitioning includes any one of following:
De Laode Deaunay triangle division method, tetra- jiaos of partitionings of Deaunay and bezier surface partitioning.
According to the another aspect of the application, a kind of Model registration method based on characteristic point is additionally provided, wherein the side
Method includes:
Source model and object module are obtained respectively;
Characteristic point and its position of preset quantity are identified on the source model, meanwhile, know on the object module
Do not go out, with the one-to-one target feature point of each characteristic point on the source model, and obtain the position of the target feature point
It sets;
Based on the preset regions partitioning, while carrying out region division to all characteristic points, to all described
Target feature point carries out region division, respectively obtains each source in the source model and divides the corresponding characteristic point in region and its position
It sets, each target in the object module divides the corresponding target feature point in region and its position;
At least one location point in region is divided to each source respectively and carries out topological coding, each source is obtained and divides region
The topological relation of at least one interior location point;
Based on the topological relation and the corresponding target feature point and its coordinate, it is registrated out corresponding with the source model
, each target in the object module divide at least one location point and its coordinate in region.
According to the another aspect of the application, a kind of non-volatile memory medium is additionally provided, being stored thereon with computer can
Reading instruction when the computer-readable instruction can be executed by processor, realizes the processor as described above based on feature
The Model registration method of point.
According to the another aspect of the application, a kind of Model registration equipment based on characteristic point is additionally provided, wherein described to set
It is standby to include:
One or more processors;
Non-volatile memory medium, for storing one or more computer-readable instructions,
When one or more of computer-readable instructions are executed by one or more of processors, so that one
Or multiple processors realize the Model registration method as described above based on characteristic point.
Compared with prior art, characteristic point and its coordinate that the application passes through the identification preset quantity on source model;It is based on
Preset regions partitioning and all characteristic points of the characteristic point and its coordinate pair carry out region division, and change respectively region
The each source obtained after point divides at least one location point in region and carries out topological coding, obtains each source and divides in region
The topological relation of at least one location point realizes the Feature point recognition to source model, characteristic point region division and each source and draws
The topology coding of all location points in subregion, and then the topological relation of each location point is obtained, so as to subsequent to the source mould
Type is corresponding, actual object module carries out that the registration that the topological relation carries out object module can be called directly with punctual, from
And effectively improve the subsequent registration efficiency to object module;It is punctual when needing to carry out object module to match, it obtains to be registered
Object module;Referring to the characteristic point identified in source model, from being identified in the object module and the spy in the source model
The target feature point and its coordinate of corresponding, the described preset quantity of sign point, and based on the preset regions partitioning to target mould
All target feature points in type carry out region division, obtain each target in object module and divide the corresponding mesh in region
Mark characteristic point and its coordinate;It is divided in region in each target of object module, is based on the topological relation and target dividing regions
The corresponding target feature point in domain and its coordinate, each target being registrated out in object module corresponding with the source model, described
At least one location point and its coordinate in region are divided, so that realizing each location point in source model based on topological relation
With the mapping one by one between each location point in object module, and then the accuracy to the registration of object module is improved.
Detailed description of the invention
By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other
Feature, objects and advantages will become more apparent upon:
Fig. 1 shows the method for registering schematic diagram between the face in the face and standard header of scanning in the prior art;
Fig. 2 shows the flow diagrams according to a kind of Model registration method based on characteristic point of the application one aspect;
Fig. 3 shows the faceform and its spy in a kind of Model registration method based on characteristic point of the application one aspect
Sign point distribution schematic diagram;
Fig. 4 show target faceform in a kind of Model registration method based on characteristic point of the application one aspect and
Its target feature point distribution schematic diagram;
Fig. 5 shows faceform and target in a kind of Model registration method based on characteristic point of the application one aspect
The registration schematic diagram of faceform.
The same or similar appended drawing reference represents the same or similar component in attached drawing.
Specific embodiment
The application is described in further detail with reference to the accompanying drawing.
In a typical configuration of this application, terminal, the equipment of service network and trusted party include one or more
Processor (CPU), input/output interface, network interface and memory.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium
Example.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable
Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM),
Digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or
Any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, computer
Readable medium does not include non-temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.
Fig. 2 shows a kind of Model registration methods based on characteristic point according to the application one aspect, wherein the method
Including step S11, step S12, step S13 and step S14, specifically include:
The step S11 identifies the characteristic point and its coordinate of preset quantity on source model;Herein, enterprising in source model
The algorithm of row Feature point recognition may include being but not limited to Scale invariant features transform (Scale-invariant
Feature transform, SIFT) etc., to realize extraction and identification to the characteristic point of the key position on source model;
Then, the step S11 is based on preset regions partitioning and the characteristic point and its all features of coordinate pair
Point carries out region division, and at least one location point in each source division region obtained after changing point to region respectively is opened up
Coding is flutterred, the topological relation at least one location point that each source divides in region is obtained, is realized by step S11 to source mould
The each source obtained after the Feature point recognition of type, the region division of the characteristic point of preset quantity and region division divides in region
All location points topology coding, closed to obtain each source in source model and divide the topology of each location point in region
System carries out with punctual corresponding with source model, actual object module so as to subsequent, can call directly the topological relation into
The registration of row object module, to effectively improve the subsequent registration efficiency to object module.
Here, the preset regions partitioning can include but is not limited to include any one of following:De Laode Deaunay tri-
Angle partitioning, tetra- jiaos of partitionings of Deaunay and bezier surface partitioning.Certainly, what other will be likely to occur from now on is used for source
The preset regions partitioning that characteristic point in model or object module carries out region division is such as applicable to the application, then wraps
It includes in the application.It is illustrated by taking Deaunay triangle division method as an example below.
It is punctual when needing to carry out object module to match, it first passes through step S12 and obtains object module to be registered;The step
S13 is referring to the characteristic point identified in source model, from identifying in the object module and the characteristic point pair in the source model
The target feature point and its coordinate of the preset quantity answer, described;In other words, in the source model in the step S11
It is special to there is a corresponding target in the object module in the step S13 in each characteristic point identified
Levy point, with guarantee object module identify with the one-to-one target feature point of characteristic point in source model, be convenient for subsequent base
All location points in object module are registrated in corresponding target feature point;Then the step S13 is based on described
Preset regions partitioning carries out region division to all target feature points in object module, obtains every in object module
A target divides the corresponding target feature point in region and its coordinate;Step S14 is divided in region in each target of object module,
The corresponding target feature point in region and its coordinate are divided based on the topological relation and target, is registrated out corresponding with the source model
, each target in the object module divide at least one location point and its coordinate in region so that being closed based on topology
System realizes the mapping one by one between each location point in each location point in source model and object module, and then improves
To the accuracy of the registration of object module.
In one embodiment of the application, the step S11 identifies the characteristic point and its coordinate of preset quantity on source model;Base
Region divisions are carried out in all characteristic points of preset regions partitioning and the characteristic point and its coordinate pair, and respectively to obtaining
Each source divide at least one location point in region and carry out topological coding, obtain each source and divide at least one in region
The topological relation of location point, including:
Preset source model is obtained, the characteristic point of identification and the preset preset quantity in the source model, and obtain
The position of each characteristic point;Here, preset source model can be faceform, human body limb model, human organ mould
Type and joint model etc., certainly, what other will be likely to occur from now on is such as applicable to the application for source model, then should be included in this
It in application, is illustrated by taking faceform as an example below, as shown in Figure 3.
For the ease of carrying out accuracy registration to the object module that is registrated of actual needs, it is preset first for on time into
The source model of row reference chooses the location point of key position by Feature point recognition algorithm from all location points in source model
As in the faceform in characteristic point, such as Fig. 3, by the key position in all location points in faceform:Such as
Eyebrows, nose, the characteristic point at the bridge of the nose, at point and labial angle etc. as faceform, and the preset characteristic point to be chosen
Quantity is preset quantity, which can be 68 characteristic points taken when human face characteristic point identification (for example, characteristic point
A1, characteristic point A2, characteristic point A3 ..., characteristic point A68), be also possible to the number of other any amount of characteristic points, with protect
Demonstrate,prove the characteristic point that identifies can accurately reflect the general profile of faceform and can be adapted to actual acquisition other
Any faceform.
Then the step S11 is based on the preset regions partitioning and the characteristic point and its coordinate, to described default
The characteristic point of quantity carries out region division, obtains each source in the source model and divides the corresponding characteristic point in region and its seat
Mark;In one preferred embodiment of the application, uses preset regions partitioning for Deaunay triangle division method as shown in Figure 3, pass through
Deaunay triangle division method simultaneously combines the characteristic point that identifies and its coordinate in faceform in Fig. 3, to the preset quantity
The characteristic point of (for example, 68) carries out region division, obtains each source in faceform and divides the corresponding characteristic point in region such as
Shown in Fig. 3, the source division region of each triangle in faceform is corresponding, and there are three characteristic points, respectively Pm1、Pm2、Pm3,
Middle m is that source divides zone number belonging to region, such as m is that m is 8 etc. when the 1, the 8th source divides region when the division region of the first source,
It realizes and region division is carried out to the characteristic point of the preset quantity in source model.
Then each source in source model divides in region the step S11 respectively, is based on the corresponding characteristic point
And its corresponding source of coordinate pair divides at least one location point in region and carries out topological coding, obtains each source dividing regions
The topological relation of at least one location point in domain.For example, the source that the number in faceform in Fig. 3 is m divides region
It is interior, the characteristic point P in region is divided based on the sourcem1、Pm2And Pm3And corresponding coordinate (Xm1, Ym1, Zm1)、(Xm2, Ym2,
Zm2) and (Xm3, Ym3, Zm3), each location point in the m of region is divided to the source and carries out topological coding, divides region to obtain the source
The topological relation of each location point P in m:P=aPm1+bPm2+cPm3, wherein the source divides each location point P in the m of region
Topological relation can be expressed as:(a, b, c), features described above can be passed through by dividing each location point in the m of region in the source
Point Pm1、Pm2And Pm3And corresponding coordinate (Xm1, Ym1, Zm1)、(Xm2, Ym2, Zm2) and (Xm3, Ym3, Zm3) linear combination come
The expression for carrying out topological relation since the topological relation is not influenced by the geometric deformation for dividing region, therefore can be opened up by this
Flutter relationship guarantee source model in each location point, can be calculated on actual object module map one by one therewith it is corresponding
Location point.
In the present embodiment, each source in source model respectively in the step S11 is divided in region, based on described right
The corresponding source of the characteristic point and its coordinate pair answered divides at least one location point in region and carries out topological coding, obtains every
A source divides the topological relation of at least one location point in region, specifically includes:
Obtain topological coding method corresponding with the preset regions partitioning;Here, the topological coding method is according to described pre-
It sets region division method and comes what corresponding selection determined, for example, if the preset regions partitioning is Deaunay triangle division method, with
The corresponding topological coding method of Deaunay triangle division method is barycentric coodinates, so that dividing appointing in region in the source of triangle
The sum of its component of the barycentric coodinates of one location point is 1.After being recompiled by the barycentric coodinates, two correspondences can be made
Delta-shaped region in all the points, also can all have the identical corresponding points of barycentric coodinates, thus can be two corresponding three
All the points in angular domain establish one-to-one mapping relations, as long as there is the one-to-one mapping relations, so that it may
Any desired Feature Mapping is done for one of delta-shaped region can be based on, is not being influenced by geometric deformation with reaching
Under, the mapping of each point in corresponding delta-shaped region can also be realized, so that one-to-one correspondence maps out and delta-shaped region
To another delta-shaped region of reply.In another example if the preset regions partitioning is bezier surface partitioning, with the shellfish
The corresponding topological coding method of Sai Er curved surface partitioning is parametric surface compiling method, so that the source of the curved form in source model divides
The target that each location point in region is capable of corresponding curved form in object module divides region memory and is corresponding
Location point.
Then each source in the source model divides in region the step S11 respectively, based on the topology coding
Method and the corresponding characteristic point and its coordinate divide at least one location point in region to the corresponding source and carry out topology
Coding obtains the topological relation at least one location point that each source divides in region, realizes and is based on dividing with preset regions
The topology that the corresponding topological coding method of method divides all location points in region to each source in source model encodes, with obtain can
Corresponding actual target is mapped to one by one with all location points divided each source in source model in region divides region
Interior topological relation.
In one embodiment of the application, the step S13 in the object module from identifying and the spy in the source model
The target feature point and its coordinate of corresponding, the described preset quantity of sign point, and based on the preset regions partitioning to all institutes
State target feature point and carry out region division, obtain each target in object module divide the corresponding target feature point in region and its
Coordinate specifically includes:
From the target for identifying preset quantity corresponding with the characteristic point in the source model, described in the object module
Characteristic point, and obtain from the object module position of the target feature point;As shown in figure 4, being identified according to source model special
Levy point Feature point recognition algorithm, identified from the target faceform of the actual scanning in Fig. 4 of acquisition in source model
One-to-one, the described preset quantity of each characteristic point target feature point, for example, the source model in Fig. 3:Face mould
There is a characteristic point A1 at nose in type, then there are one and spy at the nose in the actual target faceform in Fig. 4
The corresponding target feature point B1 of point A1 is levied, in another example, the source model in Fig. 3:There is a feature at chin in faceform
Point A2, then there are a target feature points corresponding with characteristic point A2 at the chin in the actual target faceform in Fig. 4
B2 so analogizes, Fig. 3 source model:Each characteristic point in faceform is equal in the actual target faceform in such as Fig. 4
There are a corresponding target feature points, so that source model:In characteristic point and actual target faceform in faceform
Target feature point correspond.
Then the step S13 is based on the predeterminable area partitioning and the target feature point and its coordinate, to described
The target feature point of preset quantity carries out region division, obtains each target in the object module and divides the corresponding mesh in region
Mark characteristic point and its coordinate, wherein each source in the source model divides region and has one in the object module
Corresponding target divides region.For example, referring in step S11 to source model:The characteristic point of preset quantity in faceform
The predeterminable area partitioning for carrying out region division, to the corresponding preset quantity of actual target faceform in Fig. 4
Target feature point carries out region division, can obtain each source showing such as the target face model in Fig. 4 and in Fig. 3 and draw
The corresponding target in subregion divides region and each target divides target feature point and its coordinate corresponding to region, with realization pair
The identification of the target feature point of the target faceform of actual scanning.Such as:If the source of each triangle in the faceform in Fig. 3
Division region is corresponding, and there are three characteristic points, respectively Pm1、Pm2、Pm3, wherein m is that source divides zone number belonging to region, then
The source of each triangle in actual target faceform divides region and respectively corresponds that there are three target feature points:Pn1、Pn2、
Pn3, wherein n is that target divides zone number belonging to region, here, the characteristic point P in faceformm1In actual target person
Target feature point P is corresponded in face modeln1, characteristic point P in faceformm2It is corresponded in actual target faceform
Target feature point Pn2, characteristic point P in faceformm3Target feature point P is corresponded in actual target faceformn3。
In one embodiment of the application, the step S14 be based on the topological relation and the corresponding target feature point and
Its coordinate is registrated out at least one in each target division region in object module corresponding with the source model, described
Location point and its coordinate, including:
Each target in the object module divides in region,
It is based respectively on the topological relation and the corresponding target feature point and its coordinate, the target is registrated out and divides
At least one location point and its coordinate in region,
Wherein, it is corresponding in the presence of one in target division region to divide each location point in region for the source
Location point.
For example, if source in source model divides in the m of region there are W location point, respectively P1, P2, P3 ..., PW,
And the topological relation of each location point is respectively (aP1, bP1, cP1), (aP2, bP2, cP2), (aP3, bP3, cP3) ... ... and (aPW,
bPW, cPW), if dividing region m corresponding target division region with source is that target divides region n, and the mesh in object module
It is respectively P that mark, which divides the corresponding target feature point in region,n1、Pn2、Pn3, then the target feature point in the n of region is divided according to the target
And its coordinate, and the topological relation that the target divides region n and source divides each location point in the m of region is combined, it can be one by one
It maps and calculates each location point P ' and its coordinate in target division region n, then each of in target division region n
Location point P '=aPn1+bPn2+cPn3, for example, corresponding in object module to calculate the location point P1 divided in the m of region with source
Location point P1 ', then pass through P1 '=(aP1)Pn1+(bP1)Pn2+(cP1)Pn3It is calculated, is divided in the m of region to calculate with source
The corresponding location point P2 ' in object module of location point P2, then pass through P2 '=(aP2)Pn1+(bP2)Pn2+(cP2)Pn3It calculates
It arrives ... ..., and so on, each target in object module divides in region, can be drawn by source model and target
The corresponding former topological relation for dividing each location point in region in subregion, and combining target divides the corresponding target spy in region
Sign point and its coordinate, can map one by one and be registrated out each target in object module divide region in all location points and
Its coordinate as shown in figure 5, realization each location point in source model can be mapped correspondingly in object module
A location point out, to achieve the purpose that all location points in source model and object module establishing one-to-one relationship.
According in an embodiment of the another aspect of the application, a kind of method for registering based on source model is additionally provided,
In, the method includes:Source model and object module are obtained respectively, such as source model is faceform shown in Fig. 3, target mould
Type is the target faceform that actual scanning shown in Fig. 4 obtains;For the target face in the faceform and Fig. 4 in Fig. 3
Model, using identical Feature point recognition algorithm, in source model as shown in Figure 3:Preset quantity is identified on faceform
Characteristic point and its position, meanwhile, identified on target faceform as shown in Figure 4, with the source model on each spy
The one-to-one target feature point of point is levied, and obtains the position of the target feature point, so that every in the faceform in Fig. 3
There is a corresponding target feature point in target faceform in Fig. 4 in a characteristic point;
Then, for the target faceform in the faceform and Fig. 4 in Fig. 3, identical preset regions is based on and are drawn
Point-score while carrying out region division to all characteristic points, carries out region division to all target feature points, respectively
It obtains each source in the source model and divides the corresponding characteristic point in region and its position, each target in the object module
Divide the corresponding target feature point in region and its position, wherein each source in the source model divides region can be in target
A corresponding target is found in model and divides region, so that the source in source model divides the dividing regions in region and object module
Domain corresponds, convenient for subsequent based on the corresponding mapping one by one for dividing the location point in the progress region of region and registration;
Later, at least one location point in region is divided to each source respectively and carries out topological coding, obtained each source and draw
The topological relation of at least one location point in subregion, so as to it is subsequent based on the topological relation can by in the source model
All location points that each source divides in region, which map one by one and are registrated the corresponding target into object module to be registered, to be drawn
In subregion, it can not only guarantee one by the topological relation that each source behind division region divides each location point in region
The accuracy of one registration also makes the registration of the location point carried out based on the topological relation not by each target in object module
The influence for dividing the geometry deformation in region, further increases the accuracy of registration.
Finally, being registrated out and the source mould based on the topological relation and the corresponding target feature point and its coordinate
Type is corresponding, each target in the object module divides at least one location point and its coordinate in region, so that described
Source, which divides each location point in region and divides in the target, has a corresponding location point in region, realize source mould
Each location point in type can map out correspondingly a location point in object module, to reach source model and
All location points in object module establish the purpose of one-to-one relationship.
According to the another aspect of the application, a kind of non-volatile memory medium is additionally provided, being stored thereon with computer can
Reading instruction when the computer-readable instruction can be executed by processor, realizes the processor such as based on the model of characteristic point
Method for registering.
According to the another aspect of the application, a kind of Model registration equipment based on characteristic point is additionally provided, wherein described to set
It is standby to include:
One or more processors;
Non-volatile memory medium, for storing one or more computer-readable instructions,
When one or more of computer-readable instructions are executed by one or more of processors, so that one
Or multiple processors realize such as Model registration method based on characteristic point.
Here, the detailed content of each embodiment in the Model registration equipment based on characteristic point, for details, reference can be made to this
The corresponding part of the embodiment of the method for Model registration equipment end based on characteristic point, here, repeating no more.
In conclusion characteristic point and its coordinate of the application by the identification preset quantity on source model;Based on preset area
Domain partitioning and all characteristic points of the characteristic point and its coordinate pair carry out region divisions, and must after changing point to region respectively
The each source arrived divides at least one location point in region and carries out topological coding, obtains at least one in each source division region
The topological relation of a location point realizes the Feature point recognition to source model, characteristic point region division and each source and divides region
The topology coding of interior all location points, and then the topological relation of each location point is obtained, so as to subsequent corresponding to the source model
, actual object module carries out the registration that the topological relation carries out object module capable of being called directly, thus effectively with punctual
Improve the subsequent registration efficiency to object module in ground;It is punctual when needing to carry out object module to match, obtain target mould to be registered
Type;Referring to the characteristic point identified in source model, from being identified in the object module and the characteristic point pair in the source model
The target feature point and its coordinate of the preset quantity answer, described, and based on the preset regions partitioning in object module
All target feature points carry out region division, obtain each target in object module and divide the corresponding target signature in region
Point and its coordinate;It is divided in region in each target of object module, it is corresponding to divide region based on the topological relation and target
Target feature point and its coordinate, be registrated out each target dividing regions in object module corresponding with the source model, described
At least one location point and its coordinate in domain, so that realizing each location point and the target in source model based on topological relation
The mapping one by one between each location point in model, and then improve the accuracy to the registration of object module.
It should be noted that the application can be carried out in the assembly of software and/or software and hardware, for example, can adopt
With specific integrated circuit (ASIC), general purpose computer or any other realized similar to hardware device.In one embodiment
In, the software program of the application can be executed to implement the above steps or functions by processor.Similarly, the application
Software program (including relevant data structure) can be stored in computer readable recording medium, for example, RAM memory,
Magnetic or optical driver or floppy disc and similar devices.In addition, hardware can be used to realize in some steps or function of the application, example
Such as, as the circuit cooperated with processor thereby executing each step or function.
In addition, a part of the application can be applied to computer program product, such as computer program instructions, when its quilt
When computer executes, by the operation of the computer, it can call or provide according to the present processes and/or technical solution.
And the program instruction of the present processes is called, it is possibly stored in fixed or moveable recording medium, and/or pass through
Broadcast or the data flow in other signal-bearing mediums and transmitted, and/or be stored according to described program instruction operation
In the working storage of computer equipment.Here, including a device according to one embodiment of the application, which includes using
Memory in storage computer program instructions and processor for executing program instructions, wherein when the computer program refers to
When enabling by processor execution, method and/or skill of the device operation based on aforementioned multiple embodiments according to the application are triggered
Art scheme.
It is obvious to a person skilled in the art that the application is not limited to the details of above-mentioned exemplary embodiment, Er Qie
In the case where without departing substantially from spirit herein or essential characteristic, the application can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and scope of the present application is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included in the application.Any reference signs in the claims should not be construed as limiting the involved claims.This
Outside, it is clear that one word of " comprising " does not exclude other units or steps, and odd number is not excluded for plural number.That states in device claim is multiple
Unit or device can also be implemented through software or hardware by a unit or device.The first, the second equal words are used to table
Show title, and does not indicate any particular order.
Claims (10)
1. a kind of Model registration method based on characteristic point, wherein the method includes:
The characteristic point and its coordinate of preset quantity are identified on source model;Based on preset regions partitioning and the characteristic point and its
All characteristic points of coordinate pair carry out region division, and divide at least one position in region to obtained each source respectively
Point carries out topological coding, obtains the topological relation at least one location point that each source divides in region;
Obtain object module to be registered;
From the target signature for identifying preset quantity corresponding with the characteristic point in the source model, described in the object module
Point and its coordinate, and region division is carried out to all target feature points based on the preset regions partitioning, obtain target
Each target in model divides the corresponding target feature point in region and its coordinate;
Based on the topological relation and the corresponding target feature point and its coordinate, be registrated out it is corresponding with the source model,
Each target in the object module divides at least one location point and its coordinate in region.
2. according to the method described in claim 1, wherein, each characteristic point in the source model is equal in the object module
There are a corresponding target feature points.
3. according to the method described in claim 2, wherein, the characteristic point and its seat that preset quantity is identified on source model
Mark;Region division is carried out based on all characteristic points of preset regions partitioning and the characteristic point and its coordinate pair, and respectively
At least one location point in region is divided to obtained each source and carries out topological coding, each source is obtained and divides in region extremely
The topological relation of a few location point, including:
Preset source model is obtained, the characteristic point of identification and the preset preset quantity in the source model, and obtain each
The position of the characteristic point;
Based on the preset regions partitioning and the characteristic point and its coordinate, region is carried out to the characteristic point of the preset quantity
It divides, obtains each source in the source model and divide the corresponding characteristic point in region and its coordinate;
Each source in source model divides in region respectively, corresponding described based on the corresponding characteristic point and its coordinate pair
Source divides at least one location point in region and carries out topological coding, obtains at least one location point in each source division region
Topological relation.
4. being based on according to the method described in claim 3, wherein, each source in source model respectively divides in region
The corresponding characteristic point and its corresponding source of coordinate pair divide at least one location point in region and carry out topological coding,
The topological relation at least one location point that each source divides in region is obtained, including:
Obtain topological coding method corresponding with the preset regions partitioning;
Each source in the source model divides in region respectively, is based on the topological coding method and the corresponding characteristic point
And its coordinate, at least one location point in region is divided to the corresponding source and carries out topological coding, each source is obtained and divides
The topological relation of at least one location point in region.
5. according to the method described in claim 4, wherein, it is described from identified in the object module in the source model
The target feature point and its coordinate of corresponding, the described preset quantity of characteristic point, and based on the preset regions partitioning to all
The target feature point carries out region division, obtain each target in object module divide the corresponding target feature point in region and
Its coordinate, including:
From the target signature for identifying preset quantity corresponding with the characteristic point in the source model, described in the object module
Point, and obtain from the object module position of the target feature point;
Based on the predeterminable area partitioning and the target feature point and its coordinate, to the target feature point of the preset quantity
Region division is carried out, each target in the object module is obtained and divides the corresponding target feature point in region and its coordinate,
Wherein, each source in the source model divides region and divides in the presence of a corresponding target in the object module
Region.
6. described to be based on the topological relation and the corresponding target feature point according to the method described in claim 5, wherein
And its coordinate, at least one be registrated out in each target division region in object module corresponding with the source model, described
A location point and its coordinate, including:
Each target in the object module divides in region,
It is based respectively on the topological relation and the corresponding target feature point and its coordinate, the target is registrated out and divides region
At least one interior location point and its coordinate,
Wherein, the source, which divides each location point in region and divides in the target, has a corresponding position in region
Point.
7. method according to any one of claim 1 to 6, wherein the preset regions partitioning includes following any
?:
De Laode Deaunay triangle division method, tetra- jiaos of partitionings of Deaunay and bezier surface partitioning.
8. a kind of method for registering based on source model, wherein the method includes:
Source model and object module are obtained respectively;
Characteristic point and its position of preset quantity are identified on the source model, meanwhile, identified on the object module,
With the one-to-one target feature point of each characteristic point on the source model, and the position of the target feature point is obtained;
Based on the preset regions partitioning, while carrying out region division to all characteristic points, to all targets
Characteristic point carries out region division, respectively obtains each source in the source model and divides the corresponding characteristic point in region and its position,
Each target in the object module divides the corresponding target feature point in region and its position;
At least one location point in region is divided to each source respectively and carries out topological coding, each source is obtained and divides in region
The topological relation of at least one location point;
Based on the topological relation and the corresponding target feature point and its coordinate, be registrated out it is corresponding with the source model,
Each target in the object module divides at least one location point and its coordinate in region.
9. a kind of non-volatile memory medium, is stored thereon with computer-readable instruction, the computer-readable instruction can be located
When managing device execution, the processor is made to realize such as method described in any item of the claim 1 to 8.
10. a kind of Model registration equipment based on characteristic point, wherein the equipment includes:
One or more processors;
Non-volatile memory medium, for storing one or more computer-readable instructions,
When one or more of computer-readable instructions are executed by one or more of processors, so that one or more
A processor realizes such as method described in any item of the claim 1 to 8.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810265349 | 2018-03-28 | ||
CN2018102653496 | 2018-03-28 |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108898627A true CN108898627A (en) | 2018-11-27 |
Family
ID=64350988
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810794495.8A Pending CN108898627A (en) | 2018-03-28 | 2018-07-19 | A kind of Model registration method and apparatus based on characteristic point |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108898627A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109766850A (en) * | 2019-01-15 | 2019-05-17 | 西安电子科技大学 | Fingerprint image matching method based on Fusion Features |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102663810A (en) * | 2012-03-09 | 2012-09-12 | 北京航空航天大学 | Full-automatic modeling approach of three dimensional faces based on phase deviation scanning |
CN104851123A (en) * | 2014-02-13 | 2015-08-19 | 北京师范大学 | Three-dimensional human face change simulation method |
CN105427385A (en) * | 2015-12-07 | 2016-03-23 | 华中科技大学 | High-fidelity face three-dimensional reconstruction method based on multilevel deformation model |
US20160154994A1 (en) * | 2014-12-02 | 2016-06-02 | Samsung Electronics Co., Ltd. | Method and apparatus for registering face, and method and apparatus for recognizing face |
-
2018
- 2018-07-19 CN CN201810794495.8A patent/CN108898627A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102663810A (en) * | 2012-03-09 | 2012-09-12 | 北京航空航天大学 | Full-automatic modeling approach of three dimensional faces based on phase deviation scanning |
CN104851123A (en) * | 2014-02-13 | 2015-08-19 | 北京师范大学 | Three-dimensional human face change simulation method |
US20160154994A1 (en) * | 2014-12-02 | 2016-06-02 | Samsung Electronics Co., Ltd. | Method and apparatus for registering face, and method and apparatus for recognizing face |
CN105427385A (en) * | 2015-12-07 | 2016-03-23 | 华中科技大学 | High-fidelity face three-dimensional reconstruction method based on multilevel deformation model |
Non-Patent Citations (5)
Title |
---|
全国农业资源区划办公室等: "《中国农业资源与区划》", 31 December 1997, 中国农业科技出版社 * |
姜太平等: "基于 Candide 模型的人脸深度信息生成技术研究", 《计算机技术与发展》 * |
王东等: "基于点匹配的3D人脸网格自动转换方法", 《计算机工程与设计》 * |
王成焘等: "《骨科植入物工程学》", 31 October 2016, 上海交通大学出版社 * |
郭笑寒等: "基于Candide3的三维人脸快速建模研究", 《数字技术与应用》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109766850A (en) * | 2019-01-15 | 2019-05-17 | 西安电子科技大学 | Fingerprint image matching method based on Fusion Features |
CN109766850B (en) * | 2019-01-15 | 2021-06-01 | 西安电子科技大学 | Fingerprint image matching method based on feature fusion |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Hamann | A data reduction scheme for triangulated surfaces | |
US10872439B2 (en) | Method and device for verification | |
US7760201B2 (en) | Transfer of attributes between geometric surfaces of arbitrary topologies with distortion reduction and discontinuity preservation | |
WO2017186016A1 (en) | Method and device for image warping processing and computer storage medium | |
CN109242903A (en) | Generation method, device, equipment and the storage medium of three-dimensional data | |
WO2021142996A1 (en) | Point cloud denoising method, system, and device employing image segmentation, and storage medium | |
WO2014123619A1 (en) | System and method for identifying similarities in different images | |
JP2015521767A (en) | Method for depicting the periphery of a vehicle | |
CN107507232B (en) | Stereo matching method based on multi-scale iteration | |
US20200380250A1 (en) | Image processing method and apparatus, and computer storage medium | |
WO2019169699A1 (en) | House model rendering method and apparatus, terminal device, and medium | |
CN112583620A (en) | Network topology map generation method and device, electronic equipment and storage medium | |
CN114820972B (en) | Contour line and/or contour surface generation method, system, equipment and storage medium | |
CN109726755A (en) | A kind of picture mask method, device and electronic equipment | |
ES2906626T3 (en) | Image processing device and two-dimensional image generation program | |
CN111932438A (en) | Image style migration method, equipment and storage device | |
CN108898627A (en) | A kind of Model registration method and apparatus based on characteristic point | |
CN113643446B (en) | Automatic marking method and device for mandibular neural tube and electronic equipment | |
CN111130569B (en) | Spatial information data self-adaptive fault-tolerant processing method and system | |
US20210158590A1 (en) | Facial animation retargeting using an anatomical local model | |
CN110427944A (en) | Acquisition methods, device, equipment and the storage medium of car plate detection data | |
CN111145204B (en) | Polygonal simplifying method for profile curve with settable edge number | |
US10395408B1 (en) | Systems and methods for rendering vector shapes | |
WO2020094023A1 (en) | Road thinning algorithm | |
CN113051406A (en) | Character attribute prediction method, device, server and readable storage medium |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: Room J1839, 1st floor, 1185 Huyi Highway, Jiading District, Shanghai 201800 Applicant after: Research Territory Information Technology (Shanghai) Co., Ltd. Address before: Room J1839, 1st floor, 1185 Huyi Highway, Jiading District, Shanghai 201802 Applicant before: Jingjing information technology (Shanghai) Co., Ltd. |
|
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181127 |