CN109410119A - Mask image distortion method and its system - Google Patents
Mask image distortion method and its system Download PDFInfo
- Publication number
- CN109410119A CN109410119A CN201710710931.4A CN201710710931A CN109410119A CN 109410119 A CN109410119 A CN 109410119A CN 201710710931 A CN201710710931 A CN 201710710931A CN 109410119 A CN109410119 A CN 109410119A
- Authority
- CN
- China
- Prior art keywords
- characteristic point
- image
- triangle gridding
- mask image
- mask
- 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 32
- 230000001815 facial effect Effects 0.000 claims abstract description 52
- 210000000214 mouth Anatomy 0.000 claims description 16
- 210000001508 eye Anatomy 0.000 claims description 15
- 210000004709 eyebrow Anatomy 0.000 claims description 15
- 210000001061 forehead Anatomy 0.000 claims description 15
- 238000013507 mapping Methods 0.000 claims description 15
- 210000001331 nose Anatomy 0.000 claims description 13
- 230000009466 transformation Effects 0.000 claims description 11
- 230000003190 augmentative effect Effects 0.000 claims description 10
- 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 8
- 238000012545 processing Methods 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 3
- 210000005252 bulbus oculi Anatomy 0.000 claims 2
- 239000011159 matrix material Substances 0.000 description 10
- 238000010586 diagram Methods 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 241000282693 Cercopithecidae Species 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 241000283973 Oryctolagus cuniculus Species 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000029052 metamorphosis Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000036544 posture Effects 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 241001465754 Metazoa Species 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000002939 deleterious effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000000887 face Anatomy 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 210000004209 hair Anatomy 0.000 description 1
- 210000003128 head Anatomy 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000011017 operating method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000001747 pupil Anatomy 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/18—Image warping, e.g. rearranging pixels individually
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Processing Or Creating Images (AREA)
- Image Processing (AREA)
Abstract
The present invention discloses a kind of mask image distortion method and its system, wherein determining multiple characteristic points according to the mask image this method comprises: obtain mask image；Triangle gridding concordance list is generated according to the multiple characteristic point, wherein includes multiple triangle griddings in the triangle gridding concordance list；Facial image is obtained in real time by camera, determines location information of the multiple characteristic point on the facial image；According to the triangle gridding concordance list, multiple characteristic points of the facial image are divided into triangle gridding；The triangle gridding of the mask image is deformed to the triangle gridding of the facial image, the mask image and the facial image are overlapped.It is fused together with capable of making face and mask harmony and natural through the invention.
Description
Technical field
The present invention relates to face image processing technology more particularly to a kind of mask image distortion method and its systems.
Background technique
Face recognition technology refers to and identifies face using the computer technology that analysis is compared.All the time, recognition of face is all
It is a popular computer technology research field, there is high research value and extensive market application prospect.One
As, face identification system includes: Face datection, face alignment, face characteristic extraction, face alignment etc..
Currently, using morphing can many special efficacys, be widely used in film, TV and advertising media
In.Have the method for many Facial metamorphosis at present, for example by human face five-sense-organ, face mask local directed complete set, makes its beautifulization, or
It is analogous to distorting mirror and face is distorted into amplification exaggerationization.But for by Facial metamorphosis be specific mask template at present still
It does not put forward effective solutions.
Summary of the invention
The main purpose of the present invention is to provide a kind of mask image distortion method and its systems, are deposited with solving the prior art
The above problem, in which:
A kind of mask image distortion method is proposed according to embodiments of the present invention comprising:
Mask image is obtained, multiple characteristic points are determined according to the mask image；
Triangle gridding concordance list is generated according to the multiple characteristic point, wherein includes more in the triangle gridding concordance list
A triangle gridding；
Facial image is obtained in real time by camera, determines position letter of the multiple characteristic point on the facial image
Breath；
According to the triangle gridding concordance list, multiple characteristic points of the facial image are divided into triangle gridding；
The triangle gridding of the mask image is deformed to the triangle gridding of the facial image, by the mask figure
Picture and the facial image are overlapped.
A kind of mask image deformation system is also proposed according to embodiments of the present invention comprising:
Mask image collection module determines multiple characteristic points according to the mask image for obtaining mask image；
Triangle gridding concordance list generation module, for generating triangle gridding concordance list according to the multiple characteristic point, wherein
It include multiple triangle griddings in the triangle gridding concordance list；
Characteristic point determining module determines the multiple characteristic point in institute for obtaining facial image in real time by camera
State the location information on facial image；
Triangulation module is used for according to the triangle gridding concordance list, by multiple features of the facial image
Point divides triangle gridding；
Anamorphose module, for carrying out the triangle gridding of the mask image to the triangle gridding of the facial image
Deformation, the mask image and the facial image are overlapped.
According to the technique and scheme of the present invention, fixed triangle gridding concordance list is generated by the mask image obtained in advance,
Triangle gridding is divided in the facial image that captures in real time according to the triangle gridding concordance list, and by the texture of mask image to people
Correspondence tri patch in face image is deformed, and through the invention can be passed through mask image real-time and accurately tight after deformation
It is attached on tracked face, is fused together with making its face and mask harmony and natural.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes part of this application, this hair
Bright illustrative embodiments and their description are used to explain the present invention, and are not constituted improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is the flow chart of mask image distortion method according to an embodiment of the present invention；
Fig. 2 is the schematic diagram of the characteristic point distribution of mask image according to an embodiment of the present invention；
Fig. 3 is the schematic diagram of CR_Spline curve according to an embodiment of the present invention；
Fig. 4 A and Fig. 4 B are the schematic diagram of monkey Prototype drawing and interpolation result according to an embodiment of the present invention respectively；
Fig. 5 is the schematic diagram of triangle gridding piece according to an embodiment of the present invention；
Fig. 6 is the schematic diagram of the mapping relations of triangle according to an embodiment of the present invention；
Fig. 7 is the schematic diagram of the fusion results of mask image according to an embodiment of the present invention；
Fig. 8 is the structural block diagram of mask image deformation system according to an embodiment of the present invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with the specific embodiment of the invention and
Technical solution of the present invention is clearly and completely described in corresponding attached drawing.Obviously, described embodiment is only the present invention one
Section Example, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not doing
Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
The technical solution provided below in conjunction with attached drawing, each embodiment that the present invention will be described in detail.
A kind of mask image distortion method is provided according to embodiments of the present invention, and the method is applied to camera
Terminal device.Wherein, can be for example with the terminal device of camera: mobile phone, tablet computer, electronic reader, remote controler,
The tool such as personal computer (Personal Computer, PC), laptop, mobile unit, Web TV, wearable device
There is the intelligent terminal of network function.
Fig. 1 is the flow chart of mask image distortion method according to an embodiment of the present invention, as shown in Figure 1, this method includes
Following steps:
Step S102 obtains mask image, determines multiple characteristic points according to the mask image.
In the present invention, the mask image can be the cartoons faces such as the zodiac animal, such as monkey, rabbit of cartoon type
Tool.Cartoon mask is put on to face, needed first to do artificial described point to the image of mask image template, determined the mask image
Multiple characteristic points.The characteristic point of mask image includes multiple essential characteristic points and is expanded according to the multiple essential characteristic point interpolation
The multiple augmented features points filled；Wherein, essential characteristic point has 75, in order to make deformed mask keep the nature of texture
Property, avoids being distorted or fold, needs to expand this 75 essential characteristic points, interpolation go out more augmented features points with
Enrich each region of mask.
Wherein, face mask in mask image described in the multiple essential characteristic point location, eyebrow, nose, eyes, mouth,
The location information of forehead, specifically:
The location information of 17 essential characteristic point location face masks；
The location information of 10 essential characteristic point location eyebrows；
The location information of 9 essential characteristic point location noses；
The location information of 12 essential characteristic point location eyes；
The location information of 20 essential characteristic point location mouths；
The location information of 7 essential characteristic point location foreheads.
Fig. 2 shows the distribution situation of mask image characteristic point, digital representation essential characteristic point therein and augmented features
The identification information of point.
Step S104 generates triangle gridding concordance list according to the multiple characteristic point, wherein the triangle gridding concordance list
In include multiple triangle griddings.
By essential characteristic point shown in Fig. 2 and augmented features point, according to delaunay triangulation, (triangle of point set is cutd open
Point) method composition triangular mesh.Due to the mouth that mainly show real human face of the mask image design in the present invention, and
The variation of mouth is uncertain (such as upper lower lip intersection etc.), this is all easy to cause the folding of triangle, reversion, so of the invention
All triangle griddings inside mouth are screened, the triangle about interior mouth is all deleted together, and suitably reconfigures
Some triangle griddings of outer mouth；In addition, by other parts, those are easy to appear fold since face is the changeable state of posture
Triangle gridding also carries out suitably deleting recombination, thus can get fixed triangle gridding concordance list.In face tracking below,
Mask framing to essential characteristic point and expand augmented features point out, be all to be drawn according to this grid index table
Point.
235 in upper table indicate the number of triangles, " { " following digital representation essential characteristic point and augmented features point
Identification information, wherein every a line all indicates a triangle, if the triangle of the first row is by 133,93,36 these three numbers
Feature point group at.That is, having recorded the characteristic point table by triangle gridding in the triangle gridding concordance list of the embodiment of the present invention
The information of the multiple triangle griddings shown.
Step S106 obtains facial image by camera in real time, determines the multiple characteristic point in the facial image
On location information.
The positioning of characteristic point is exactly face alignment in fact, this is the previous very hot research direction of mesh, the energy on face
Such as made up, worn glasses, U.S. pupil, skin makeup and sex, race differentiate etc..The purpose of facial modeling is in people
On the basis of face detection, the position of facial feature points (eyes, eyebrow, nose, mouth, face's outer profile) is further determined that.But
Due to different postures, expression, illumination and the influence for the factors such as blocking, it is accurately located out each characteristic point but and does not allow
Easily.The basic ideas of location algorithm are: the position constraint between the textural characteristics of face and each characteristic point combines.Classic algorithm
There are active shape model (ASM), active apparent model (AAM), supervision gradient former (SDM) etc..
If Fig. 2 illustrates the distribution situation of key point, description face mask is gone with the point of number 0 to 16, is arrived with number 17
26 point removes description left and right eyebrow, describes nose with the point of number 27 to 35, describes left and right eye, mouth with the point of number 36 to 47
Description include Wai Zui and interior mouth, the point of number 48 to 59 and the point of number 60 to 67 are separately included, with the point of number 68 to 75
Forehead is described.
In order to reduce error, the accuracy of deformation is improved, for mask template image, by mask by size in proportion
It is attached to one and 75 characteristic points of upper figure is found out by manual described point according to the layout of human face referring on face, remember
Record their coordinate position.For being tracked for person, that just needs a face alignment algorithm that real-time calibration is gone to go out these spies
Point is levied, the embodiment of the present invention can use SDM method, its theory is simple, and calculation amount is small, and stability is also good.
The core content of SDM face alignment is exactly the mapping of feature to offset:
Wherein, Δ x indicates offset, and R indicates mapping matrix,Indicate feature vector.SDM method is to pass through training sample
It goes to obtain mapping matrix R, the specific steps are as follows:
(1) sample is normalized, keeps the scale of sample unified；
(2) mean value face is calculated；
(3) it by mean value face, is placed on sample as estimation face, makes mean value center and original face shape center pair
Together；
(4) feature of the mark point based on each mean value face, such as hog are calculated；
(5) together by the feature string of all characteristic points, sample characteristics matrix is formed
(6) offset between estimation face and real human face is calculated, and forms matrix Δ x；
(7) linear equation is solved.
Estimation shape is exactly as a result, in a practical situation, needing to train multilayer, needing after two layers plus prediction drift amount
Use the result of upper one layer of alignment as estimation shape.After obtaining mapping matrix R by training sample, in testing by following
Formula finds out each layer of face shape:
Wherein, xkIndicate the face shape of kth layer, Rk-1Indicate -1 layer of kth of mapping matrix,Indicate -1 layer of kth
Feature vector.Here the number of plies is consistent with the number of plies of training stage.
After the alignment of SDM face, accurate 68 eyebrows facial key point below can be obtained, but since mask is
Aughty lovely and beautification the effect for preferably highlighting the Chinese zodiac, much includes forehead area.In addition Chinese zodiac dragon, rabbit
The ear of equal mask templates devises some sequence movements for them, has also required to calculate forehead feature so all on head
The necessity of point.7 points in the present invention on forehead are found out according to existing 68 coordinates.Specific calculating process is such as
Under:
(1) left eye central point lefEye is calculated by number 36 to 41, right eye central point is calculated by the point of number 42 to 47
RightEye, then eyes central point eyeCen is calculated by them；
(2) by meeting face golden ratio relationship: the distance of place between the eyebrows to forehead is approximately equal to place between the eyebrows to the distance of nose bottom,
I.e. | eyeCen-P33|=| eyeCen-P71|, the point position of number 71 can be calculated；
(3) P is calculated separately71To P0With P16Between transverse direction, fore-and-aft distance, and be divided into quarter respectively, then can calculate
The coordinate of other 6 points (i.e. 68-70,72-74).
Again since mask image has abundant texture, if the affine change of grid is done with the 75 above control points
It changes, that deformed texture is extremely difficult to fine and smooth, smooth degree.Therefore, it is necessary to do interpolation expansion with existing 75 points.Face
Have image difference, it is desirable that the position of interpolation and expand out control points also require it is different.For example, Catmull-Rom sample
Bar line, is a more special Bezier curve, and this article of Bezier curve can guarantee that it can be across from the of control point
Two points are to all the points between the point second from the bottom at control point.Thus, Catmull-Rom line transect at least needs 4 controls
System point is to be controlled.
Following formula provides the point how calculated between certain two specified point:
OutP=P0*(-0.5*t3+t2-0.5*t)+P1*(1.5*t3-2.5*t2+1.0)+P2*(-1.5*t3+2.0*t2+
0.5*t)+P3*(0.5*t3-0.5*t2)；
Here P0, P1, P2, P3All it is the point of curve, as shown in figure 3, can calculate from point P1To P2Between point.
T value range in formula is [0,1], and when t linear change from 0 to 1, curve will be from point P1(t at this time
=0), slowly it is moved to P2(t=1 at this time).Another feature of this curve be exactly the point P calculated tangent vector and this
The tangent vector of two beginning and ends is parallel around a point.
In Fig. 3 although control point has 4, but the curve drawn but is merely able to through two intermediate points.If it is desired to
Curve crosses this four points simultaneously, and the method for that processing is to create a ground zero and new terminal, to constitute four control points.
Such as now with P0, P1, P2, P3If with [P0, P1, P2, P3] curve construction, curve will be merely able to pass through P1And P2, in order to allow
Curve can pass through P0And P3, can taking human as construct following control point [2P0-P1, P0, P1, P2], and [P1, P2, P3,
2P3-P2].By such method, it will be able to draw a curve by all control points.
By above method, the control point of extendible monkey mask counts to 141, wherein eyes be closed area, every 2
Control point interpolation goes out 1 point, and extendible 12 points, spline curve out can pass through entire point well；Because here forehead with
Face profile separates, so eyebrow, face profile, mouth and nose etc. require manually to make ground zero and terminal.
Multiple characteristic points of the facial image are divided the triangulation network according to the triangle gridding concordance list by step S108
Lattice.
It will include multiple essential characteristic points and multiple augmented features point in facial image, according to the triangle gridding
Concordance list divides triangle gridding.As shown in Figure 4 B, the point that "+" indicates is previously described 75 characteristic points, and the point that " * " is indicated is
Expand point out by Catmull-Rom interpolation.The point that interpolation is expanded out can group by searching for fixed mesh concordance list
Synthesize fixed triangle number, as shown in Figure 5 in include 235 triangle griddings.
Step S110 deforms the mask image to the facial image according to the location information of characteristic point, will
The mask image and the facial image are overlapped.
Distortion of the mesh algorithm is very simple and quick, but when deformation extent is slightly larger, the straight flange of triangle gridding can be to deformation
Effect causes deleterious effect, than distorting as previously described.Therefore, before distortion of the mesh, need to be added a thin plate
Batten process is mapped characteristic point using thin plate spline, smoothed out characteristic point, according still further to fixed triangle grid index table, group
At suitable triangle gridding, correspond with mask template grid affine, can make that deformed template is finer and smoother, forces
Very, naturally, make synthesis after mask more enrich it is interesting.
Thin plate spline mapping, is that a warping function is determined according to the Control point collection in two width associated pictures.It
Find the smallest smooth surface of rich degree by all set points.Its specific principle is that selection two images (set image A and figure
As two set of respective coordinates information in B), the corresponding position being mapped to from the Pixel Information in image A in image B.It is thin
Plate batten is accurately to match corresponding coordinate information, and by anamorphose energy limit in minimum.Image mapping function
It is set asThe coefficient of mapping transformation is determined by selected registering control points number.
Wherein P=(x, y) is image space coordinate, and the matrix W of n+3 × 2 is defined as follows:
W=(ω1ω2…ωnα1αxαy)T=L-1M
Wherein ω1,ω2,…ωn,α1,αxAnd αyIt is the key coefficient in thin plate spline transformation.
Wherein hi=(Xi,Yi), i=1,2 ..., n is the control point coordinates in image B.Matrix L is by matrix K, and Q and one 3
× 3 null matrix composition.
Wherein U (r)=r2Logr, and rij=(| pi-pj|) it is the distance between any two control point of image A.In Q
(xi,yi), i=1 ..., n indicates the control point coordinates of image A.
Key coefficient W can be calculated by above formula, then by the coordinate of any point P in image A and point P and its
The distance of its each point | P-Pj|, then the coordinate of corresponding points in image B can be obtained.
Mask template will be deformed to the face of current pose, in fact be exactly affine transformation process.Affine transformation refers to one
Vector space carries out once linear transformation and connects a translation, is transformed to the process of another vector space.One arbitrary
Affine transformation generally includes rotation (linear transformation), translation and scaling, and what is typically represented is a kind of mapping pass between two width figures
System.In the present invention, the mask and the facial image selected is divided into equal number of triangles, their deformation is exactly three
An angular shape transformation be another shape, by it includes texture do one-to-one mapping, as shown in Figure 6.
Assuming that tri- coordinates of image1 are (x '1,y′1)、(x′2,y′2)、(x′3,y′3), three coordinates in image2
(x1,y1)、(x2,y2) and (x3,y3), then can find out in image2 any point (x, y) in image2 corresponding position (x ',
Y '):
Alpha=-y*x3+y3*x-x*y2+x2*y-x2*y3+x3*y2；
Beta=y*x3-x1*y-x3*y1-y3*x+x1*y3+x*y1；
Gamma=x*y2-x*y1-x1*y2-x2*y+x2*y1+x1*y；
X '=alpha*x '1+beta*x′2+gamma*x′3；
Y '=alpha*y '1+beta*y′2+gamma*y′3；
Point (x ', y ') pixel can be gone out in image1 interpolation as a result, and be assigned to the midpoint image2 (x, y), so i.e.
All pixels in fillable full image2.And mask and facial image be all numerous triangle sets at, thus can in face phase
The part answered realizes the fusion of mask image, as shown in Figure 7 by mask Texture mapping.
According to embodiments of the present invention, a kind of mask image deformation system is also provided, the system can be to be mounted on terminal
Application program (App) in equipment, the terminal device includes but is not limited to: mobile phone or smart phone, smart television, plate
The terminal devices such as computer (PAD), desktop computer, laptop.The terminal device may include: at least one processor, penetrate
Frequency module, antenna, memory, input/output module (including audio input output module, keyboard input module and display
Deng) and user interface.In some embodiments of the invention, these components can be connected by bus or other way.Wherein,
Radio-frequency module also includes at least radio-frequency front-end (RFFE) and RF IC (RFIC).
With reference to Fig. 8, the mask image deformation system includes:
Mask image collection module 81 determines multiple characteristic points according to the mask image for obtaining mask image；
Triangle gridding concordance list generation module 82, for generating triangle gridding concordance list according to the multiple characteristic point,
In, it include multiple triangle griddings in the triangle gridding concordance list；
Characteristic point determining module 83 is determined for obtaining facial image in real time by the camera of the intelligent terminals such as mobile phone
Location information of the multiple characteristic point on the facial image；
Triangulation module 84 is used for according to the triangle gridding concordance list, by multiple spies of the facial image
Sign point divides triangle gridding；
Anamorphose module 85, for by triangle gridding from the triangle gridding of the mask image to the facial image into
Row deformation, the mask image and the facial image are overlapped.
Wherein, the multiple characteristic point that the mask image collection module 81 determines include: multiple essential characteristic points with
And the multiple augmented features points expanded according to the multiple essential characteristic point interpolation；Wherein, the multiple essential characteristic point
Position face mask in the mask image, eyebrow, nose, eyes, mouth, forehead location information.
Further, the essential characteristic point has 75, using the location information of 17 essential characteristic point location face masks,
The location information of 10 essential characteristic point location eyebrows, the location information of 9 essential characteristic point location noses, 12 essential characteristics
The position of the location information of point location eyes, the location information of 20 essential characteristic point location mouths, 7 essential characteristic point location foreheads
Information.
In the present invention, the triangulation module 84 is also used to, and multiple features of the facial image are clicked through
The mapping of row thin plate spline, carries out characteristic point smoothing processing；Smoothed out characteristic point generates more according to the triangle gridding concordance list
A triangle gridding.
In the present invention, described image deformation module 85 is also used to, and the mask image and the facial image are divided
For the triangle gridding of equal amount, by triangle gridding from the triangle gridding of the mask image to the corresponding facial image into
Row affine transformation.
The operating procedure of method of the invention and the structure feature of system are corresponding, can be cross-referenced, no longer repeat one by one.
It will be understood by those skilled in the art that the embodiment of the present invention can provide as method, system or computer program product.
Therefore, complete hardware embodiment, complete software embodiment or embodiment combining software and hardware aspects can be used in the present invention
Form.It is deposited moreover, the present invention can be used to can be used in the computer that one or more wherein includes computer usable program code
The shape for the computer program product implemented on storage media (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
Formula.
The above description is only an embodiment of the present invention, is not intended to restrict the invention, for those skilled in the art
For member, the invention may be variously modified and varied.All within the spirits and principles of the present invention, it is made it is any modification,
Equivalent replacement, improvement etc., should be included within scope of the presently claimed invention.
Claims (10)
1. a kind of mask image distortion method characterized by comprising
Mask image is obtained, multiple characteristic points are determined according to the mask image；
Triangle gridding concordance list is generated according to the multiple characteristic point, wherein includes multiple three in the triangle gridding concordance list
Angle grid；
Facial image is obtained in real time by camera, determines location information of the multiple characteristic point on the facial image；
According to the triangle gridding concordance list, multiple characteristic points of the facial image are divided into triangle gridding；
The triangle gridding of the mask image is deformed to the triangle gridding of the facial image, by the mask image and
The facial image is overlapped.
2. the method according to claim 1, wherein further include:
The multiple characteristic point includes multiple essential characteristic points and is expanded according to the multiple essential characteristic point interpolation
Multiple augmented features points；Wherein, face mask, eyebrow, nose, eye in mask image described in the multiple essential characteristic point location
Eyeball, mouth, forehead location information.
3. according to the method described in claim 2, it is characterized in that, in mask image described in the multiple essential characteristic point location
Face mask, eyebrow, nose, eyes, mouth, forehead location information, comprising:
The location information of 17 essential characteristic point location face masks；
The location information of 10 essential characteristic point location eyebrows；
The location information of 9 essential characteristic point location noses；
The location information of 12 essential characteristic point location eyes；
The location information of 20 essential characteristic point location mouths；
The location information of 7 essential characteristic point location foreheads.
4. the method according to claim 1, wherein multiple characteristic points by the facial image divide three
Angle grid, comprising:
Multiple characteristic points of the facial image are subjected to thin plate spline mapping, carry out characteristic point smoothing processing；
Smoothed out characteristic point generates multiple triangle griddings according to the triangle gridding concordance list.
5. according to the method described in claim 4, it is characterized in that, the triangle gridding by the mask image is to the people
The triangle gridding of face image is deformed, comprising:
The mask image and the facial image are divided into the triangle gridding of equal amount, by the triangle of the mask image
Grid carries out affine transformation to the triangle gridding of the corresponding facial image.
6. a kind of mask image deformation system characterized by comprising
Mask image collection module determines multiple characteristic points according to the mask image for obtaining mask image；
Triangle gridding concordance list generation module, for generating triangle gridding concordance list according to the multiple characteristic point, wherein described
It include multiple triangle griddings in triangle gridding concordance list；
Characteristic point determining module determines the multiple characteristic point in the people for obtaining facial image in real time by camera
Location information in face image；
Triangulation module is used for according to the triangle gridding concordance list, by multiple feature click and sweep of the facial image
Divide triangle gridding；
Anamorphose module, for becoming the triangle gridding of the mask image to the triangle gridding of the facial image
The mask image and the facial image are overlapped by shape.
7. system according to claim 6, which is characterized in that the mask image collection module is also used to:
The multiple characteristic point includes multiple essential characteristic points and is expanded according to the multiple essential characteristic point interpolation
Multiple augmented features points；Wherein, face mask, eyebrow, nose, eye in mask image described in the multiple essential characteristic point location
Eyeball, mouth, forehead location information.
8. system according to claim 7, which is characterized in that mask image collection module uses:
The location information of 17 essential characteristic point location face masks；
The location information of 10 essential characteristic point location eyebrows；
The location information of 9 essential characteristic point location noses；
The location information of 12 essential characteristic point location eyes；
The location information of 20 essential characteristic point location mouths；
The location information of 7 essential characteristic point location foreheads.
9. system according to claim 6, which is characterized in that the triangulation module is also used to, by the people
Multiple characteristic points of face image carry out thin plate spline mapping, carry out characteristic point smoothing processing；Smoothed out characteristic point is according to
Triangle gridding concordance list generates multiple triangle griddings.
10. system according to claim 9, which is characterized in that described image deformation module is also used to, by the mask figure
Picture and the facial image are divided into the triangle gridding of equal amount, by the triangle gridding of the mask image to corresponding described
The triangle gridding of facial image carries out affine transformation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710710931.4A CN109410119A (en) | 2017-08-18 | 2017-08-18 | Mask image distortion method and its system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710710931.4A CN109410119A (en) | 2017-08-18 | 2017-08-18 | Mask image distortion method and its system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109410119A true CN109410119A (en) | 2019-03-01 |
Family
ID=65455256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710710931.4A Pending CN109410119A (en) | 2017-08-18 | 2017-08-18 | Mask image distortion method and its system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109410119A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110263671A (en) * | 2019-05-30 | 2019-09-20 | 量子动力（深圳）计算机科技有限公司 | It is a kind of can Fast Calibration muscular features mask and its processing method |
| CN110728271A (en) * | 2019-12-19 | 2020-01-24 | 恒信东方文化股份有限公司 | Method for generating human expression aiming at face recognition |
| CN111031234A (en) * | 2019-11-20 | 2020-04-17 | 维沃移动通信有限公司 | Image processing method and electronic equipment |
| CN111047511A (en) * | 2019-12-31 | 2020-04-21 | 维沃移动通信有限公司 | Image processing method and electronic equipment |
| CN111652022A (en) * | 2019-06-26 | 2020-09-11 | 广州虎牙科技有限公司 | Image data display method, image data live broadcast device, electronic equipment and storage medium |
| CN111767876A (en) * | 2020-07-02 | 2020-10-13 | 北京爱笔科技有限公司 | Method and device for generating face image with shielding |
| CN112099712A (en) * | 2020-09-17 | 2020-12-18 | 北京字节跳动网络技术有限公司 | Face image display method and device, electronic equipment and storage medium |
| JP7214926B1 (en) | 2019-12-18 | 2023-01-30 | 北京字節跳動網絡技術有限公司 | Image processing method, device, electronic device and computer-readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN106709931A (en) * | 2015-07-30 | 2017-05-24 | 中国艺术科技研究所 | Method of mapping facial makeup to face, and facial makeup mapping device |
-
2017
- 2017-08-18 CN CN201710710931.4A patent/CN109410119A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN106709931A (en) * | 2015-07-30 | 2017-05-24 | 中国艺术科技研究所 | Method of mapping facial makeup to face, and facial makeup mapping device |
Non-Patent Citations (1)
| Title |
|---|
| 原娜: "视频中人脸的京剧脸谱映射", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110263671B (en) * | 2019-05-30 | 2023-03-31 | 量子动力（深圳）计算机科技有限公司 | Mask capable of quickly calibrating muscle characteristics and processing method thereof |
| CN110263671A (en) * | 2019-05-30 | 2019-09-20 | 量子动力（深圳）计算机科技有限公司 | It is a kind of can Fast Calibration muscular features mask and its processing method |
| CN111652022A (en) * | 2019-06-26 | 2020-09-11 | 广州虎牙科技有限公司 | Image data display method, image data live broadcast device, electronic equipment and storage medium |
| CN111652022B (en) * | 2019-06-26 | 2023-09-05 | 广州虎牙科技有限公司 | Image data display method, image data live broadcast device, electronic equipment and storage medium |
| CN111031234B (en) * | 2019-11-20 | 2021-09-03 | 维沃移动通信有限公司 | Image processing method and electronic equipment |
| CN111031234A (en) * | 2019-11-20 | 2020-04-17 | 维沃移动通信有限公司 | Image processing method and electronic equipment |
| JP7214926B1 (en) | 2019-12-18 | 2023-01-30 | 北京字節跳動網絡技術有限公司 | Image processing method, device, electronic device and computer-readable storage medium |
| JP2023504926A (en) * | 2019-12-18 | 2023-02-07 | 北京字節跳動網絡技術有限公司 | Image processing method, device, electronic device and computer-readable storage medium |
| US11651529B2 (en) | 2019-12-18 | 2023-05-16 | Beijing Bytedance Network Technology Co., Ltd. | Image processing method, apparatus, electronic device and computer readable storage medium |
| CN110728271B (en) * | 2019-12-19 | 2020-04-07 | 恒信东方文化股份有限公司 | Method for generating human expression aiming at face recognition |
| CN110728271A (en) * | 2019-12-19 | 2020-01-24 | 恒信东方文化股份有限公司 | Method for generating human expression aiming at face recognition |
| CN111047511A (en) * | 2019-12-31 | 2020-04-21 | 维沃移动通信有限公司 | Image processing method and electronic equipment |
| CN111767876A (en) * | 2020-07-02 | 2020-10-13 | 北京爱笔科技有限公司 | Method and device for generating face image with shielding |
| CN112099712A (en) * | 2020-09-17 | 2020-12-18 | 北京字节跳动网络技术有限公司 | Face image display method and device, electronic equipment and storage medium |
| CN112099712B (en) * | 2020-09-17 | 2022-06-07 | 北京字节跳动网络技术有限公司 | Face image display method and device, electronic equipment and storage medium |
| US11935176B2 (en) | 2020-09-17 | 2024-03-19 | Beijing Bytedance Network Technology Co., Ltd. | Face image displaying method and apparatus, electronic device, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109410119A (en) | Mask image distortion method and its system | |
| CN107610209A (en) | Human face countenance synthesis method, device, storage medium and computer equipment | |
| CN109978930A (en) | A kind of stylized human face three-dimensional model automatic generation method based on single image | |
| CN110399809A (en) | The face critical point detection method and device of multiple features fusion | |
| CN110738161A (en) | face image correction method based on improved generation type confrontation network | |
| CN102567716B (en) | Face synthetic system and implementation method | |
| CN109801380A (en) | A kind of method, apparatus of virtual fitting, storage medium and computer equipment | |
| US11562536B2 (en) | Methods and systems for personalized 3D head model deformation | |
| CN108363973A (en) | A kind of unconfined 3D expressions moving method | |
| CN112348937A (en) | Face image processing method and electronic equipment | |
| US11587288B2 (en) | Methods and systems for constructing facial position map | |
| JP2014211719A (en) | Apparatus and method for information processing | |
| CN110264396A (en) | Video human face replacement method, system and computer readable storage medium | |
| JP7462120B2 (en) | Method, system and computer program for extracting color from two-dimensional (2D) facial images | |
| WO2023066120A1 (en) | Image processing method and apparatus, electronic device, and storage medium | |
| CN110288715A (en) | Virtual necklace try-in method, device, electronic equipment and storage medium | |
| CN111815768B (en) | Three-dimensional face reconstruction method and device | |
| WO2022267653A1 (en) | Image processing method, electronic device, and computer readable storage medium | |
| JP2024506170A (en) | Methods, electronic devices, and programs for forming personalized 3D head and face models | |
| CN110717978A (en) | Three-dimensional head reconstruction method based on single image | |
| CN111652795A (en) | Face shape adjusting method, face shape adjusting device, live broadcast method, live broadcast device, electronic equipment and storage medium | |
| CN115546361A (en) | Three-dimensional cartoon image processing method and device, computer equipment and storage medium | |
| CN111275610A (en) | Method and system for processing face aging image | |
| CN111652807B (en) | Eye adjusting and live broadcasting method and device, electronic equipment and storage medium | |
| CN115631516A (en) | Face image processing method, device and equipment and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190301 |
|
| WD01 | Invention patent application deemed withdrawn after publication |