CN100386778C - Human face image age changing method based on average face and senile proportional image - Google Patents
Human face image age changing method based on average face and senile proportional image Download PDFInfo
- Publication number
- CN100386778C CN100386778C CNB2006100429898A CN200610042989A CN100386778C CN 100386778 C CN100386778 C CN 100386778C CN B2006100429898 A CNB2006100429898 A CN B2006100429898A CN 200610042989 A CN200610042989 A CN 200610042989A CN 100386778 C CN100386778 C CN 100386778C
- Authority
- CN
- China
- Prior art keywords
- old
- average
- face
- image
- feeble
- 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
Images
Abstract
The present invention discloses a human face image age conversion method based on an average face and an ageing proportion chart. Dense characteristics are used for expression, human face images are analyzed into a shape vector and a texture vector by means of human face cartoon synthesis technique by calculating the difference between a specific human face image and the average face, and thereby, the difference degree is reduced to complete the young conversion of a specific human face. A human face vectorization modeling and human face image ageing conversion are combined, and human face image ageing simulation which is in accordance with physiological characteristics is achieved. Definite modification is carried out to characteristic lines of human face images according to the variation of human face shape characteristics caused by ageing, and the aging simulation of shapes is realized. For the variation of textural characteristics, people manufacture an ageing proportion chart in a proportion chart mode, and the ageing simulation of textural characteristics of human face images is realized according to texture mapping. The method fully considers the uniform variation of shapes and textures of human faces when ages are changed, and the sense of reality of composite images is enhanced greatly.
Description
Technical field
The present invention relates to computer vision and image processing field, be particularly related to and a kind ofly carry out fast age transform method true to nature, comprise based on the young transform method of the facial image of average face with based on the old and feeble transform method of the facial image of old and feeble scale map at front face image.
Background technology
People's face is the strongest part of human expressive force, and its variation has comprised many important information.The research of computer simulation human face animation can be traced back to eighties of last century seventies, continue to bring out out afterwards breakthrough achievement in research, this field also day by day becomes many researchists' of industries such as police criminal detection, production of film and TV, Entertainment interest place, the age conversion of facial image has profound significance as the pith of human face animation.
Though the macroscopic features of people's face is relevant with the age, even but also can there be very big difference in the macroscopic features that the people of same age presented because of the influence of various factorss such as individual living environment, habits and customs, so we are difficult to judge from the width of cloth facial image people's the accurate age.People self is not inherent to the perception and the judgement at age, and the people can constantly obtain the various information of the environmental feedback that comes from the outside in self developmental process, and utilizes these information constantly to revise own roughly judgement to people's face age.Because the concrete forming process of this ability varies with each individual, therefore be difficult to express, so the research of this respect does not also have a gratifying complete explanation so far with the knowledge of some quantification.For computing machine, distinguish the age problem of a difficulty especially.And on the other hand, for two width of cloth facial images, it is younger or who is more old whom we but can recognize easily, and human young with old and feeble essential characteristic can be described out, the smooth degree of skin for example, wrinkle and spot how much, the shape of bone, the relax level of muscle etc.Utilize image processing techniques to realize that the youth and the old and feeble conversion of facial image are the advanced problems of facial image process field, this technology depends on interdisciplinary theoretical support the such as physiology, cognitive science, psychology greatly and merges with mutual.Though the appearance changing features highly significant by change of age causes has abundant researching value, has only less researchist that the change of age of facial image is studied.
Thereby the shape that Thompson proposition in 1961 uses method of coordinates transform to change biologic-organ is simulated other individual organ characteristics in the similar biology.Based on this thought, many researchists use different coordinate transformation methods to simulate the change of the people's face macroscopic features that causes because of change of age.Wherein main transform method has cardioidal strain conversion, and it passes through to change the simulating shape of people's face and skull along with the caused characteristic change of change of age.The experimental result explanation uses cardioidal strain method can change the age characteristics of 2D facial image profile preferably.Mark and Todd are applied to 3D people's face data to its further extension with this method afterwards.This old and feeble analogy method based on coordinate transform has only considered that people's face shape changes, and does not specify for the texture variations that causes.
Nineteen ninety-five Burt is divided into 7 different sample groups with Perrett with the age-based section of sample image, calculates the average facial image of corresponding these 7 groups respectively, utilizes the caricature synthetic technology to simulate old and feeble the variation then.From experimental result as can be seen, the age of the age of most of synthetic facial image with the sample object that is used to calculate average face is consistent, and this explanation is in the process of calculating average facial image, and the age information of each age group has all kept.They have also described two kinds of old and feeble analogue techniques: first method, the at first heterochromia between average man's face that average man's face of the more old age group correspondence of calculating is corresponding with all images in the sample set.Then use the caricature synthetic technology to enlarge this species diversity and increase age in composograph.Second method, calculate 25-29 year and 50-54 the year age group correspondence the shape and the color distortion of average man's face.Realize increasing and change age thereby this species diversity is applied to new facial image and increases this species diversity.From they experimental result as can be seen, two kinds of all better old and feeble conversion that must realize facial image of method.The method that Burt and Perrett propose has been considered the people's face shape that old and feeble variation causes and the variation of texture information simultaneously, and effect also is very significant.
O ' Toole etc. sets up 3D people's face parameter model according to 3D people's face information.Use the caricature synthetic technology that significant 3D face characteristic is exaggerated or eliminated then.Description of test, the notable feature that comprises in the older facial image is abundanter also more obvious, therefore makes the increase or the minimizing that can realize the implicit age in the image in this way.This method depends on the specific descriptions to people's face appearance, because 3D people's face data obtain by the 3D scanner, equipment is very expensive, and needs a large amount of processing times, and therefore the practicality of this method is not strong.
Lanitis etc. express training sample image again based on the statistical model of a facial image, and getting in touch between the expression data by the learning training sample and its actual age set up old and feeble variation model then.After the modelling, can estimate from the people's who does not see age according to the expression data of new image, otherwise, also can the picture of a people of reconstruct under any age.They have used different training samples to set up different old and feeble changing patteries.Experimental results show that this method can carry out rational estimation of Age to the image of the unknown, and can select only old and feeble changing method, simulate aging course automatically for the new image of any width of cloth.
Also have a large amount of semi-automatic change of age disposal systems to be mainly used to simulate the present looks of losing children.Basic operation mainly is according to the caused characteristic change of change of age the face characteristic of correspondence to be made amendment.The appearance of considering the people has certain heredity, so generally all be according to the photo of all ages and classes section of losing the children close relative, children's appearance is made amendment.
The other correlative study mainly is around the modeling of people's face skin and the simulation of old and feeble wrinkle.Boissieux etc. see skin as elastic membrane, further it are extended to the stereoscopic model that multilayer tissue constitutes, and the change shape of wrinkle uses Finite Element Method to calculate skin deformation during according to aging, thereby produce the wrinkle texture.People such as Viaud have developed a kind of how much mixture models, with batten piecewise function simulation fold, can produce expression and age wrinkle, and use the age parameter to control the generation of age wrinkle.Z-Correct bump mapping Z-correct is shone upon and the color map technology in addition, texture synthesis language, and little geometric model etc., these technology can be used for simulating different dermatoglyph patterns.
Because in the conversion of the age of facial image, the variation complexity of the architectural feature of people's face and skin histology makes the anthropomorphic face age conversion of computer mould have very big challenge.The sense of reality of many again more complicated of existing human face image age transform method and composograph is not fully up to expectations at present, therefore the present invention is directed to the deficiency of algorithm in the past, proposed based on the young conversion of the facial image of average face and based on the old and feeble transform method of the facial image of old and feeble scale map.
Summary of the invention
The objective of the invention is to overcome the shortcoming of prior art, propose a kind of facial image age transform method based on average face and old and feeble scale map.Calculated amount of the present invention is little, and conversion process is simple, and synthetic result is level and smooth, nature, true, has very strong validity and versatility.
Technical scheme of the present invention is achieved in that this method may further comprise the steps:
A) based on the young shift step of average face:
Suppose one group of front sample facial image { I is arranged
i, the corresponding reference face is I
Ref, new facial image is I
N, be I through the image that obtains after the conversion
R
The first step: with average face as reference face I
RefSituation under, with new facial image I
NCarry out dense feature representation, obtain corresponding shape and texture vector, be designated as (S
N, T
N),
Second step: calculate new facial image I
NAspect shape and the texture and the difference between the average characteristics, be designated as (Δ S, Δ T), wherein Δ S=S
N-S
Average, Δ T=T
N-T
Average(S
Average, T
Average) be average face shape and texture vector,
The 3rd step: utilize parameter alpha 1, β 1 to adjust new facial image and shape facility between the average face and the difference degree between the textural characteristics, Δ S '=α 1 Δ S, Δ T '=β 1 Δ T, wherein 0＜α 1, and β 1＜1, and α 1, β 1 are more little, then the people's face after the conversion is young more
The 4th step: the shape vector and the texture vector (S of the new facial image after the calculated difference degree changes
R, T
R), computing formula is S
R=Δ S '+S
Average, T
R=Δ T '+T
Average,
The 5th step: according to (S
R, T
R), and with reference to face I
Ref, reconstruct facial image I after the young conversion by forward direction warp
R
B) based on the old and feeble shift step of old and feeble scale map:
The old and feeble conversion of facial image is divided into off-line phase and online stage:
Off-line phase: mainly finish S
OldWith the calculating of R,
The first step:, adjust the characteristic curve L of average face according to old and feeble Changing Pattern
YoungMake canthus, the corners of the mouth, outside eyebrow, drooping nose obtain L
Old
Second step: with L
YoungBe reference, to L
OldUse obtains corresponding S based on the Warp operation of characteristic curve
Old
The 3rd step: other parts except that eye and mouth in the old and feeble sample image are carried out low-pass filtering, obtain corresponding young image;
The 4th step: old and feeble sample image is carried out vector quantization express, obtain texture vector T
Old, young image is carried out vector quantization express, obtain texture vector T
Young
The 5th step: calculate old and feeble scale map R=T
Old/ T
Young
The online stage: the old and feeble image of synthetic new images correspondence;
The first step: with the average face is reference, with facial image I
nAgain express, be designated as (S
n, T
n);
Second step: go on foot the S that obtains according to off-line phase second
Old, calculate the shape vector S ' after the old and feeble conversion
n
The 3rd step: the R that uses the 5th step of off-line phase to obtain calculates the texture vector T ' after the old and feeble conversion
n
The 4th step: according to (S '
n, T '
n), utilize the image I after forward direction Warp technology is synthesized conversion
n
Described with the reference face of average face as facial image, promptly the characteristic curve of one group of sample image being distributed obtains the distribution of standard feature line as weighted mean, and note is L
Average, according to the mapping relations of special characteristic line and standard feature line, can finish the decomposition of shape and textural characteristics to each width of cloth sample image, obtain corresponding image expression (S
i, T
i), this group shape of sample and texture vector average, i.e. average face I
AverageImage expression (S
Average, T
Average), all images all is that the standard feature line at average face is out of shape, shape vector is described is relative variation with the average face feature, so the mean value of shape vector is 0, so the vector quantization of average face be expressed as (0, T
Average).
Shape vector S ' after the old and feeble conversion of described calculating
nComputing formula be S '
n=S
n+ α 2S
Old, wherein α 2＞0.
The R that described the 5th step of use off-line phase obtains calculates the texture vector T ' after the old and feeble conversion
nComputing formula be T '
n=T
n(β 2 (R-1)+1), wherein β 2＞0.
The present invention has following several characteristics:
1), finds that average face is implying young sensation to a certain extent according to the related conclusions of psychology and cognitive science.The present invention passes through to calculate the difference between specific facial image and the average face, and then reduces this difference degree, finishes the young conversion of specific people's face.
2) in the young conversion, based on the caricature synthetic technology, adopt dense characteristic to express to facial image, the point in every width of cloth facial image is all as a feature, seek the difference of each point and average face image corresponding point in the image, finally reduce the disparity range of specific people's face and average face feature.
3) the old and feeble conversion of modeling of facial image vector quantization and facial image is combined, realized meeting the old and feeble prediction of facial image of physiological characteristic.Adjust change of shape in people's face aging course according to the notion of otherness in the psychological study (Distinctiveness) and human-face cartoon synthetic technology, adjust the texture variations of people's face in aging course by old and feeble scale map.This method need not to set up complicated model, and calculated amount is little, and fully takes into account that shape and the unified of texture change after the aging of people's face, has improved the validity and the versatility of this method.
4) notion of otherness (Distinctiveness) is directly introduced the age conversion of 2D facial image, the extensive application category of aesthstic notion, synthetic image has the very strong sense of reality.
Description of drawings
Fig. 1 is the old and feeble Changing Pattern figure of inventor's face;
Fig. 2 is a characteristic curve distortion displayed map of the present invention;
Fig. 3 is the old and feeble scale map exemplary plot of the present invention;
Below in conjunction with accompanying drawing content of the present invention is described in further detail.
Embodiment
With reference to shown in Figure 1, enumerated the Physiologic Studies conclusion of people's face aging.The people is in aging course, and face comprises that each ingredient such as skin, eyebrow, face all may present as scheming cited typical change feature, and as cutis laxa, wrinkle produces, and the canthus corners of the mouth is sagging etc.We have simulated the characteristic feature of people's face agings such as the canthus here, the corners of the mouth, eyebrows and drooping nose in follow-up algorithm.
With reference to shown in Figure 2, the arithmetic mean of all sample institute marker characteristic lines is the L here
Young, it has reflected the shape facility of average face; According to the biology rule of people's face aging that Fig. 1 disclosed, to L
YoungEvery line segment carry out manual setting, obtain general old and feeble pattern L respectively
Old, thin partially old and feeble pattern L
OldthinWith fat partially old and feeble pattern L
Oldfat
With reference to shown in Figure 3, the implication of scale map is the merchant of two width of cloth images.The old and feeble scale map here then is the merchant between old and feeble facial image and its low-pass filtering image, and it has reflected the dermatoglyph feature of old and feeble people's face.Three width of cloth image table global change of face texture in aging course of leting others have a look in first row, three width of cloth images in second row are people's face texture localized variation in aging course, respectively corresponding forehead, eyes and cheek.
From the angle of computer graphics, caricature is meant: the exaggeration of notable feature with emphasize.Typical facial image caricature synthetic technology is exactly to calculate specific face characteristic and the corresponding on average difference between the face characteristic, by enlarging this species diversity, realizes the exaggeration of feature.Also this species diversity can be diminished, this processing means can be summed up as a kind of synthetic category of caricature of broad sense.People's face otherness mainly is the feature difference degree of people's face on people's face of portrayal and " on average " meaning.
With the facial image vector representation, i.e. gray values of pixel points in each element correspondence image in the vector, then each image vector can regard as on the higher dimensional space a bit.All facial images are distributed in the zone in the higher dimensional space, claim this zone to be people's face image space.From the angle of Neuropsychology, in such high-dimensional feature space, a people's principal character is by the direction decision between 2 of specific face and the average faces.Therefore move along this direction, change be exactly the otherness of people's face.Specifically, the caricature synthetic technology of facial image is equivalent between average face and specific face to draw a line, then along this line towards average face or the direction that deviates from average face move, obtain the corresponding new point of facial image in the space.
People's face otherness defines at average face, and therefore calculating average face is the first step of finishing this work.The image expression mode that the present invention has adopted shape to separate with texture vector, what therefore should calculate here is the shape and the texture vector of average face.
The characteristic curve of one group of sample image distributed obtains the distribution of standard feature line as weighted mean, note is L
AverageAccording to the mapping relations of special characteristic line and standard feature line, can finish the decomposition of shape and textural characteristics to each width of cloth sample image, obtaining corresponding image expression is (S
i, T
i).This group shape of sample and texture vector average is average face I
AverageImage expression (S
Average, T
Average).In fact, all images all is that the standard feature line at average face is out of shape, and shape vector is described is relative variation with the average face feature, so the mean value of shape vector is 0, thus the vector quantization of average face be expressed as (0, T
Average).In young the 3rd step of conversion, parameter alpha 1, β 1 are independent of each other, control the shape and the texture variations of people's face respectively.Can see that value is in [0,1] scope, what that is to say realization here is reducing of feature difference, rather than exaggeration, and some is different for this and caricature synthetic technology.The present invention has also used a power exponential function foundation contact between the two when realizing, realize both interlocks.
For old and feeble mapping algorithm, with advancing age, the skin of people's face, fat, muscle and bone all can change.Its principal character changes sees Fig. 1.Change of shape and grain details that these features have comprised people's face change, and therefore will realize the old and feeble conversion of facial image, except aspect shape of face, will changing, and also will be aspect texture to putting forth effort to portray such as grain details such as wrinkle, senile plaque expellings.
Use the caricature synthetic technology can't depict and change, so the present invention adopts the method based on old and feeble scale map to realize that the aging of facial image changes such as textural characteristics such as wrinkles.The variation of the people's face shape feature that causes according to aging is carried out certain modification to the characteristic curve of facial image, realizes the aging simulation of shape; For the change of textural characteristics, adopt the scale map mode, make old and feeble scale map, according to texture, realize the old and feeble simulation of textural characteristics of facial image.
About the aging simulation of shape, with the characteristic curve L of average face
YoungBe reference, the old man's face characteristic curve L that uses the Warp technology based on characteristic curve that manual setting is got
OldDistortion obtains optical flow field S
OldThe shape vector of supposing a width of cloth new images is S
n, then can obtain the shape vector S ' after aging is simulated
n=S
n+ S
Old
Finish old and feeble texture vector simulation by means of the notion of scale map.Therefore scale map is based on that two width of cloth image calculation obtain, but obtains a people in the facial image of all ages and classes section similarity condition difficulty comparatively, and old and feeble scale map is based on that single image carries out.In people's face sample image of aging to except eye, low-pass filtering is carried out at other positions beyond the mouth, the smoothed image that obtains like this is considered as the young image of this person's correspondence, and then finishes texture transformation in the aging course by means of the thought of texture.
Before calculating scale map, be reference with the average face, old and feeble sample image and corresponding young image are carried out the vector quantization expression, finish the aligning between the unique point, obtain corresponding texture vector T
OldAnd T
Young, then old and feeble scale map is by formula R=T
Old/ T
YoungCalculate.For the new facial image of a width of cloth, its texture vector is T
n, the texture vector after the then old and feeble simulation is T '
n=T
n(β 2 (R-1)+1).Use such computing formula can guarantee that when texture non-crumple zone does not change and wrinkle can be deepened gradually along with the change of β 2.
Claims (4)
1. based on the facial image age transform method of average face and old and feeble scale map, it is characterized in that this method may further comprise the steps:
A) based on the young shift step of average face:
Suppose one group of front sample facial image { I is arranged
i, the corresponding reference face is I
Ref, new facial image is I
N, be I through the image that obtains after the conversion
R
The first step: with average face as reference face I
RefSituation under, with new facial image I
NCarry out dense feature representation, obtain corresponding shape and texture vector, be designated as (S
N, T
N),
Second step: calculate new facial image I
NAspect shape and the texture and the difference between the average characteristics, be designated as (Δ S, Δ T), wherein Δ S=S
N-S
Average, Δ T=T
N-T
Average, (S
Average, T
Average) be average face shape and texture vector,
The 3rd step: utilize parameter alpha 1, β 1 to adjust new facial image and shape facility between the average face and the difference degree between the textural characteristics, Δ S '=α 1 Δ S, Δ T '=β 1 Δ T, wherein 0＜α 1, and β 1＜1, and α 1, β 1 are more little, then the people's face after the conversion is young more
The 4th step: the shape vector and the texture vector (S of the new facial image after the calculated difference degree changes
R, T
R), computing formula is S
R=Δ S '+S
Average, T
R=Δ T '+T
Average,
The 5th step: according to (S
R, T
R), and with reference to face I
Ref, reconstruct facial image I after the young conversion by forward direction warp
R
B) based on the old and feeble shift step of old and feeble scale map:
The old and feeble conversion of facial image is divided into off-line phase and online stage:
Off-line phase: mainly finish S
OldWith the calculating of R,
The first step:, adjust the characteristic curve L of average face according to old and feeble Changing Pattern
YoungMake canthus, the corners of the mouth, outside eyebrow, drooping nose obtain L
Old
Second step: with L
YoungBe reference, to L
OldUse obtains corresponding S based on the Warp operation of characteristic curve
Old
The 3rd step: other parts except that eye and mouth in the old and feeble sample image are carried out low-pass filtering, obtain corresponding young image;
The 4th step: old and feeble sample image is carried out vector quantization express, obtain texture vector T
Old, young image is carried out vector quantization express, obtain texture vector T
Young
The 5th step: calculate old and feeble scale map R=T
Old/ T
Young
The online stage: the old and feeble image of synthetic new images correspondence;
The first step: with the average face is reference, with facial image I
nAgain express, be designated as (S
n, T
n);
Second step: go on foot the S that obtains according to off-line phase second
Old, calculate the shape vector S after the old and feeble conversion
n';
The 3rd step: the R that uses the 5th step of off-line phase to obtain calculates the texture vector T after the old and feeble conversion
n';
The 4th step: according to (S
n', T
n'), utilize the image I after forward direction Warp technology is synthesized conversion
n
2. according to right 1 described method, it is characterized in that, described with average face as the reference face, promptly the characteristic curve of one group of sample image being distributed obtains the distribution of standard feature line as weighted mean, note is L
Average, according to the mapping relations of special characteristic line and standard feature line, can finish the decomposition of shape and textural characteristics to each width of cloth sample image, obtain corresponding image expression (S
i, T
i), this group shape of sample and texture vector average, i.e. average face I
AverageImage expression (S
Average, T
Average), all images all is that the standard feature line at average face is out of shape, shape vector is described is relative variation with the average face feature, so the mean value of shape vector is 0, so the vector quantization of average face be expressed as (0, T
Average).
3. according to right 1 described method, it is characterized in that the shape vector S after the old and feeble conversion of described calculating
n' computing formula be S
n'=S
n+ α 2S
Old, wherein α 2＞0.
4. according to right 1 described method, it is characterized in that the R that described the 5th step of use off-line phase obtains calculates the texture vector T after the old and feeble conversion
n' computing formula be T
n'=T
n(β 2 (R-1)+1), wherein β 2＞0.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100429898A CN100386778C (en) | 2006-06-15 | 2006-06-15 | Human face image age changing method based on average face and senile proportional image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100429898A CN100386778C (en) | 2006-06-15 | 2006-06-15 | Human face image age changing method based on average face and senile proportional image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1870047A CN1870047A (en) | 2006-11-29 |
| CN100386778C true CN100386778C (en) | 2008-05-07 |
Family
ID=37443704
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2006100429898A Expired - Fee Related CN100386778C (en) | 2006-06-15 | 2006-06-15 | Human face image age changing method based on average face and senile proportional image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100386778C (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8391639B2 (en) * | 2007-07-23 | 2013-03-05 | The Procter & Gamble Company | Method and apparatus for realistic simulation of wrinkle aging and de-aging |
| US20100289912A1 (en) * | 2009-05-14 | 2010-11-18 | Sony Ericsson Mobile Communications Ab | Camera arrangement with image modification |
| TWI467498B (en) | 2011-12-19 | 2015-01-01 | Ind Tech Res Inst | Method and system for recognizing images |
| CN102799276B (en) * | 2012-07-18 | 2016-06-01 | 上海量明科技发展有限公司 | The method of avatar icon age conversion, client terminal and system in instant messaging |
| CN104063842A (en) * | 2014-05-30 | 2014-09-24 | 小米科技有限责任公司 | Image processing method and device and terminal |
| KR101758825B1 (en) * | 2015-09-22 | 2017-07-18 | 한국과학기술연구원 | Method for facial age simulation based on age of each part and environmental factors, recording medium and device for performing the method |
| CA3015492C (en) | 2016-03-21 | 2021-11-23 | The Procter & Gamble Company | Systems and methods for providing customized product recommendations |
| CN106651978B (en) * | 2016-10-10 | 2021-04-02 | 讯飞智元信息科技有限公司 | Face image prediction method and system |
| US10614623B2 (en) | 2017-03-21 | 2020-04-07 | Canfield Scientific, Incorporated | Methods and apparatuses for age appearance simulation |
| US10621771B2 (en) | 2017-03-21 | 2020-04-14 | The Procter & Gamble Company | Methods for age appearance simulation |
| WO2018222812A1 (en) | 2017-05-31 | 2018-12-06 | The Procter & Gamble Company | System and method for guiding a user to take a selfie |
| CN110709856B (en) | 2017-05-31 | 2023-11-28 | 宝洁公司 | System and method for determining apparent skin age |
| CN113761985A (en) * | 2020-06-05 | 2021-12-07 | 中国科学院上海营养与健康研究所 | Method and apparatus for determining local regions affecting the degree of facial aging |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN1617169A (en) * | 2003-11-12 | 2005-05-18 | 德鑫科技股份有限公司 | Spot eliminating method for digital image |
| CN1645406A (en) * | 2005-02-24 | 2005-07-27 | 北京工业大学 | Identity discriminating method based on eyebrow identification |
-
2006
- 2006-06-15 CN CNB2006100429898A patent/CN100386778C/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN1617169A (en) * | 2003-11-12 | 2005-05-18 | 德鑫科技股份有限公司 | Spot eliminating method for digital image |
| CN1645406A (en) * | 2005-02-24 | 2005-07-27 | 北京工业大学 | Identity discriminating method based on eyebrow identification |
Non-Patent Citations (2)
| Title |
|---|
| 人脸的表情与年龄变换和非完整信息的重构技术(上). 郑南宁，付昀，张婷，卓峰.电子学报，第31卷第12A期. 2003 |
| 人脸的表情与年龄变换和非完整信息的重构技术(上). 郑南宁，付昀，张婷，卓峰.电子学报，第31卷第12A期. 2003 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1870047A (en) | 2006-11-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100386778C (en) | Human face image age changing method based on average face and senile proportional image | |
| CN101556701A (en) | Human face image age changing method based on average face and aging scale map | |
| CN101755288B (en) | Method and apparatus for realistic simulation of wrinkle aging and de-aging | |
| CN106023288B (en) | A kind of dynamic scapegoat's building method based on image | |
| Cheng et al. | Parametric modeling of 3D human body shape—A survey | |
| Yang et al. | Semantic parametric reshaping of human body models | |
| CN101751689B (en) | Three-dimensional facial reconstruction method | |
| CN108288072A (en) | A kind of facial expression synthetic method based on generation confrontation network | |
| Choe et al. | Analysis and synthesis of facial expressions with hand-generated muscle actuation basis | |
| CN103473801A (en) | Facial expression editing method based on single camera and motion capturing data | |
| CN101739719A (en) | Three-dimensional gridding method of two-dimensional front view human face image | |
| CN111950430A (en) | Color texture based multi-scale makeup style difference measurement and migration method and system | |
| CN110443872B (en) | Expression synthesis method with dynamic texture details | |
| Patterson et al. | Comparison of synthetic face aging to age progression by forensic sketch artist | |
| Li et al. | Realistic wrinkle generation for 3D face modeling based on automatically extracted curves and improved shape control functions | |
| CN116740240A (en) | Real-time garment animation generation method with various styles | |
| Kawai et al. | Data-driven speech animation synthesis focusing on realistic inside of the mouth | |
| Li et al. | Computer-aided 3D human modeling for portrait-based product development using point-and curve-based deformation | |
| CN114581288A (en) | Image generation method and device, electronic equipment and storage medium | |
| Mao et al. | A sketch-based approach to human body modelling | |
| CN106960467A (en) | A kind of face reconstructing method and system with bone information | |
| CN110309554A (en) | Video human three-dimensional rebuilding method and device based on clothes modeling and simulation | |
| Agianpuye et al. | Synthesizing neutral facial expression on 3D faces using Active Shape Models | |
| CN115797569B (en) | Dynamic generation method and system for high-precision degree twin facial expression action subdivision | |
| CN106504308A (en) | Face three-dimensional animation generation method based on 4 standards of MPEG |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080507 Termination date: 20210615 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |