CN1889129A - Fast human face model building method and system based on single-sheet photo - Google Patents
Fast human face model building method and system based on single-sheet photo Download PDFInfo
- Publication number
- CN1889129A CN1889129A CNA2006100888579A CN200610088857A CN1889129A CN 1889129 A CN1889129 A CN 1889129A CN A2006100888579 A CNA2006100888579 A CN A2006100888579A CN 200610088857 A CN200610088857 A CN 200610088857A CN 1889129 A CN1889129 A CN 1889129A
- Authority
- CN
- China
- Prior art keywords
- model
- human face
- reference mark
- profile
- dimensional
- 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
- 210000000056 organ Anatomy 0.000 claims abstract description 38
- 230000001815 facial effect Effects 0.000 claims description 7
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 2
- 210000003462 vein Anatomy 0.000 abstract 1
- 239000000284 extract Substances 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 5
- 230000003993 interaction Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 210000004709 eyebrow Anatomy 0.000 description 1
- 239000012467 final product Substances 0.000 description 1
- 210000003128 head Anatomy 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000010561 standard procedure Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
Images
Landscapes
- Processing Or Creating Images (AREA)
Abstract
The invention discloses a fast modeling method and system of human face which based on sheet photo. First, importing a photo of facade human face and a universal model of three-dimensional human face, then in allusion to the photo of facade human face, fixing the outline position of the true human face organ, and ascertaining the outline control point; in allusion to universal model of three-dimensional human face, demarcating the model control point according to the said outline control point, establishing the connection between the said model control point and outline control point corresponding to the said, according to the connection between the said model control point and outline control point, fitting the model by making use of universal model of three-dimensional human face. Adjusting the model control point according to the location of the outline control point, to make the entire location of organ similar between three-dimensional human face and human face; the fitted model adds vein. This invention can build the human face model rapidly in front human face; the operation is easy, the achievable human face model have very strong sense of reality within certain angle range.
Description
Technical field
The present invention relates to the facial image process field, relate in particular to the method and system of people's face modeling.
Background technology
The method of present people's face modeling based on still picture can be divided into two classes: a class is traditional shape method for reconstructing based on many sheet photos, and an other class is based on the method for the model fitting of quadrature picture.
Traditional shape method for reconstructing based on many sheet photos generally comprises following steps: the triangulation network, texture are rebuild, made up to camera calibration, Feature Points Matching, three-dimensional point.The theoretical foundation comparative maturity of traditional method for reconstructing, but implement the comparison difficulty.
The purpose of camera calibration is the elements of interior orientation (focal length, principal point, distortion parameter etc.) and the elements of exterior orientation (locus of camera and attitude parameter) of computing camera, and demarcating generally is to utilize special utility appliance to finish.Though the self-calibrating method of Ti Chuing can need the number of characteristics corresponding point without any need for utility appliance in recent years.And Feature Points Matching (promptly seeking the feature corresponding point) is a very problem of difficulty, does not also have very effective method so far.After demarcating and Feature Points Matching finishes,, generally understand specimen beam method (Bundle Adjustment) all unknown numbers are carried out global optimization in order to obtain accurate reconstructed results.
An other class method for reconstructing has been avoided demarcation and the Feature Points Matching in the classic method.Its thought is that people's face is taken front and side image, because direct picture and side image quadrature, so can from image, directly extract the three-dimensional coordinate of human face characteristic point.Come a general three-dimensional face model is carried out match with the three-dimensional coordinate that extracts then, just can obtain the model of human face similarity in a profile and the image.Method principle based on orthogonal image is very simple, but generally needs manual interaction to determine the position of human face characteristic point, and needs the user to shoot orthogonal image, and this operates cumbersome in practice.
Summary of the invention
Technical matters to be solved by this invention is to provide a kind of fast human face model building and system based on single-sheet photo, to set up the faceform simply, rapidly.
For solving the problems of the technologies described above, the invention provides a kind of fast human face model building based on single-sheet photo, comprise the steps:
(1) a front face photograph of input and a universal three-dimensional human face model;
(2), orient real human face organ contours position, and determine profile reference mark wherein at the front face photograph;
(3), calibrate the model reference mark corresponding with described profile reference mark at universal three-dimensional human face model;
(4) set up getting in touch between described model reference mark and the described corresponding profile reference mark;
(5) according to getting in touch between described model reference mark and the profile reference mark, utilize universal three-dimensional human face model to carry out model fitting, come the adjustment model reference mark according to the position at profile reference mark, make the organ of whole three-dimensional face model distribute similar with the organ site in the real human face;
(6) model after the match is added texture.
Wherein, in the described step (2), orient the step of human face outline position in the front face image, comprising:
(21), go out the initial position of human face profile in the front face image based on the ASM model orientation at the front face photograph;
(22) in conjunction with the facial contour actual position in the photograph, the initial position of orienting is finely tuned, obtain the accurate position of human face profile.
Wherein, in the described step (5), in the model fitting process, further, further make the organ of whole three-dimensional face model distribute similar with the organ site in the real human face by the non-reference mark in the method adjustment universal three-dimensional human face model of interpolation.
Wherein, the adjustment at described non-reference mark comprises:
In universal three-dimensional human face model, determine the triangulation network formed by each reference mark, and in the real human face organ contours definite triangulation network of forming by corresponding reference mark;
According to the corresponding triangle in universal three-dimensional human face model and real human face organ contours, set up the affine relation between per two corresponding triangles;
Judge which delta-shaped region the non-reference mark in the universal three-dimensional human face model falls in, utilize described affine relation, determine this adjusted position, non-reference mark according to the corresponding triangle in the real human face organ contours.
Wherein, in the described step (6),, directly the texture coordinate assignment at the reference mark in the human face profile of correspondence is given the reference mark in the model after the match, obtain texture for the reference mark.
Wherein, in the described step (6),,, go out the texture coordinate at adjusted non-reference mark by the texture interpolation calculation according to the texture coordinate of triangular apex corresponding in the real human face organ contours for non-reference mark.
For solving the problems of the technologies described above, the present invention and then a kind of fast face modeling based on single-sheet photo is provided comprises:
True locations of contours device is used for individual front face photograph at input, orients real human face organ contours position, and determines profile reference mark wherein;
The model calibration device is used for the universal three-dimensional human face model at input, calibrates the model reference mark corresponding with described profile reference mark, and sets up getting in touch between described model reference mark and the corresponding profile reference mark;
The model fitting device, be used for according to getting in touch between described model reference mark and the profile reference mark, utilize universal three-dimensional human face model to carry out model fitting, come the adjustment model reference mark according to the position at profile reference mark, make the organ of whole three-dimensional face model distribute similar with the organ site in the real human face;
The texture adding set is used for the model after the match is added texture.
Wherein, described true locations of contours device, at the front face photograph, go out the initial position of human face profile in the front face image based on the ASM model orientation, and in conjunction with the facial contour actual position in the photograph, the initial position of orienting is finely tuned, obtain the accurate position of human face profile.
Wherein, described model fitting device in the model fitting process, further carries out the interpolation adjustment to the non-reference mark in the universal three-dimensional human face model, further makes the organ of whole three-dimensional face model distribute similar with the organ site in the real human face.
Wherein, described texture adds module, and the texture coordinate corresponding by computation model point adds texture.
The present invention can extract the initial profile of human face automatically from individual direct picture, the user only need carry out a spot ofly just can accurate in locating going out the human face profile alternately, can set up the triangle gridding that covers human face region according to the profile that extracts, by setting up the contact of corresponding point in human face profile reference mark and the general three-dimensional model, and carry out the coordinate built at non-reference mark in the three-dimensional model according to triangle gridding, finish the match of universal model to real human face.By setting up the contact of pixel corresponding in three-dimensional model point and the direct picture, finish texture, obtain realistic faceform.The present invention can set up the faceform fast from front face, simple to operate, and the faceform who obtains has the very strong sense of reality in certain angular range.
Description of drawings
Fig. 1 is according to the described modeling process flow diagram of the embodiment of the invention;
Fig. 2 is a front face organ contours positioning result synoptic diagram;
Fig. 3 is a three-dimensional face editing interface synoptic diagram;
Fig. 4 is based on leg-of-mutton model fitting synoptic diagram;
The people face modeling result synoptic diagram of Fig. 5 for finishing according to the embodiment of the invention.
Embodiment
According to the described fast face modeling of the embodiment of the invention, structurally can comprise true locations of contours device, model calibration device, model fitting device, texture adding set based on single-sheet photo.
Wherein, true locations of contours device is used for individual front face photograph at input, orients real human face organ contours position, and determines profile reference mark wherein; The model calibration device is used for the universal three-dimensional human face model at input, calibrates the model reference mark corresponding with described profile reference mark, and sets up getting in touch between described model reference mark and the corresponding profile reference mark; The model fitting device, be used for according to getting in touch between described model reference mark and the profile reference mark, utilize universal three-dimensional human face model to carry out model fitting, come the adjustment model reference mark according to the position at profile reference mark, make the organ of whole three-dimensional face model distribute similar with the organ site in the real human face; The texture adding set is used for the model after the match is added texture.
In the system of the present invention, described locations of contours device can extract the initial profile of human face based on ASM (Active ShapeModel) automatically from individual picture.The ASM model is a kind of expression to people's face shape variation of sample space representative, therefore can be used for controlling people's face shape variation of different attitudes in the human face locations of contours.In fact become the standard method of human face locations of contours based on the human face locations of contours method of ASM.Add the accurate profile that obtains human face after the simple man-machine interaction, its profile reference mark and model calibration device of determining has corresponding relation between the three-dimensional point that calibrates on the universal three-dimensional human face model, based on corresponding relation, this model fitting device can carry out match to the general three-dimensional model automatically according to profile, makes it similar with the real human face model.Come three-dimensional model is added the three-dimensional model that texture just can obtain having the real human face outward appearance according to the image coordinate of the point of extracting by the texture adding set at last.Show that for convenience three-dimensional model can store together with corresponding texture image with the PLY form.
With reference to figure 1, for according to the described modeling process flow diagram of the embodiment of the invention.On the one hand, the embodiment of the invention is at first imported a front face photograph (step 101).At the front face photograph, can orient the profile (step 102) of human face in the front face image earlier based on the ASM model, but be subjected to the influence of factors such as illumination, the result of location can be very inaccurate, therefore can also finely tune (step 103) by man-machine interaction to the profile of location again, obtain the accurate profile of each organ of people's face.The profile that extracts is for model fitting provides foundation, and the result behind the location can be with reference to shown in Figure 2.
In the step 101, the front face photograph can be taken with digital camera and obtain, and also can extract from the video data of USB camera input.Because our general three-dimensional universal model generally is not express one's feelings, so the front face photograph of input should not have expression yet.Last texture card diagram data is to obtain from the front face image of input, so the quality of image will have influence on the display effect of final real human face, therefore should guarantee the front face image clean mark of importing as far as possible.
Well known, the ASM model of mentioning in step 102 can obtain by training, concrete training method has a lot, for example number be: 200610012233.9 in Chinese patent application, name is called as described in the application for a patent for invention of " based on head pose real-time estimating method and the system of AAM ", training ASM model need mark the principal character zone (eyebrow, eyes, nose, face and chin) that each people's face sample carries out people's face, can mark n the profile of putting the principal character zone of representing people's face in a specific embodiment.If the coordinate of the demarcation point in people's face sample is: (x
i, y
i), i=0,1 ..., n, n is the number of point, the planimetric coordinates tissue of being had a few can be called one one n dimensional vector n: S so
j=[x
0, y
0..., x
n, y
n], S
jIt is exactly a sample of training.Planar point in the facial contour of all demarcation all is organized into a n dimensional vector n, then all samples is carried out the PCA conversion and just can obtain the ASM model:
Wherein, s
0Be the average of training sample, mean profile face just, s is exactly the ASM model.
Certainly, the ASM that obtains of other training method also be applicable to of the present invention.
Human face locations of contours based on ASM in the step 102 is a kind of very effective method among the Face Alignment, the specific implementation method is as follows: the first step is to utilize the detection of people's face and eyes, face location technology to orient the eyes of the front face photograph that step 1 provides, the position of face, concrete eyes, face location technology also have much at present, for example Chinese patent application numbers 200610011673.2, and the localization method that provides in " a kind of man face characteristic point positioning method " is provided name; Second step was posting the mean profile face of ASM on the front face photograph on the basis of the position of orienting eyes, face, with the profile of the human face of the mean profile face of the ASM first slightly profile as the front face photograph; The 3rd step was to utilize the edge matching technique further to search for the marginal point of organ according to just omiting the position, and the marginal point that searches was projected to the proper vector s of ASM
iThe subspace of being opened, the result after the projection just can obtain new profile through inverse.Just can access one and the real human face profile similarity while to continuous iterative computation of the 3rd step and be subjected to the model constrained initial profile of ASM again, these profiles are made up of the profile reference mark.But because ASM is a statistical model, the change of shape that therefore can only reflect the human face profile in the certain limit, simultaneously the location on facial image also can be subjected to the influence of illumination etc., so can not point-devicely orient the profile of human face based on the method for ASM.This defective can remedy by step 103.
On the other hand, the embodiment of the invention also will be imported a universal three-dimensional human face model (step 104), and it is demarcated reference mark (step 105).Universal three-dimensional human face model is an average three-dimensional model, can be used as the basis of model fitting.In order to carry out model fitting, also need to set up the contact (step 106) of human face profile and general three-dimensional model corresponding point.Set up the contact of corresponding point for convenience, can design a three-dimensional model editing system, as shown in Figure 3, use this system can finish selection to each point in the model, and can in advance the corresponding three-dimensional model point of each point in the profile all be chosen, and leave in the middle of the corresponding file.
In the step 105, the general three-dimensional model generally is to be made by 3D sculpting software (such as 3DMax, Maya etc.).Do not use complete three-dimensional headform among the present invention, but only used the three-dimensional data of people face part.Its data obtain the depth buffer mechanism of having utilized OpenGL, concrete grammar is: projection mode is set to orthogonal projection under the three-dimensional programmed environment of OpenGL, and then the three-dimensional headform in front being drawn and preserve its depth information is an amplitude deepness image.Choose near the point the human face profile from depth image, with the Z value of its depth value as three-dimensional data, the ranks of image number just can only be comprised the three-dimensional data of human face region as Y, X value.
In the step 106, set up the contact of corresponding point for convenience, can utilize the three-dimensional model editing system, as shown in Figure 3, use this system can finish selection to each point in the model, and can in advance the corresponding three-dimensional model point of each point in the profile all be chosen, and leave in the middle of the corresponding file.
In having obtained human face profile and general three-dimensional model, after the point of correspondence, just can carry out the match (step 107) of model.The principle of match is exactly to adjust the point corresponding with point in the universal model, makes the distribution of organ of whole three-dimensional face model similar with the organ site in the real human face.
In the step 107, except with the point of point correspondence, also have a large amount of non-reference mark to need to adjust in the general three-dimensional model.The reposition of these points can be finished by the method for interpolation.As shown in Figure 4, A, B, C, P point are the positions at original point (corresponding to the model reference mark of demarcating in the general three-dimensional model), and A ', B ', C ', P ' point (the human face profile reference mark that extracts from front face) are adjusted positions.The p point is the non-reference mark in the general three-dimensional model, and it can be by leg-of-mutton three summit A, B, and C comes unique expression.Owing to can set up two corresponding relations (affine relation) between the triangle in advance, so the adjusted position p ' of p point just can uniquely decide.To each the non-reference mark in the universal model, can judge at first which triangle scope it drops within, just can utilize above-mentioned method to determine adjusted position then.
Need at last promptly to calculate the texture coordinate of each model points correspondence for the model after the match adds texture (step 108).Its method and model fitting are similar, to the reference mark in the model, directly the texture coordinate of the point of correspondence are composed to it and are got final product.To non-reference mark, its adjusted position p ' can obtain, and the texture coordinate on three summits of triangle A ' B ' C ' is known, utilizes simple texture interpolation method just can calculate p ' texture coordinate.
As shown in Figure 5, for utilizing the result after the present invention carries out the modeling of people's face, in certain rotating range, can obtain good effect.
Claims (10)
1, a kind of fast human face model building based on single-sheet photo is characterized in that, comprises the steps:
(1) a front face photograph of input and a universal three-dimensional human face model;
(2), orient real human face organ contours position, and determine profile reference mark wherein at the front face photograph;
(3), calibrate the model reference mark corresponding with described profile reference mark at universal three-dimensional human face model;
(4) set up getting in touch between described model reference mark and the described corresponding profile reference mark;
(5) according to getting in touch between described model reference mark and the profile reference mark, utilize universal three-dimensional human face model to carry out model fitting, come the adjustment model reference mark according to the position at profile reference mark, make the organ of whole three-dimensional face model distribute similar with the organ site in the real human face;
(6) model after the match is added texture.
2, the method for claim 1 is characterized in that, in the described step (2), orients the step of human face outline position in the front face image, comprising:
(21), go out the initial position of human face profile in the front face image based on the ASM model orientation at the front face photograph;
(22) in conjunction with the facial contour actual position in the photograph, the initial position of orienting is finely tuned, obtain the accurate position of human face profile.
3, the method for claim 1, it is characterized in that, in the described step (5), in the model fitting process, further, further make the organ of whole three-dimensional face model distribute similar with the organ site in the real human face by the non-reference mark in the method adjustment universal three-dimensional human face model of interpolation.
4, method as claimed in claim 3 is characterized in that, the adjustment at described non-reference mark comprises:
In universal three-dimensional human face model, determine the triangulation network formed by each model reference mark, and in the real human face organ contours definite triangulation network of forming by the profile reference mark of correspondence;
According to the corresponding triangle in universal three-dimensional human face model and real human face organ contours, set up the affine relation between per two corresponding triangles;
Judge which delta-shaped region the non-reference mark in the universal three-dimensional human face model falls in, utilize described affine relation, determine this adjusted position, non-reference mark according to the corresponding triangle in the real human face organ contours.
5, the method for claim 1 is characterized in that, in the described step (6), for the reference mark, directly the texture coordinate assignment at the reference mark in the human face profile of correspondence is given the reference mark in the model after the match, obtains texture.
6, method as claimed in claim 5 is characterized in that, in the described step (6), for non-reference mark, according to the texture coordinate of triangular apex corresponding in the real human face organ contours, goes out the texture coordinate at adjusted non-reference mark by the texture interpolation calculation.
7, a kind of fast face modeling based on single-sheet photo is characterized in that, comprising:
True locations of contours device is used for individual front face photograph at input, orients real human face organ contours position, and determines profile reference mark wherein;
The model calibration device is used for the universal three-dimensional human face model at input, calibrates the model reference mark corresponding with described profile reference mark, and sets up getting in touch between described model reference mark and the corresponding profile reference mark;
The model fitting device, be used for according to getting in touch between described model reference mark and the profile reference mark, utilize universal three-dimensional human face model to carry out model fitting, come the adjustment model reference mark according to the position at profile reference mark, make the organ of whole three-dimensional face model distribute similar with the organ site in the real human face;
The texture adding set is used for the model after the match is added texture.
8, system as claimed in claim 7, it is characterized in that, described true locations of contours device, at the front face photograph, go out the initial position of human face profile in the front face image based on the ASM model orientation, and in conjunction with the facial contour actual position in the photograph, the initial position of orienting is finely tuned, obtain the accurate position of human face profile.
9, system as claimed in claim 7, it is characterized in that, described model fitting device, in the model fitting process, further the interpolation adjustment is carried out at the non-reference mark in the universal three-dimensional human face model, further made the organ of whole three-dimensional face model distribute similar with the organ site in the real human face.
10, system as claimed in claim 7 is characterized in that, described texture adds module, and the texture coordinate corresponding by computation model point adds texture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2006100888579A CN1889129A (en) | 2006-07-20 | 2006-07-20 | Fast human face model building method and system based on single-sheet photo |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2006100888579A CN1889129A (en) | 2006-07-20 | 2006-07-20 | Fast human face model building method and system based on single-sheet photo |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1889129A true CN1889129A (en) | 2007-01-03 |
Family
ID=37578404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2006100888579A Pending CN1889129A (en) | 2006-07-20 | 2006-07-20 | Fast human face model building method and system based on single-sheet photo |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1889129A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751689B (en) * | 2009-09-28 | 2012-02-22 | 中国科学院自动化研究所 | Three-dimensional facial reconstruction method |
| CN103116902A (en) * | 2011-11-16 | 2013-05-22 | 华为软件技术有限公司 | Three-dimensional virtual human head image generation method, and method and device of human head image motion tracking |
| CN103646416A (en) * | 2013-12-18 | 2014-03-19 | 中国科学院计算技术研究所 | Three-dimensional cartoon face texture generation method and device |
| CN104361362A (en) * | 2014-11-21 | 2015-02-18 | 江苏刻维科技信息有限公司 | Method for obtaining locating model of human face outline |
| CN104463777A (en) * | 2014-11-11 | 2015-03-25 | 厦门美图之家科技有限公司 | Human-face-based real-time depth of field method |
| CN106652015A (en) * | 2015-10-30 | 2017-05-10 | 深圳超多维光电子有限公司 | Virtual figure head portrait generation method and apparatus |
| CN106652037A (en) * | 2015-10-30 | 2017-05-10 | 深圳超多维光电子有限公司 | Face mapping processing method and apparatus |
| CN108305312A (en) * | 2017-01-23 | 2018-07-20 | 腾讯科技（深圳）有限公司 | The generation method and device of 3D virtual images |
| CN108776983A (en) * | 2018-05-31 | 2018-11-09 | 北京市商汤科技开发有限公司 | Based on the facial reconstruction method and device, equipment, medium, product for rebuilding network |
| CN109191505A (en) * | 2018-08-03 | 2019-01-11 | 北京微播视界科技有限公司 | Static state generates the method, apparatus of human face three-dimensional model, electronic equipment |
| CN109447043A (en) * | 2018-12-23 | 2019-03-08 | 广东腾晟信息科技有限公司 | A kind of face method for automatic modeling |
| WO2019201027A1 (en) * | 2018-04-18 | 2019-10-24 | 腾讯科技（深圳）有限公司 | Face model processing method and device, nonvolatile computer-readable storage medium and electronic device |
| CN111862329A (en) * | 2020-07-16 | 2020-10-30 | 广州安廷数字技术有限公司 | Independent static data city modeling method and system |
-
2006
- 2006-07-20 CN CNA2006100888579A patent/CN1889129A/en active Pending
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751689B (en) * | 2009-09-28 | 2012-02-22 | 中国科学院自动化研究所 | Three-dimensional facial reconstruction method |
| CN103116902A (en) * | 2011-11-16 | 2013-05-22 | 华为软件技术有限公司 | Three-dimensional virtual human head image generation method, and method and device of human head image motion tracking |
| CN103646416A (en) * | 2013-12-18 | 2014-03-19 | 中国科学院计算技术研究所 | Three-dimensional cartoon face texture generation method and device |
| CN104463777B (en) * | 2014-11-11 | 2018-11-06 | 厦门美图之家科技有限公司 | A method of the real time field depth based on face |
| CN104463777A (en) * | 2014-11-11 | 2015-03-25 | 厦门美图之家科技有限公司 | Human-face-based real-time depth of field method |
| CN104361362A (en) * | 2014-11-21 | 2015-02-18 | 江苏刻维科技信息有限公司 | Method for obtaining locating model of human face outline |
| CN106652037B (en) * | 2015-10-30 | 2020-04-03 | 深圳超多维光电子有限公司 | Face mapping processing method and device |
| CN106652015A (en) * | 2015-10-30 | 2017-05-10 | 深圳超多维光电子有限公司 | Virtual figure head portrait generation method and apparatus |
| CN106652037A (en) * | 2015-10-30 | 2017-05-10 | 深圳超多维光电子有限公司 | Face mapping processing method and apparatus |
| CN106652015B (en) * | 2015-10-30 | 2020-10-30 | 深圳超多维光电子有限公司 | Virtual character head portrait generation method and device |
| CN108305312A (en) * | 2017-01-23 | 2018-07-20 | 腾讯科技（深圳）有限公司 | The generation method and device of 3D virtual images |
| WO2019201027A1 (en) * | 2018-04-18 | 2019-10-24 | 腾讯科技（深圳）有限公司 | Face model processing method and device, nonvolatile computer-readable storage medium and electronic device |
| US11257299B2 (en) | 2018-04-18 | 2022-02-22 | Tencent Technology (Shenzhen) Company Limited | Face model processing for facial expression method and apparatus, non-volatile computer-readable storage-medium, and electronic device |
| CN108776983A (en) * | 2018-05-31 | 2018-11-09 | 北京市商汤科技开发有限公司 | Based on the facial reconstruction method and device, equipment, medium, product for rebuilding network |
| CN109191505A (en) * | 2018-08-03 | 2019-01-11 | 北京微播视界科技有限公司 | Static state generates the method, apparatus of human face three-dimensional model, electronic equipment |
| CN109447043A (en) * | 2018-12-23 | 2019-03-08 | 广东腾晟信息科技有限公司 | A kind of face method for automatic modeling |
| CN111862329A (en) * | 2020-07-16 | 2020-10-30 | 广州安廷数字技术有限公司 | Independent static data city modeling method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1889129A (en) | Fast human face model building method and system based on single-sheet photo | |
| CN109872397B (en) | Three-dimensional reconstruction method of airplane parts based on multi-view stereo vision | |
| CN110211043B (en) | Registration method based on grid optimization for panoramic image stitching | |
| US9942535B2 (en) | Method for 3D scene structure modeling and camera registration from single image | |
| CN110717494B (en) | Android mobile terminal indoor scene three-dimensional reconstruction and semantic segmentation method | |
| CN101383055B (en) | Three-dimensional human face constructing method and system | |
| CN107240129A (en) | Object and indoor small scene based on RGB D camera datas recover and modeling method | |
| CN107160395A (en) | Map constructing method and robot control system | |
| CN105701455A (en) | Active shape model (ASM) algorithm-based face characteristic point acquisition and three dimensional face modeling method | |
| CN104463969B (en) | A kind of method for building up of the model of geographical photo to aviation tilt | |
| CN102521586B (en) | High-resolution three-dimensional face scanning method for camera phone | |
| CN106780573B (en) | A kind of method and system of panorama sketch characteristic matching precision optimizing | |
| WO2021249401A1 (en) | Model generation method and apparatus, image perspective determining method and apparatus, device, and medium | |
| CN107631700B (en) | The three-dimensional vision information method that spatial digitizer is combined with total station | |
| CN112132876B (en) | Initial pose estimation method in 2D-3D image registration | |
| JP2024525703A (en) | Three-dimensional dynamic tracking method, device, electronic device and storage medium | |
| CN115631317B (en) | Tunnel lining ortho-image generation method and device, storage medium and terminal | |
| CN107578469A (en) | A kind of 3D human body modeling methods and device based on single photo | |
| CN111274944A (en) | Three-dimensional face reconstruction method based on single image | |
| CN104200476B (en) | The method that camera intrinsic parameter is solved using the circular motion in bimirror device | |
| CN107330980A (en) | A kind of virtual furnishings arrangement system based on no marks thing | |
| CN104596442A (en) | Assisted three-dimensional scanning device and method | |
| CN105321181A (en) | Method for calibrating parabolic catadioptric camera by using separate image of double balls and image of circular point | |
| Lee et al. | Interactive 3D building modeling using a hierarchical representation | |
| CN110348344A (en) | A method of the special facial expression recognition based on two and three dimensions fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |