CN100416611C - Network topology model construction based three-dimensional human face cartoon making method - Google Patents
Network topology model construction based three-dimensional human face cartoon making method Download PDFInfo
- Publication number
- CN100416611C CN100416611C CNB2006100533891A CN200610053389A CN100416611C CN 100416611 C CN100416611 C CN 100416611C CN B2006100533891 A CNB2006100533891 A CN B2006100533891A CN 200610053389 A CN200610053389 A CN 200610053389A CN 100416611 C CN100416611 C CN 100416611C
- Authority
- CN
- China
- Prior art keywords
- dimensional
- face
- curved surface
- human face
- dimension
- 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.)
- Expired - Fee Related
Links
- 238000010276 construction Methods 0.000 title claims abstract description 17
- 238000000034 method Methods 0.000 title claims description 42
- 238000006073 displacement reaction Methods 0.000 claims abstract description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000003287 optical effect Effects 0.000 claims description 6
- 239000003550 marker Substances 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract 1
- 239000011159 matrix material Substances 0.000 description 5
- 210000003484 anatomy Anatomy 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013481 data capture Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000001815 facial effect Effects 0.000 description 1
- 230000008921 facial expression Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 230000004118 muscle contraction Effects 0.000 description 1
- 210000003491 skin Anatomy 0.000 description 1
- 210000003625 skull Anatomy 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Landscapes
- Processing Or Creating Images (AREA)
Abstract
The invention relates to a method for producing three-dimension face cartoon, based on network topology model construction, wherein it comprises: first obtaining rare three-dimension motion catch data and three-dimension face curvature grid; then marking relative control point of motion catch data on the three-dimension face grid; using short distance as the shortest distance on the three-dimension curvature to build the models of topology structure and seam of three-dimension curvature; in the deformation of free curvature, using short distance to calculate the distance between control point and other points; using the displacement of control point as parameter and the radial basic function algorism matched with short distance to deform the three-dimension curvature, to generate the three-dimension face cartoon. The invention can obtain high-accuracy three-dimension face cartoon, effectively separate the seam and continuous curvature of face, and improve the producing efficiency.
Description
Technical field
The present invention relates to the computer three-dimensional animation technical field, relate in particular to a kind of three-dimensional human face cartoon making method based on network topology model construction.
Background technology
The work the earliest of 3 D human face animation technology starts from 1972.Thereafter scholar has done a lot of work, in the hope of generating lively 3 D human face animation true to nature.But because the anatomical structure of people's face is very complicated, trickle non-rigid motion is difficult to use mathematical modeling, and simultaneously, people make this problem very difficult to the susceptibility of people's face outward appearance.The work of sense of reality 3 D human face animation has at present obtained some progress, can be divided into following a few class basically:
Method based on interpolation: human face animation the earliest is based on the method for interpolation.At first make some specific human face expression key frames, between these key frames, set up smooth transition with linear interpolation then by the animation teacher.This method mainly relies on animation teacher's profile, and more consuming time.
Method based on Free Transform: the method for Free Transform is converted into 3 D human face animation the problem of a surf deform.Formed an enclosure body by some reference mark, the motion of enclosure body directly drives the distortion of people's face grid.This method import-restriction is few, but often needs people's face is carried out area dividing, and this process is relatively more difficult and loaded down with trivial details concerning the general user; And because deformable body is not considered the topological structure of people's face grid itself, so distortion tends to produce some flaws.
Method based on physical modeling: at first to the anatomical structure of people's face, carry out modeling as muscle, skull and skin, motion capture data then can be converted into the parameter of contraction of muscle, the human face animation of driving model then.The kinetic characteristic that meets people's face that this method produces, relatively truer, but calculation cost is than higher, and be difficult to fast new role be set up physical model.
Method based on example: 3 D human face animation is counted as the linear combination of some particular instances.Motion vector in the motion capture data directly is used to estimate the parameter and the interpolation coefficient of linear combination.Predefined example then is mixed together by these parameters and coefficient.These class methods are more stable and accurate, but the process of setting up example is relatively more expensive and loaded down with trivial details.
Expression clone's method: for existing 3D grid exercise data is applied on the new three-dimensional face mould, moving displacement need be rotated at the local geometric features of grid and convergent-divergent.The motion that expression clone can solve between the different three-dimensional face surface meshes is redirected problem.
The real-time face animation is caught: by computer vision algorithms make, the 3D shape of the role in the video and color texture are by real-time reconstruction.This method can be carried out lively animation creation true to nature to specific role, but the animation that produces is difficult to transfer on other roles.
Summary of the invention
The object of the present invention is to provide a kind of method of the 3 D face animation based on network topology model construction.The method can be carried out effective modeling to the topological structure of people's face, guarantees the sense of reality of human face animation, has avoided simultaneously people's face gap area is carried out manual divide loaded down with trivial details, has good real-time performance and extensibility.
For achieving the above object, the present invention adopts following technical scheme:
Three-dimensional human face cartoon making method based on network topology model construction: at first obtain sparse three-dimensional motion and catch data and three-dimensional face surface mesh; On the three-dimensional face grid, demarcate the corresponding reference mark of motion capture data then; Adopt short range from, promptly the shortest path on the three-dimension curved surface carries out modeling to topological structure on the three-dimension curved surface and slit; In the process of deformation of free, with the distance of short range between calculation control point and other points; By the displacement at reference mark as parameter, by mated short range from the radial basis function algorithm three-dimension curved surface is out of shape, generate realistic 3 D human face animation.
Describedly obtain sparse three-dimensional motion and catch data: by the fast optical video camera, people's face motion that will be tied with reflective marker is gathered, and uses the mode record of file then.
Obtain three-dimensional face surface mesh data: adopt the optical three-dimensional scanning instrument,, and be plotted on the display with the three-dimensional drawing software package with the three-dimensional information and the color texture information record of specific people's face.
On the three-dimensional face grid, demarcate the corresponding reference mark of motion capture data:, set up the corresponding relation of motion capture data and three-dimension curved surface grid with on the three-dimensional face surface mesh, naming of all or part of gauge point correspondence in the motion capture data.
Adopt short range from, be that shortest path on the three-dimension curved surface carries out modeling to topological structure on the three-dimension curved surface and slit: adopt the shortest path first in the graph theory, with the shortest path between 2 o'clock as the distance metric between them, thereby distinguish continuous curve surface and slit on the three-dimension curved surface.
In the process of deformation of free, with short range from the distance between calculation control point and other point: adopt shortest path on the three-dimension curved surface measure the reference mark to curved surface on the driving force of other points, thereby obtain three-dimension curved surface distortion of the mesh correctly on the topology.
The displacement that the reference mark is arranged is as parameter, by mated short range from the radial basis function algorithm three-dimension curved surface is out of shape, generate realistic 3 D human face animation: with the displacement at reference mark and short range in being embedded into the radial basis function algorithm.
Beneficial effect of the present invention:
1, three-dimensional motion is caught data and is provided sequential assurance accurately to 3 D human face animation, and the three-dimensional face grid data then provides space assurance accurately.The combination of data and three-dimensional face grid data is caught in three-dimensional motion, can obtain high accuracy three-dimensional human face animation on the space-time;
2, introduce short range from coming that the three-dimensional face surface mesh is carried out topology model construction, can effectively distinguish slit and continuous curve surface on people's face curved surface, avoid complexity that the people's face functional area in the Free Transform method divides and loaded down with trivial details, improved the make efficiency of human face animation;
3, with the reference mark displacement in the motion capture data as parameter, by mated short range from the radial basis function algorithm people's face curved surface is out of shape, realistic 3 D human face animation be can generate in real time, the simulation of people's face topological structure and the requirement of real-time guaranteed simultaneously.
Description of drawings
The present invention is further illustrated with concrete enforcement below in conjunction with accompanying drawing;
Fig. 1 (a) catches data for three-dimensional motion of the present invention;
Fig. 1 (b) is a three-dimensional face model of the present invention;
Fig. 2 is that short range of the present invention is from the synoptic diagram on the three-dimensional face curved surface.
Embodiment
Based on the three-dimensional human face cartoon making method of network topology model construction, can guarantee the sense of reality of human face animation, avoided simultaneously people's face gap area is carried out manual divide loaded down with trivial details, have good real-time performance and extensibility.At first obtain sparse three-dimensional motion and catch data and three-dimensional face surface mesh; On the three-dimensional face grid, demarcate the corresponding reference mark of motion capture data then: adopt short range from, promptly the shortest path on the three-dimension curved surface carries out modeling to topological structure on the three-dimension curved surface and slit; In the process of deformation of free, with the distance of short range between calculation control point and other points; By the displacement at reference mark as parameter, by mated short range from the radial basis function algorithm three-dimension curved surface is out of shape, generate realistic 3 D human face animation.This method has following steps:
(1) obtaining of data caught in sparse three-dimensional motion, promptly by the fast optical video camera, the people's face motion that is tied with reflective marker gathered, and uses the mode record of file then;
(2) obtaining of three-dimensional face surface mesh data promptly adopted the optical three-dimensional scanning instrument, with the three-dimensional information and the color texture information record of specific people's face, and is plotted on the display with the three-dimensional drawing software package;
(3) on the three-dimensional face grid, demarcate the corresponding reference mark of motion capture data, be about on the three-dimensional face surface mesh, the indicating of all or part of gauge point correspondence in the motion capture data, set up the corresponding relation of motion capture data and three-dimension curved surface grid:
(4) adopt short range from topological structure on the three-dimension curved surface and slit are carried out modeling, promptly adopt the shortest path first in the graph theory, with the shortest path between 2 o'clock as the distance metric between them, thereby distinguish continuous curve surface and slit on the three-dimension curved surface;
(5) in the process of deformation of free, with the distance of short range between calculation control point and other points, promptly adopt the shortest path on the three-dimension curved surface to measure the driving force of reference mark, thereby obtain three-dimension curved surface distortion of the mesh correctly on the topology other points on the curved surface;
(6) displacement that the reference mark is arranged is as parameter, by mated short range from the radial basis function algorithm three-dimension curved surface is out of shape, generate realistic 3 D human face animation, be about to the displacement at reference mark and short range in being embedded into the radial basis function algorithm.
Embodiment:
At first can gather the exercise data of people's face by the three-dimensional motion data capture device, performing artist's facial principal character point makes marks with reflectorized material, shown in Fig. 1 (left side), perhaps reads the three-dimensional motion data of having caught from computer memory device; In addition, set up 3 d human face mesh model, perhaps from computer memory device, read existing 3 d human face mesh model, adopt the three-dimensional software bag that 3 d human face mesh model is plotted on the display by spatial digitizer.Send out the corresponding reference mark that is equipped with demarcation motion capture data on the three-dimensional face grid by the computing machine input, soon all or part of gauge point correspondence in the motion capture data indicates on the three-dimensional face surface mesh, set up the corresponding relation of motion capture data and three-dimension curved surface grid, shown in Fig. 1 (right side).
After having set up the three-dimensional face grid, we with short range from coming topological structure to carry out modeling to people's face.We regard people's face grid of input as a non-directed graph, and wherein the node of non-directed graph and arc are represented by the summit and the limit of people's face grid.According to graph theory, can construct an adjacency matrix by the non-directed graph of people's face cancellated structure.Adjacency matrix by three-dimensional face cancellated structure shown in Figure 2 is made of 5832 row and 5832 row, has only 22984 nonzero elements, that is, nonzero element has only 0.068%.Node on the non-directed graph
With
Between short range from δ
IjCan be by the shortest path first of adjacency matrix, promptly dijkstra's algorithm obtains:
We calculated on the three-dimensional face grid the reference mark to other short ranges of having a few from, and store with the form of matrix, can quicken the deformation process of three-dimensional face grid.
In the generative process of human face animation, radial basis function is as energy-optimised mapping function.Radial basis function has various ways, and we have adopted many secondaries radial basis function (Multi-quadrics RBF), and its form is as follows:
Wherein
Be the vector representation of input data, i.e. the initial position of three-dimensional face grid vertex, W=[w
1, w
2..., w
n]
TBe weighting parameter, wherein n is the number at reference mark, γ
JiBe the distance between input vertex j and i the reference mark, σ
iIt is coefficient of relaxation.. in actual moving process, Euclidean distance γ
JiBy short range from δ
IjReplace, and can from the matrix of prior storage, inquire about and obtain.According to the displacement at reference mark, can learn to obtain weighting parameter W, relaxation factor can be got this summit to the Euclidean distance apart from its nearest reference mark.
When the three-dimensional face grid was more sparse, the short range defection became discrete, brought some undesired flaw and distortions.In this case, can adopt mix Euclidean distance and short range from method, i.e. δ
Ij=w * δ
Ij+ (1-w) * γ
Ij, wherein w ∈ [0,1] is an interpolation coefficient.The value of W is defined as 0.3 by experiment.
Present embodiment is by adopting the method based on the 3 D face animation of network topology model construction, catch data with three-dimensional motion and provide sequential assurance accurately to 3 D human face animation, then providing accurately with the three-dimensional face grid data, the space guarantees, the combination of data and three-dimensional face grid data is caught in three-dimensional motion, obtained high accuracy three-dimensional human face animation on the space-time: introduce short range simultaneously from coming that the three-dimensional face surface mesh is carried out topology model construction, effectively distinguish slit and continuous curve surface on people's face curved surface, avoided complexity that the people's face functional area in the Free Transform method divides and loaded down with trivial details, improved the make efficiency of human face animation: with the reference mark displacement in the motion capture data as parameter, by mated short range from the radial basis function algorithm people's face curved surface is out of shape, generate realistic 3 D human face animation in real time, guaranteed the simulation of people's face topological structure and the requirement of real-time simultaneously.
Claims (7)
1. the three-dimensional human face cartoon making method based on network topology model construction is characterized in that: at first obtain sparse three-dimensional motion and catch data and three-dimensional face surface mesh; On the three-dimensional face surface mesh, demarcate the corresponding reference mark of motion capture data then; Adopt short range from, promptly the shortest path on the three-dimension curved surface carries out modeling to topological structure on the three-dimension curved surface and slit; In the process of deformation of free, with the distance of short range between calculation control point and other points; By the displacement at reference mark as parameter, by mated short range from the radial basis function algorithm three-dimension curved surface is out of shape, generate realistic 3 D human face animation.
2. a kind of three-dimensional human face cartoon making method according to claim 1 based on network topology model construction, it is characterized in that, describedly obtain sparse three-dimensional motion and catch data: by the fast optical video camera, people's face motion that will be tied with reflective marker is gathered, and uses the mode record of file then.
3. a kind of three-dimensional human face cartoon making method according to claim 1 based on network topology model construction, it is characterized in that: the described three-dimensional face surface mesh data of obtaining: adopt the optical three-dimensional scanning instrument, with the three-dimensional information and the color texture information record of specific people's face, and be plotted on the display with the three-dimensional drawing software package.
4. a kind of three-dimensional human face cartoon making method according to claim 1 based on network topology model construction, it is characterized in that, the described corresponding reference mark of on the three-dimensional face surface mesh, demarcating motion capture data:, set up the corresponding relation of motion capture data and three-dimension curved surface grid with on the three-dimensional face surface mesh, indicating of all or part of gauge point correspondence in the motion capture data.
5. a kind of three-dimensional human face cartoon making method according to claim 1 based on network topology model construction, it is characterized in that: described employing short range from, be that shortest path on the three-dimension curved surface carries out modeling to topological structure on the three-dimension curved surface and slit: adopt the shortest path first in the graph theory, with the shortest path between 2 o'clock as the distance metric between them, thereby distinguish continuous curve surface and slit on the three-dimension curved surface.
6. a kind of three-dimensional human face cartoon making method according to claim 1 based on network topology model construction, it is characterized in that, described in the process of deformation of free, with short range from the distance between calculation control point and other point: adopt shortest path on the three-dimension curved surface measure the reference mark to curved surface on the driving force of other points, go up correct three-dimension curved surface distortion of the mesh thereby obtain topology.
7. a kind of three-dimensional human face cartoon making method according to claim 1 based on network topology model construction, it is characterized in that, described displacement by the reference mark is as parameter, by mated short range from the radial basis function algorithm three-dimension curved surface is out of shape, generate realistic 3 D human face animation: with the displacement at reference mark and short range in being embedded into the radial basis function algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100533891A CN100416611C (en) | 2006-09-14 | 2006-09-14 | Network topology model construction based three-dimensional human face cartoon making method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100533891A CN100416611C (en) | 2006-09-14 | 2006-09-14 | Network topology model construction based three-dimensional human face cartoon making method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1920885A CN1920885A (en) | 2007-02-28 |
| CN100416611C true CN100416611C (en) | 2008-09-03 |
Family
ID=37778603
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2006100533891A Expired - Fee Related CN100416611C (en) | 2006-09-14 | 2006-09-14 | Network topology model construction based three-dimensional human face cartoon making method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100416611C (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100562895C (en) * | 2008-01-14 | 2009-11-25 | 浙江大学 | A kind of method of the 3 D face animation based on Region Segmentation and speced learning |
| CN101241601B (en) * | 2008-02-19 | 2010-06-02 | 深圳先进技术研究院 | Graphic processing joint center parameter estimation method |
| CN103325086B (en) * | 2013-07-10 | 2016-03-09 | 中国科学院苏州纳米技术与纳米仿生研究所 | A kind of deformation method of the three-dimensional picture based on tetrahedral coordinate system |
| CN104077798B (en) * | 2014-07-01 | 2017-05-03 | 中国科学技术大学 | High-reality-sense animation synthesis method for deformable object |
| CN105631920B (en) * | 2014-10-31 | 2018-07-06 | 北京临近空间飞行器***工程研究所 | A kind of sample compressing method of radial basis function supporting point |
| CN116843804A (en) * | 2018-12-29 | 2023-10-03 | 华为技术有限公司 | Method for generating animation expression and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6208357B1 (en) * | 1998-04-14 | 2001-03-27 | Avid Technology, Inc. | Method and apparatus for creating and animating characters having associated behavior |
| CN1404016A (en) * | 2002-10-18 | 2003-03-19 | 清华大学 | Establishing method of human face 3D model by fusing multiple-visual angle and multiple-thread 2D information |
| US6552729B1 (en) * | 1999-01-08 | 2003-04-22 | California Institute Of Technology | Automatic generation of animation of synthetic characters |
-
2006
- 2006-09-14 CN CNB2006100533891A patent/CN100416611C/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6208357B1 (en) * | 1998-04-14 | 2001-03-27 | Avid Technology, Inc. | Method and apparatus for creating and animating characters having associated behavior |
| US6552729B1 (en) * | 1999-01-08 | 2003-04-22 | California Institute Of Technology | Automatic generation of animation of synthetic characters |
| CN1404016A (en) * | 2002-10-18 | 2003-03-19 | 清华大学 | Establishing method of human face 3D model by fusing multiple-visual angle and multiple-thread 2D information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1920885A (en) | 2007-02-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100562895C (en) | A kind of method of the 3 D face animation based on Region Segmentation and speced learning | |
| Wang et al. | Feature based 3D garment design through 2D sketches | |
| CN104376594B (en) | Three-dimensional face modeling method and device | |
| US7515155B2 (en) | Statistical dynamic modeling method and apparatus | |
| US8902232B2 (en) | Facial performance synthesis using deformation driven polynomial displacement maps | |
| CN101303772A (en) | Method for modeling non-linear three-dimensional human face based on single sheet image | |
| US7307633B2 (en) | Statistical dynamic collisions method and apparatus utilizing skin collision points to create a skin collision response | |
| CN100416611C (en) | Network topology model construction based three-dimensional human face cartoon making method | |
| EP2043049A2 (en) | Facial animation using motion capture data | |
| CN106023288A (en) | Image-based dynamic substitute construction method | |
| CN103035022B (en) | Facial expression synthetic method based on feature points | |
| CN103208133A (en) | Method for adjusting face plumpness in image | |
| CN102945561B (en) | Based on the motion synthesis of motion capture data and edit methods in a kind of computing machine skeleton cartoon | |
| Miranda et al. | Sketch express: A sketching interface for facial animation | |
| Fratarcangeli | Position‐based facial animation synthesis | |
| CN103914873B (en) | A kind of sculpture threedimensional model driven according to human action generates method | |
| US20240135620A1 (en) | Morph target animation | |
| Orvalho et al. | Transferring the rig and animations from a character to different face models | |
| Kus et al. | A comparative study of 3D scanning in engineering, product and transport design and fashion design education | |
| CN115908664B (en) | Animation generation method and device for man-machine interaction, computer equipment and storage medium | |
| CN102855652B (en) | Method for redirecting and cartooning face expression on basis of radial basis function for geodesic distance | |
| Çetinaslan | Position manipulation techniques for facial animation | |
| JP2739447B2 (en) | 3D image generator capable of expressing wrinkles | |
| CN115023742A (en) | Facial mesh deformation with detailed wrinkles | |
| Li et al. | Animating cartoon faces by multi‐view drawings |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080903 Termination date: 20120914 |