CN105719326A - Realistic face generating method based on single photo - Google Patents

Abstract

The invention discloses a realistic face generating method based on a single photo. The method comprises steps of firstly constructing a standard face model base, interactively selecting face feature points of an input picture and based on face profile features, matching an optimal model; secondly, via the triangle deformation and the bilinearity interpolation method, realizing texture mapping from the picture to a three-dimensional model, and by introducing an Alpha graph, realizing fusion transition from face covering textures to model neutral textures; and at last, by adopting the grid adjusting method, hierarchically adjusting the model from the whole to the details so as to achieve an objective of a realistic three-dimensional face. The method is user-friendly and requires quite few feature points; and just with one full face photo and quite few feature points, a user can participate in an autonomously finished realistic face establishing method.

Description

A kind of Realistic Human face generating method based on single photo

Technical field

The present invention relates to a kind of three-dimensional face rapid generation, specifically, relate to a kind of Realistic Human face generating method based on single photo.

Background technology

Face is the pith of emotional expression and identification.In life, we identify identity by face, express happiness, anger, grief and joy by face, and face plays very important effect in our life at ordinary times.Along with the development of computer graphics, the human face rebuilding for special pattern starts to be widely used in a lot of fields.The interest strengthening user is had positive effect with participating in initiative by the three-dimensional role of one realistic face of structure.At game animation, film advertisement, Three Dimensional Campus, the fields such as digital travelling scenic spot, virtual dressing room are as one of means improving user's sensory experience, sense of reality degree is better, manufacturing cost is cheap in research, and generation process faceform's generation technique quickly and easily has great importance.

Existing face rapid generation mainly has by equipment acquisition face three-dimensional datas such as spatial digitizers, and the model fidelity of generation is high, but relies on substantial amounts of cloud data, and technology requires higher, expensive；In addition with using multiple fronts and side image to realize faceform's reconstruction by visual correlation is theoretical with algorithm, the method needs are manual demarcates more characteristic point, and initial conditions is complex.So the problem such as conventional face's model generating method exists apparatus expensive, realizes process complexity, real-time is not enough, the demand of real-time, interactive in virtual scene can not be fully met.

Summary of the invention

It is an object of the invention to the defect overcoming above-mentioned technology to exist, a kind of Realistic Human face generating method based on single photo is provided, the method that this technology uses is to user friendly, required characteristic point is less, only need a full face and less characteristic point, can be participated in, by user, the Realistic Face creation method that independently completes.User is incorporated into the establishment process of face by the method, a small amount of characteristic point is chosen by interactive mode, the faceform that coupling is suitable, realize from covering the texture mapping to neutral texture, and the colour of skin of facial boundary merges, finally by the model adjustment of entirety to level of detail time, generate the faceform with personalization.

Its concrete technical scheme is:

A kind of Realistic Human face generating method based on single photo, comprises the following steps:

Step1. characteristic point is demarcated: obtains front two-dimension human face photo, 7 parts such as eyes, nose, mouth is chosen 13 characteristic points and carries out labelling, it is determined that the approximate location of whole face face；

Step2. Model Matching: compare the characteristic point distance between photo and model, the model that matching similarity is maximum in model library；

Step3. texture maps: the texture mapping simulation of facial true colors according to face, realize the mapping covering texture to neutral texture, neutral texture is the texture of model acquiescence, covering texture is human face photo, by the mapping of 13 characteristic points and peripheral region thereof in two layer texture, add the textural characteristics of human face photo to Matching Model；

Step4. the colour of skin merges: uses one Alpha figure, and the boudary portion in face fusion is formed from covering texture seamlessly transitting to neutrality texture, meanwhile, utilizes Alpha figure that the details such as eyebrow, bridge of the nose position is carried out trickle adjustment；

Step5. model regulates: by the method for vector differentials, the characteristics such as face size, position such as the face of Matching Model, eyes, shape of face are adjusted, and generates the personalized model meeting its facial characteristics.

Preferably, Step5 constructs a Model Regulator that can be used for adjusting human face five-sense-organ slight change, generates personalized face true to nature；

If the summit V in face grid_i=(x, y, z)^TRegard Vector Groups H as₀=(V₁, V₂, V₃..., V_n), n represents the quantity on summit, and (as changed size) after eyes are converted, the position of some points changes, and forms Vector Groups H₁=(V₁', V₂', V₃' ... V_n'), this conversion is increased a weights W, then, from H₀To H₁Middle all of transition state H_xCan be transferred through linear interpolation to calculate；

H_x=H₀+W*(H₁-H₀)(1)

A kind of conversion is difficult to meet the demand of faceform, it is necessary to the feature according to face, selects multiple conversion, gives a weighted value W for each conversion, then the mixed formulation of m conversion is:

$H_x=H_0+{\Σ}_{i=1}^m{W}_i*(H_i-H_1)---\left(2\right)$

Wherein the quantity of m is as the criterion with the quantity needed in practical application, in an experiment, define the grid actuator of 23 kinds of different conversion, including: the covering of eye-level, eye distance, eyes size, eyebrow height, the eyebrow degree of depth, nose size, the nose degree of depth, nose height, the bridge of the nose degree of depth, nose bridge widths, nostril width, face height, face thickness, face width, lip, the cheekbone degree of depth, cheekbone width, cheekbone height, buccal width, chin height, the chin degree of depth, chin width, lower jaw width.Actuator realizes distortion of the mesh by control weight.

Preferably, in Step2, the method for Model Matching is as follows: first, and photo is normalized, and calculates the distance between principal character point and proportionate relationship；Secondly, according to characteristic point distance and proportionate relationship, it is determined that the rough classification of shape of face；Finally, Euclidean distance between Matching Model and photo individual features point in the model library of corresponding shape of face, adopt formula (3) to search out closest threedimensional model；

$E={\mathrm{\Σ}}_{i=1}^n{(D_i-{D_i}^{,})}^2*{\mathrm{\λ}}_i---\left(3\right)$

Wherein, D_iFor distance between certain characteristic point in photo, D_i' for distance between individual features point in threedimensional model, λ_iFor the weighted value of each characteristic point distance, determine according to its percentage contribution size and empirical equation.N is the distance quantity compared.

Preferably, Step3 assuming, the characteristic point position of neutral texture and covering texture labelling respectively forms V respectively₁, V₂Two two-dimentional point sets, due to the difference of manual mark, V₁, V₂In corresponding point necessarily will not be exactly accurate at same position, it is necessary to a kind of mapping method makes its natural fusion, takes P₁∈V₁, P₂∈V₂.By P₁, P₂Expand to three-dimensional vector, P₁=(X₁, Y₁, 1), P₂=(X₂, Y₂, 1), and assume there is matrix M, then can obtain:

$M=P_1*P_2^{-1}---\left(4\right)$

After obtaining the transformation matrix of each point, the conversion of characteristic point is applied in the pixel around characteristic point, this flow process uses common triangle mode of texturing that covering texture is applied to model texture, characteristic point is connected, form the grid covering face, grid now has concurrently triangle and tetragon, by splitting, polygon is all reduced to triangle, is in the pixel in the middle of triangle and is subject to the impact on Atria summit simultaneously；

Three summits assuming triangle are P₁, P₂, P₃, using they three row as matrix, form 3*3 matrix (P₁, P₂, P₃), it is assumed that the triangle after conversion is (P₁', P₂', P₃'), then:

M*(P₁, P₂, P₃)=(P₁', P₂', P₃’)(5)

M=(P₁', P₂', P₃’)*(P₁, P₂, P₃)^-1(6)

Certain i point in the middle of diabolo, uses M*P_iJust can obtain P_i, determine the coordinate after mapping a little in triangle with this, for unmapped point, coordinate figure is all integer, but the coordinate after mapping is likely to not be integer, and causing cannot with the pixel one_to_one corresponding on figure.

Preferably, in actual fused process in Step4, model neutrality texture, Alpha figure, covering texture will be processed into the same size, and the pixel in Alpha figure participates in mixing as the interpolation weights of mixing, ultimately forming the texture of natural fusion, the mixed formulation of three figure is:

C (x, y)=C_base(x, y) * (255-C_Alpha(x, y))+C_overlay(x, y) * C_Alpha(x, y) wherein 0≤C_Alpha(x, y)≤255 (7)

C_base(x, y), C_Alpha(x, y), C_overlay(x y) represents neutral texture, Alpha figure respectively, covers texture in (x, y) color value at coordinate place.

Compared with prior art, the invention have the benefit that

The method that the present invention uses is to user friendly, and required characteristic point is less, only needs a full face and less characteristic point, can be participated in, by user, the Realistic Face creation method that independently completes.User is incorporated into the establishment process of face by the method, a small amount of characteristic point is chosen by interactive mode, the faceform that coupling is suitable, realize from covering the texture mapping to neutral texture, and the colour of skin of facial boundary merges, finally by the model adjustment of entirety to level of detail time, generate the faceform with personalization.

Accompanying drawing explanation

Fig. 1 is that Realistic Face builds flow process；

Fig. 2 is characteristic point position；

Fig. 3 is that characteristic point connects；

Fig. 4 is Alpha figure.

Detailed description of the invention

For the technological means making the present invention realize, creation characteristic, reach purpose and effect and be easy to understand, the present invention is expanded on further below in conjunction with accompanying drawing and instantiation.

First, building standardized Face Image Database, interactive mode chooses the face feature point of input photo, and based on shape of face feature, mates best model；Secondly, realize photo by the method for triangle deformation and bilinear interpolation and map to the texture of threedimensional model, and introduce Alpha figure and realize the fusion transition to model neutrality texture of the face covering texture；Finally, the adjustment model that employing grid control method is secondary from entirety to level of detail is to reach to generate the purpose of sense of reality three-dimensional face.

1, interactive generation method step

Sense of reality three-dimensional face interactive generation method first has to, for Asia face feature, build Face Image Database；Then two dimension full face is mated with faceform, utilize texture to map and colour of skin blending algorithm, by the fine adjustments of Model Regulator, generate sense of reality threedimensional model.

1.1 model libraries build

More existing standardized three-dimensional face model storehouses, such as UND, BU-3DFE, BJUT etc. now.These model libraries are widely used in fields such as face generation and recognitions of face.Race, age, or sex, illumination etc. are respectively provided with different standards by different 3 d model libraries, and the inventive method, mainly for Asia ethnic group, sets up three-dimensional face storehouse targetedly.

Conventional Aisan's front shape of face classification has multiple standards: such as waveforms method, font method, Aisan's method etc..Shape of face for ease of model can be mated preferably with photo, model library according to the obvious degree of feature and conventional degree shape of face is divided into ellipse, del, elongated, square, circular 5 kinds.Generating with reference to the image studies tissue ICG of the Linkoping,Sweden university Nature face wire-frame model CANDIDE-3 developed of face neutral model^[8], its model simply and is gratuitously opened to the outside world use.But this model is only applicable to face, and this method, with reference to CANDIDE-3 Nature face model, has been especially tailored the head neutral model of applicable Chinese's shape of face feature.After model library construction completes, it is necessary to model is carried out pretreatment, including repairing leak, binding neutral skin tone texture, normalization size, feature point for calibration etc..

1.2 generation steps

Three-dimensional face mainly embodies its sense of reality from the facial characteristics of model and two aspects of texture, and user participates in the appointment of texture and the adjustment of model.Concrete grammar flow process is as it is shown in figure 1, be essentially sub-divided into following 5 steps:

Step1. characteristic point is demarcated: obtains front two-dimension human face photo, 7 parts such as eyes, nose, mouth is chosen 13 characteristic points and carries out labelling, it is determined that the approximate location of whole face face.

Step2. Model Matching: compare the characteristic point distance between photo and model, the model that matching similarity is maximum in model library.

Step3. texture maps: the texture mapping simulation of facial true colors according to face, it is achieved cover the texture mapping to neutral texture.Neutral texture is the texture of model acquiescence, and covering texture is human face photo.By the mapping of 13 characteristic points and peripheral region thereof in two layer texture, add the textural characteristics of target face to Matching Model.

Step4. the colour of skin merges: use an Alpha figure, and the boudary portion in face fusion is formed from covering texture seamlessly transitting to neutral texture.Meanwhile, utilize Alpha figure that the details such as eyebrow, bridge of the nose position is carried out trickle adjustment.

Step5. model regulates: by the method for vector differentials, the characteristics such as face size, position such as the face of Matching Model, eyes, shape of face are adjusted, and generates the personalized model meeting its facial characteristics.

1.3 Model Regulator designs

The effect of Model Regulator is that the layout to human face five-sense-organ carries out local directed complete set, to reach to be more nearly the purpose of real human face.Conventional 3 d modeling software, such as Maya, 3DSMax etc., is provided which the function of blendshapes (or claiming morph), for being converted between two or more states by model, forms different face layouts or human face expression.The present invention uses this technological thought, by constructing a Model Regulator that can be used for adjusting human face five-sense-organ slight change, generates personalized face true to nature.

If the summit V in face grid_i=(x, y, z)^TRegard Vector Groups H as₀=(V₁, V₂, V₃..., V_n), n represents the quantity on summit, and (as changed size) after eyes are converted, the position of some points changes, and forms Vector Groups H₁=(V₁', V₂', V₃' ... V_n'), this conversion is increased a weights W.So, from H₀To H₁Middle all of transition state H_xCan be transferred through linear interpolation to calculate.

H_x=H₀+W*(H₁-H₀)(1)

A kind of conversion is difficult to meet the demand of faceform, it is necessary to the feature according to face, selects multiple conversion, gives a weighted value W for each conversion, then the mixed formulation of m conversion is:

$H_x=H_0+{\Σ}_{i=1}^m{W}_i*(H_i-H_1)---\left(2\right)$

Wherein the quantity of m is as the criterion with the quantity needed in practical application.In an experiment, define the grid actuator of 23 kinds of different conversion, including: the covering of eye-level, eye distance, eyes size, eyebrow height, the eyebrow degree of depth, nose size, the nose degree of depth, nose height, the bridge of the nose degree of depth, nose bridge widths, nostril width, face height, face thickness, face width, lip, the cheekbone degree of depth, cheekbone width, cheekbone height, buccal width, chin height, the chin degree of depth, chin width, lower jaw width.Actuator realizes distortion of the mesh by control weight.

2, Realistic Face generates key technology

The key technology that Realistic Face generates mainly has the matching technique of two-dimension human face photo and threedimensional model, mapping technology and colour of skin integration technology.

2.1 characteristic points are demarcated and coupling

After obtaining two-dimension human face photo, it is necessary to the key position in photo is carried out characteristic point demarcation, to mate best faceform.Mpeg 4 standard defines 84 characteristic points on Nature face.With reference to mpeg 4 standard, need for faceform's coupling and texture mapping, for reducing the complexity of manual intervention, it is proposed on input photo and faceform, choose the characteristic point labelling as coupling and mapping of 13 key positions, as shown in Figure 2.Achievement in research according to three-dimensional face identification, the mutual relation between these key feature points can substantially determine position and the shape of face of human face five-sense-organ, thus matching optimal model.At experimental section by comparing with the matching process of different characteristic point quantity, demonstrate the effectiveness of 13 characteristic points.

The method of Model Matching is as follows: first, and photo is normalized, and calculates the distance between principal character point and proportionate relationship；Secondly, according to characteristic point distance and proportionate relationship, it is determined that the rough classification of shape of face；Finally, Euclidean distance between Matching Model and photo individual features point in the model library of corresponding shape of face, adopt formula (3) to search out closest threedimensional model.

$E={\mathrm{\Σ}}_{i=1}^n{(D_i-{D_i}^{,})}^2*{\mathrm{\λ}}_i---\left(3\right)$

Wherein, D_iFor distance between certain characteristic point in photo, D_i' for distance between individual features point in threedimensional model, λ_iFor the weighted value of each characteristic point distance, determine according to its percentage contribution size and empirical equation.N is the distance quantity compared.

2.2 textures map

The outer research worker of Present Domestic proposes multiple texture mapping method.Model after parametrization and corresponding texture according to the corresponding relation of characteristic point, are carried out triangulation, set up corresponding mapping relations by Kraevoy et al., and this algorithm needs to be based upon on the basis of a large amount of iteration, and complexity is higher.The present invention uses for reference the method for Kraevoy, and the methods such as triangle deformation, bilinearity difference that introduce realize covering texture and map to the characteristic point between neutral texture.

Assume that the characteristic point position of neutral texture and covering texture labelling respectively forms V respectively₁, V₂Two two-dimentional point sets.Due to the difference of manual mark, V₁, V₂In corresponding point necessarily will not be exactly accurate at same position, it is necessary to a kind of mapping method makes its natural fusion.Take P₁∈V₁, P₂∈V₂.By P₁, P₂Expand to three-dimensional vector, P₁=(X₁, Y₁, 1), P₂=(X₂, Y₂, 1), and assume there is matrix M, then can obtain:

$M=P_1*P_2^{-1}---\left(4\right)$

After obtaining the transformation matrix of each point, the conversion of characteristic point is applied in the pixel around characteristic point.This flow process uses common triangle mode of texturing that covering texture is applied to model texture.As it is shown on figure 3, characteristic point is connected, form the grid covering face, grid now has concurrently triangle and tetragon, by splitting, polygon is all reduced to triangle.It is in the pixel in the middle of triangle and is subject to the impact on Atria summit simultaneously.

Three summits assuming triangle are P₁, P₂, P₃.Using they three row as matrix, form 3*3 matrix (P₁, P₂, P₃).Assume that the triangle after conversion is (P₁', P₂', P₃'), then:

M*(P₁, P₂, P₃)=(P₁', P₂', P₃’)(5)

M=(P₁', P₂', P₃’)*(P₁, P₂, P₃)^-1(6)

Certain i point in the middle of diabolo, uses M*P_iJust can obtain P_i, determine the coordinate after mapping a little in triangle with this.For unmapped point, coordinate figure is all integer, but the coordinate after mapping is likely to not be integer, and causing cannot with the pixel one_to_one corresponding on figure.Now, utilize the bilinear interpolation method of sampling that the color on figure is sampled, good effect can be obtained.

2.3 colours of skin merge

After texture maps, due to the color distortion between neutral texture and covering texture, it is necessary to certain blending algorithm eliminates visual difference, makes the effect that the colour of skin merges more natural.In existing image processing software and projection splicing fusion software, conventional AlphaTransition method realizes seamlessly transitting of superimposed images, not only merges speed fast, and effect is better.Using this thought, the present invention is distributed according to faceform's feature and texture maps, makes the Alpha figure of standard, eliminates border sudden change by arranging the transition region at edge.The effect adopting Alpha figure has two: first, produces the Texture Boundaries seamlessly transitted；Secondly, to other position details of face, such as eyebrow, bridge of the nose part has been also carried out gradual change adjustment, to reduce the impact covering texture centering texture.Fig. 4 is 10 Alpha figure of Freehandhand-drawing, and in figure, the color of white portion takes from the color covering texture, and black part takes from the color of neutral texture.Part in the middle of white and black will be that power carries out linear interpolation and obtains by the pixel on Alpha figure, by coverage test, is numbered the Alpha figure of 10 and can realize the syncretizing effect of optimum.

In actual fused process, model neutrality texture, Alpha figure, covering texture will be processed into the same size.Pixel in Alpha figure participates in mixing as the interpolation weights of mixing, ultimately forms the texture of natural fusion.The mixed formulation of three figure is:

C (x, y)=C_base(x, y) * (255-C_Alpha(x, y))+C_overlay(x, y) * C_Alpha(x, y) wherein 0≤C_Alpha(x, y)≤255 (7)

C_base(x, y), C_Alpha(x, y), C_overlay(x y) represents neutral texture, Alpha figure respectively, covers texture in (x, y) color value at coordinate place.

The above, be only best mode for carrying out the invention, and any those familiar with the art is in the technical scope of present disclosure, and the simple change of the technical scheme that can become apparent to or equivalence are replaced and each fallen within protection scope of the present invention.

Claims (5)

1. the Realistic Human face generating method based on single photo, it is characterised in that comprise the following steps:

Step1. characteristic point is demarcated: obtains front two-dimension human face photo, 7 parts is chosen 13 characteristic points and carries out labelling, it is determined that the approximate location of whole face face；

Step2. Model Matching: compare the characteristic point distance between photo and model, the model that matching similarity is maximum in model library；

Step3. texture maps: the texture mapping simulation of facial true colors according to face, realize the mapping covering texture to neutral texture, neutral texture is the texture of model acquiescence, covering texture is human face photo, by the mapping of 13 characteristic points and peripheral region thereof in two layer texture, add the textural characteristics of target face to Matching Model；

Step4. the colour of skin merges: using an Alpha figure, the boudary portion in face fusion is formed from covering texture seamlessly transitting to neutral texture, meanwhile, utilizes Alpha figure that the details position of eyebrow, the bridge of the nose is carried out trickle adjustment；

Step5. model regulates: by the method for vector differentials, the characteristic of the face size of Matching Model, position is adjusted, and generates the personalized model meeting its facial characteristics.

2. the Realistic Human face generating method based on single photo according to claim 1, it is characterised in that construct a Model Regulator for adjusting human face five-sense-organ slight change in Step5, generate personalized face true to nature；

If the summit V in face grid_i=(x, y, z)^TRegard Vector Groups H as₀=(V₁, V₂, V₃..., V_n), n represents the quantity on summit, and after eyes are converted, the position of some points changes, and forms Vector Groups H₁=(V₁', V₂', V₃' ... V_n'), this conversion is increased a weights W, then, from H₀To H₁Middle all of transition state H_xCan be transferred through linear interpolation to calculate；

H_x=H₀+W*(H₁-H₀)(1)

A kind of conversion is difficult to meet the demand of faceform, it is necessary to the feature according to face, selects multiple conversion, gives a weighted value W for each conversion, then the mixed formulation of m conversion is:

$H_x=H_0+{\Σ}_{i=1}^m{W}_i*(H_i-H_1)---\left(2\right)$

Wherein the quantity of m is as the criterion with the quantity needed in practical application, in an experiment, define the grid actuator of 23 kinds of different conversion, including: the covering of eye-level, eye distance, eyes size, eyebrow height, the eyebrow degree of depth, nose size, the nose degree of depth, nose height, the bridge of the nose degree of depth, nose bridge widths, nostril width, face height, face thickness, face width, lip, the cheekbone degree of depth, cheekbone width, cheekbone height, buccal width, chin height, the chin degree of depth, chin width, lower jaw width, actuator realizes distortion of the mesh by control weight.

3. the Realistic Human face generating method based on single photo according to claim 1, it is characterised in that in Step2, the method for Model Matching is as follows: first, and photo is normalized, and calculates the distance between principal character point and proportionate relationship；Secondly, according to characteristic point distance and proportionate relationship, it is determined that the rough classification of shape of face；Finally, Euclidean distance between Matching Model and photo individual features point in the model library of corresponding shape of face, adopt formula (3) to search out closest threedimensional model；

$E={\Σ}_{i=1}^n{(D_i-{D_i}^{,})}^2*{\mathrm{\λ}}_i---\left(3\right)$

Wherein, D_iFor distance between certain characteristic point in photo, D_i' for distance between individual features point in threedimensional model, λ_iFor the weighted value of each characteristic point distance, determining according to its percentage contribution size and empirical equation, n is the distance quantity compared.

4. the Realistic Human face generating method based on single photo according to claim 1, it is characterised in that assume in Step3 that the characteristic point position of neutral texture and covering texture labelling respectively forms V respectively₁, V₂Two two-dimentional point sets, due to the difference of manual mark, V₁, V₂In corresponding point necessarily will not be exactly accurate at same position, it is necessary to a kind of mapping method makes its natural fusion, takes P₁∈V₁, P₂∈V₂, by P₁, P₂Expand to three-dimensional vector, P₁=(X₁, Y₁, 1), P₂=(X₂, Y₂, 1), and assume there is matrix M, then:

M=P₁*P₂ ^-1(4)

After obtaining the transformation matrix of each point, the conversion of characteristic point is applied in the pixel around characteristic point, this flow process uses common triangle mode of texturing that covering texture is applied to model texture, characteristic point is connected, form the grid covering face, grid now has concurrently triangle and tetragon, by splitting, polygon is all reduced to triangle, is in the pixel in the middle of triangle and is subject to the impact on Atria summit simultaneously；

Three summits assuming triangle are P₁, P₂, P₃, using they three row as matrix, form 3*3 matrix (P₁, P₂, P₃), it is assumed that the triangle after conversion is (P₁', P₂', P₃'), then:

M*(P₁, P₂, P₃)=(P₁', P₂', P₃’)(5)

M=(P₁', P₂', P₃’)*(P₁, P₂, P₃)^-1(6)

Certain i point in the middle of diabolo, uses M*P_iJust obtain P_i, determine the coordinate after mapping a little in triangle with this, for unmapped point, coordinate figure is all integer, but the coordinate after mapping is likely to not be integer, and causing cannot with the pixel one_to_one corresponding on figure.

5. the Realistic Human face generating method based on single photo according to claim 1, it is characterized in that, in actual fused process in Step4, model neutrality texture, Alpha figure, covering texture will be processed into the same size, pixel in Alpha figure participates in mixing as the interpolation weights of mixing, ultimately forming the texture of natural fusion, the mixed formulation of three figure is:

C (x, y)=C_base(x, y) * (255-C_Alpha(x, y))+C_overlay(x, y) * C_Alpha(x, y) wherein 0≤C_Alpha(x, y)≤255 (7)

C_base(x, y), C_Alpha(x, y), C_overlay(x y) represents neutral texture, Alpha figure respectively, covers texture in (x, y) color value at coordinate place.