CN109087240A - Image processing method, image processing apparatus and storage medium - Google Patents
Image processing method, image processing apparatus and storage medium Download PDFInfo
- Publication number
- CN109087240A CN109087240A CN201810955579.5A CN201810955579A CN109087240A CN 109087240 A CN109087240 A CN 109087240A CN 201810955579 A CN201810955579 A CN 201810955579A CN 109087240 A CN109087240 A CN 109087240A
- Authority
- CN
- China
- Prior art keywords
- face
- information
- image
- image processing
- processing method
- 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.)
- Granted
Links
- 238000012545 processing Methods 0.000 title claims abstract description 71
- 238000003672 processing method Methods 0.000 title claims abstract description 71
- 238000000034 method Methods 0.000 claims abstract description 50
- 230000008569 process Effects 0.000 claims abstract description 32
- 238000010606 normalization Methods 0.000 claims description 25
- 238000004422 calculation algorithm Methods 0.000 claims description 21
- 238000004590 computer program Methods 0.000 claims description 11
- 241000208340 Araliaceae Species 0.000 claims description 4
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 4
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 4
- 235000008434 ginseng Nutrition 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 20
- 238000004891 communication Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 9
- 210000000887 face Anatomy 0.000 description 6
- 230000001815 facial effect Effects 0.000 description 6
- 238000001514 detection method Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 210000001508 eye Anatomy 0.000 description 3
- 210000004709 eyebrow Anatomy 0.000 description 3
- 230000029052 metamorphosis Effects 0.000 description 3
- 208000002874 Acne Vulgaris Diseases 0.000 description 2
- 241001396014 Priacanthus arenatus Species 0.000 description 2
- 229930003448 Vitamin K Natural products 0.000 description 2
- 206010000496 acne Diseases 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 210000003128 head Anatomy 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- SHUZOJHMOBOZST-UHFFFAOYSA-N phylloquinone Natural products CC(C)CCCCC(C)CCC(C)CCCC(=CCC1=C(C)C(=O)c2ccccc2C1=O)C SHUZOJHMOBOZST-UHFFFAOYSA-N 0.000 description 2
- 235000019168 vitamin K Nutrition 0.000 description 2
- 239000011712 vitamin K Substances 0.000 description 2
- 150000003721 vitamin K derivatives Chemical class 0.000 description 2
- 229940046010 vitamin k Drugs 0.000 description 2
- 230000002087 whitening effect Effects 0.000 description 2
- 210000000216 zygoma Anatomy 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000712 assembly Effects 0.000 description 1
- 238000000429 assembly Methods 0.000 description 1
- 210000005252 bulbus oculi Anatomy 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 210000004209 hair Anatomy 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000007493 shaping process Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Classifications
-
- G06T3/04—
-
- G06T5/80—
-
- G06T5/90—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
Abstract
A kind of image processing method, image processing apparatus and storage medium, the image processing method include: to obtain to refer to the information of face with most like at least one of the first face in image to be processed；Based on it is described at least one refer to face information, to first face executor's face deformation process in the image to be processed with obtain output image.The image processing method can be adjusted the relative position of the human face five-sense-organ in image, and it is controllable to adjust degree, have U.S. face effect.
Description
Technical field
Embodiment of the disclosure is related to a kind of image processing method, image processing apparatus and storage medium.
Background technique
Face U.S. face is a common function of many mobile phone photograph softwares, and it is thin that big eye usually may be implemented in U.S. face algorithm
Face adjusts the functions such as the colour of skin, mill skin whitening, speckle removing, acne removing, Vitamin K, thin nose.For example, will include people by U.S. face algorithm
The picture of face is handled, and the effect of U.S. face can be realized in the picture after handling.These pictures, which can be, passes through cell-phone camera
The picture of head acquisition or the picture obtained by other means.
The relative position of human face five-sense-organ plays an important role to the height of face face value.Common U.S. face algorithm can not be right
The relative position of human face five-sense-organ is adjusted, limited to the landscaping effect of face.How the relative position of human face five-sense-organ is adjusted, with
The landscaping effect to face is improved, urgent problem to be solved is become.
Summary of the invention
At least one embodiment of the disclosure provides a kind of image processing method, comprising: obtains and the in image to be processed
Most like at least one of one face refers to the information of face；Based on it is described at least one refer to face information, to it is described to
First face executor's face deformation process in processing image is to obtain output image.
For example, in the image processing method that one embodiment of the disclosure provides, it is described based at least one described reference man
The information of face, to first face executor's face deformation process in the image to be processed to obtain output image, comprising: be based on
At least one described information for referring to face obtains the information of average face；According to the information of the average face to described wait locate
First face executor's face deformation process in image is managed to obtain the output image.
For example, in the image processing method that one embodiment of the disclosure provides, the acquisition and the in image to be processed
Most like at least one of one face refers to the information of face, comprising: Face datection is carried out to the image to be processed, to obtain
Multiple key points of first face；Multiple key points based on first face obtain the information of first face；
Based on the information of first face, face database is referred to from scheduled, is obtained most like at least with first face
One with reference to face and it is described at least one refer to face information, wherein it is described with reference to face database include multiple references
Face and each information with reference to face.
For example, in the image processing method that one embodiment of the disclosure provides, it is described based on the multiple of first face
Key point obtains the information of first face, comprising: executes trigonometric ratio processing to multiple key points of first face, obtains
To multiple triangle surfaces；Based on the multiple triangle surface, the information of first face is obtained.
For example, in the image processing method that one embodiment of the disclosure provides, it is described to be based on the multiple triangle surface,
Obtain the information of first face, comprising: by the first side length of each of the multiple triangle surface divided by described first
The face area of face is to be normalized, and using the first side length of each after normalization as first face
Information.
For example, in the image processing method that one embodiment of the disclosure provides, the information based on first face,
Face database is referred to from scheduled, obtains and refers to face and described at least one with most like at least one of first face
A information with reference to face, comprising: obtain each information with reference to face；By each information and institute with reference to face
The information for stating the first face is compared, thus to obtain described with reference to institute most like with first face in face database
State at least one with reference to face and it is described at least one refer to face information.
For example, in the image processing method that one embodiment of the disclosure provides, it is described by each letter with reference to face
Breath be compared with the information of first face, thus to obtain it is described refer to face database in the first face most phase
As it is described at least one with reference to face and it is described at least one refer to the information of face, comprising: according to first face
Information and each information with reference to face obtain first face with each and described refer to the corresponding multiple differences of face
Value, wherein the multiple difference indicates first face and each difference with reference to face；According to first face
With it is each described with reference to the corresponding the multiple difference of face, calculate first face with each described to refer to face corresponding
Fiducial value；Select at least one the smallest fiducial value it is corresponding at least one with reference to face as most like with first face
At least one refer to face.
For example, in the image processing method that one embodiment of the disclosure provides, the information according to first face
With each information with reference to face, obtain first face with it is each described with reference to the corresponding multiple differences of face, packet
It includes: trigonometric ratio processing being executed to first face and each reference face respectively, to respectively obtain multiple gores
Piece；By the first side length of each of multiple triangle surfaces of first face divided by the face area of first face with
It is normalized, by the second side length of each of each multiple triangle surfaces with reference to face divided by each described
The second side after the first side length and normalization with reference to the face area of face to be normalized, after being normalized
It is long, wherein the first side length of each in first face corresponds to each second side of each with reference in face
It is long；The first side length of each and each second side length with reference to after normalization corresponding in face after calculating normalization it
Difference, thus obtain first face with it is each described with reference to the corresponding multiple differences of face.
For example, in the image processing method that one embodiment of the disclosure provides, it is described according to first face and each
It is described to refer to the corresponding the multiple difference of face, first face is calculated compared with each reference face is corresponding
Value, comprising: the sum of the multiple difference is calculated, it will be described and as the fiducial value.
For example, in the image processing method that one embodiment of the disclosure provides, it is described according to first face and each
It is described to refer to the corresponding the multiple difference of face, first face is calculated compared with each reference face is corresponding
Value, comprising: distribute weight to each difference；The weighted sum for calculating the multiple difference, using the weighted sum as described in
Fiducial value.
For example, in the image processing method that one embodiment of the disclosure provides, the information based on first face,
Face database is referred to from scheduled, obtains and refers to face and described at least one with most like at least one of first face
A information with reference to face, further includes: before calculating the multiple difference, select it is described with reference in face database with it is described
First face has the reference face of identical gender, and the information of the reference face with identical gender is used for described more
The calculating of a difference.
For example, in the image processing method that one embodiment of the disclosure provides, it is described based at least one described reference man
The information of face obtains the information of average face, comprising: to it is described at least one with reference to face execute trigonometric ratio handle to obtain it is multiple
Triangle surface；To with it is each it is described carry out average computation with reference to the corresponding triangle surface of face, obtain the average face
Information.
For example, in the image processing method that one embodiment of the disclosure provides, the information according to the average face
To first face executor's face deformation process in the image to be processed to obtain the output image, comprising: obtain described
Multiple key points of average face；Using multiple key points of first face, multiple key points of the average face, obtain
To multiple key points of target face；According to multiple key points of the target face to the first in the image to be processed
Face executes face deformation process to obtain output image, wherein the output image includes the target face.
For example, utilizing multiple keys of first face in the image processing method that one embodiment of the disclosure provides
Point, the multiple key points for being averaged face obtain multiple key points of the target face using following formula:Wherein, pi,...,pNIndicate the target person
Multiple key points of face, eijIndicate each side of triangle surface, q'iIndicate multiple key points of the average face, λ table
Show adjustment factor, qiIndicate multiple key points of first face, i and j indicate the serial number of multiple key points, and i is not equal to
j。
For example, in the image processing method that one embodiment of the disclosure provides, it is described according to the multiple of the target face
Key point obtains output image to first face executor's face deformation process in the image to be processed, comprising: will be described
Multiple key points of target face hold the first face in the image to be processed as control point, using distortion of the mesh algorithm
Pedestrian's face deformation process is to obtain output image.
At least one embodiment of the disclosure also provides a kind of image processing apparatus, comprising: refers to face acquiring unit, configuration
For the first face in acquisition and image to be processed, at least one most like refers to the information of face；Anamorphose unit, matches
Be set to based on it is described at least one refer to face information, at first face executor's face deformation in the image to be processed
Reason is to obtain output image.
At least one embodiment of the disclosure also provides a kind of image processing apparatus, comprising: processor；Memory；One or
Multiple computer program modules, one or more of computer program modules are stored in the memory and are configured as
It is executed by the processor, one or more of computer program modules include for realizing described in disclosure any embodiment
Image processing method instruction.
At least one embodiment of the disclosure also provides a kind of storage medium, for storing the computer-readable finger of non-transitory
It enables, can execute and be realized described in disclosure any embodiment when the non-transitory computer-readable instruction is executed as computer
Image processing method instruction.
Detailed description of the invention
In order to illustrate more clearly of the technical solution of the embodiment of the present disclosure, the attached drawing to embodiment is simply situated between below
It continues, it should be apparent that, the accompanying drawings in the following description merely relates to some embodiments of the present disclosure, rather than the limitation to the disclosure.
Fig. 1 is a kind of schematic flow chart for image processing method that one embodiment of the disclosure provides；
Fig. 2 is a kind of specific flow chart for image processing method that one embodiment of the disclosure provides；
Fig. 3 is the schematic diagram for the triangle surface that face key point and trigonometric ratio are handled；
An exemplary process of step S120 in a kind of image processing method that Fig. 4 provides for one embodiment of the disclosure
Figure；
An exemplary process of step S130 in a kind of image processing method that Fig. 5 provides for one embodiment of the disclosure
Figure；
An exemplary process of step S132 in a kind of image processing method that Fig. 6 provides for one embodiment of the disclosure
Figure；
Fig. 7 is a kind of triangle surface that image processing method intermediate cam is handled that one embodiment of the disclosure provides
Schematic diagram；
An exemplary process of step S220 in a kind of image processing method that Fig. 8 provides for one embodiment of the disclosure
Figure；
Fig. 9 is a kind of schematic block diagram for image processing apparatus that one embodiment of the disclosure provides；
Figure 10 is the schematic block diagram for another image processing apparatus that one embodiment of the disclosure provides；
Figure 11 is the schematic block diagram for another image processing apparatus that one embodiment of the disclosure provides；And
Figure 12 is a kind of schematic diagram for storage medium that one embodiment of the disclosure provides.
Specific embodiment
To keep the purposes, technical schemes and advantages of the embodiment of the present disclosure clearer, below in conjunction with the embodiment of the present disclosure
Attached drawing, the technical solution of the embodiment of the present disclosure is clearly and completely described.Obviously, described embodiment is this public affairs
The a part of the embodiment opened, instead of all the embodiments.Based on described embodiment of the disclosure, ordinary skill
Personnel's every other embodiment obtained under the premise of being not necessarily to creative work, belongs to the range of disclosure protection.
Unless otherwise defined, the technical term or scientific term that the disclosure uses should be tool in disclosure fields
The ordinary meaning for thering is the personage of general technical ability to be understood." first ", " second " used in the disclosure and similar word are simultaneously
Any sequence, quantity or importance are not indicated, and are used only to distinguish different component parts.Equally, "one", " one " or
The similar word such as person's "the" does not indicate that quantity limits yet, but indicates that there are at least one." comprising " or "comprising" etc. are similar
Word mean to occur element or object before the word cover the element for appearing in the word presented hereinafter or object and its
It is equivalent, and it is not excluded for other elements or object.The similar word such as " connection " or " connected " be not limited to physics or
The connection of person's machinery, but may include electrical connection, it is either direct or indirect."upper", "lower", " left side ",
" right side " etc. is only used for indicating relative positional relationship, after the absolute position for being described object changes, then the relative positional relationship
May correspondingly it change.
The relative position of human face five-sense-organ plays an important role to the height of face face value.Common U.S. face algorithm, such as
The thin face of big eye, the algorithm for adjusting the functions such as the colour of skin, mill skin whitening, speckle removing, acne removing, Vitamin K, thin nose can be achieved, it can not be right
The relative position of human face five-sense-organ is adjusted, limited to the landscaping effect of face.
A disclosure at least embodiment provides a kind of image processing method, image processing apparatus and storage medium.The image
Processing method can be adjusted the relative position of the human face five-sense-organ in image, and it is controllable to adjust degree, have U.S. face effect
Fruit.
In the following, embodiment of the disclosure will be described in detail with reference made to the accompanying drawings.It should be noted that identical in different attached drawings
Appended drawing reference will be used to refer to the identical element that has described.
A disclosure at least embodiment provides a kind of image processing method, the image processing method include: obtain with wait locate
Manage the information that most like at least one of the first face in image refers to face；Based on it is described at least one refer to face letter
Breath, to first face executor's face deformation process in the image to be processed to obtain output image.
Fig. 1 is a kind of schematic flow chart for image processing method that one embodiment of the disclosure provides.With reference to Fig. 1, the figure
As processing method the following steps are included:
Step S10: the information (example that face is referred to most like at least one of the first face in image to be processed is obtained
Such as, the information that face is referred to most like at least one of the first face is obtained from reference face database)；
Step S20: referring to the information of face based at least one, becomes to first face executor's face in image to be processed
Shape processing is to obtain output image.
For example, in step slo, image to be processed can be the various images comprising face, such as character image etc..To
Handling image for example can be gray level image, or color image.For example, image to be processed can pass through image appropriate
Acquisition device obtains.The image collecting device can for digital camera, the camera of smart phone, tablet computer camera,
Camera, IP Camera, monitoring camera or other applicable components of personal computer, embodiment of the disclosure to this not
It limits.
For example, image to be processed can be the original image that image collecting device directly collects, it is also possible to original
The image that beginning image obtains after being pre-processed.For example, before step S10, at the image that embodiment of the disclosure provides
Reason method can also include carrying out pretreated operation to image to be processed, in order to detect the first face in image to be processed
Human face region.Pretreatment can eliminate irrelevant information or noise information in image to be processed, in order to preferably treat place
It manages image and carries out Face datection.For example, in the case where image to be processed is photo, above-mentioned pretreatment may include to photo into
The processing such as row image scaling, compression or format conversion, color gamut conversion, gamma (Gamma) correction, image enhancement or noise reduction filtering；
In the case where image to be processed is video, above-mentioned pretreatment may include the processing such as key frame for extracting video.
For example, the first face is the face handled, that is, need to carry out the face of face relative position adjustment.Example
It such as, may include one or more first faces in image to be processed, which can be in image to be processed
One or more first faces carry out the adjustment of face relative position.
For example, more meeting target with reference to face face value with higher for facial image gathered in advance with reference to face
The aesthetic conceptions of spectators (group), the target audience can be spectators, the spectators of certain culture background, certain duty in some area
The spectators etc. of industry.For example, may include male's face and women face with reference to face, it is saved to reference to face with reference to face number
According in library, in order to inquire and obtain.The face relative positional relationship with reference to face is represented with reference to the information of face, it can be by
It is saved together to reference in face database.For example, being pre-established above-mentioned with reference to face before executing the image processing method
Database can be inquired when executing the image processing method every time by reference to face database, to obtain needs
At least one refers to face and its information.
Here, " at least one most like refers to face " indicates: if what is obtained is one with reference to face, the reference
Face will be closer to the first face with reference to face compared to other in reference face database；If what is obtained is multiple ginsengs
Examine face, i.e., one group refers to face, then it is multiple with reference to face compared to other in reference face database with reference to face all
Will be closer to the first face, but multiple degree of closeness with reference to face and the first face is not necessarily the same.
For example, in step S20, to first face executor's face deformation process in image to be processed, to adjust first
The face relative position of face carries out image procossing, to obtain output image.In the output image, the face of the first face
Relative position is optimized according to the information of reference face, that is, realizes the micro-shaping effect of face relative position, therefore real
The effect of the U.S. face for the first face in output image is showed.For example, Facial metamorphosis processing can use common grid
Warping algorithm realize, embodiment of the disclosure to this with no restriction.
The image processing method is based on the information for referring to face with most like at least one of the first face, to the first face
Face deformation process is executed, i.e. progress face relative position adjustment, to realize the effect of U.S. face.
Fig. 2 is a kind of specific flow chart for image processing method that one embodiment of the disclosure provides.With reference to Fig. 2, the image
Processing method the following steps are included:
Step S110: Face datection is carried out to image to be processed, to obtain multiple key points of the first face；
Step S120: multiple key points based on the first face obtain the information of the first face；
Step S130: the information based on the first face refers to face database from scheduled, obtains and the first face most phase
As at least one with reference to face and this at least one refer to the information of face；
Step S210: the information of average face is obtained with reference to the information of face based at least one；
Step S220: according to the information of average face to first face executor's face deformation process in image to be processed with
Obtain output image.
For example, the step S10 in Fig. 1 includes the steps that S110, S120 and S130 in the present embodiment；Step S20 in Fig. 1
Include the steps that S210 and S220 in the present embodiment.
For example, in step s 110, Face datection can use the method based on template, the method based on model or nerve
Network method etc. is realized.Method based on template for example may include eigenface method, linear discriminant analysis method, singular value point
Solution method, Dynamic link library matching process etc..Method based on model for example may include Hidden Markov Model, active shape mould
The methods of type and active appearance models.Neural network method for example may include convolutional neural networks (Convolutional
Neural Network, CNN) etc..For example, Face datection can also other faces existing using this field or occurring in the future
Detection method.
Critical point detection algorithm is obtained for example, it is also possible to be trained using machine learning algorithm.For example, can receive in advance
Collect a large amount of facial image (such as 10000 or more), manually marks out the key point in every facial image.
Then it is trained using machine learning algorithm (such as deep learning algorithm or the regression algorithm based on local feature etc.), thus
Obtain critical point detection algorithm.
For example, multiple key points of the first face can be the strong key point of some characterization abilities of face, for example, eye
The key points such as eyeball, canthus, eyebrow, cheekbone highest point, nose, mouth, chin and face outer profile.In embodiment of the disclosure
In, multiple key points of the first face refer to multiple key points of the first face in image to be processed.
For example, with women face for the first face in example shown in Fig. 3 (1), available 81 by Face datection
Key point, these key points include eyes outer profile, eyebrow outer profile, nose, nostril, mouth outer profile, face outer profile etc..
It should be noted that face representated by key point are unrestricted in embodiment of the disclosure, can extract indicates eyes, eyebrow
The key point of any one or more face in the face such as hair, nose, mouth, can also extract indicates cheekbone position, face
The key point of outer profile, depending on this can according to need the U.S. face effect of realization, embodiment of the disclosure to this with no restriction.It closes
The quantity of key point is also unrestricted, can be any numbers such as 63,81,83,109, the quantity of key point for example can be according to fortune
Depending on calculation ability, detection accuracy.
For example, in the step s 120, multiple key points of the first face can be executed with trigonometric ratio processing, to obtain the
The information of one face.For example, in one example, as shown in figure 4, step S120 may comprise steps of:
Step S121: trigonometric ratio processing is executed to multiple key points of the first face, obtains multiple triangle surfaces；
Step S122: multiple triangle surfaces are based on, the information of the first face is obtained.
For example, trigonometric ratio processing can extract the face relative positional relationship of the first face in step S121.Trigonometric ratio
Processing can use delaunay triangulation (Delaunay Triangulation) algorithm, can also use other applicable points
The triangulation of collection, embodiment of the disclosure to this with no restriction.For example, using Delaunay triangulation algorithm to Fig. 3
(1) multiple key points in are handled, available multiple as shown in Fig. 3 (2) by connecting three adjacent key points
Triangle surface (or triangle).These triangle surfaces contain face relative position information.For example, each key point is joined
With construct above-mentioned multiple triangle surfaces, also, each triangle surface is each other without intersecting or overlapping, that is, arbitrarily
The side of two triangle surfaces is non-cross.It can refer to conventional side about trigonometric ratio processing and the detailed description of triangle surface
Method, details are not described herein again.
For example, in step S122, it can be by the first side length of each of multiple triangle surfaces divided by the first face
Face area is to be normalized, and using the first side length of each after normalization as the information of the first face.As a result,
The information of first face can represent the face relative positional relationship of the first face.For example, the first face of analysis can be passed through
Multiple key points obtain the face area of the first face.In this step, it is normalized, different pictures can be coped with
Resolution ratio and the first face ratio shared in image to be processed, so that the information of the first face and other faces (such as join
Examine face) information be comparable, consequently facilitating the processing of subsequent step.Here, " the first side length " is referred to the first
Face carries out each side length of the multiple triangle surfaces obtained after trigonometric ratio processing, " first " and is not specific to certain gores
Piece or certain side lengths, and it is only intended to the side length with the triangle surface in other faces (such as with reference to face) described below
It distinguishes.
For example, as shown in Fig. 2, in step s 130, the available information with reference to the reference face in face database,
And be compared with the information of the first face, thus obtain with most like at least one of the first face with reference to face and this at least
One information with reference to face.For example, in one example, as shown in figure 5, step S130 may comprise steps of:
Step S131: each information with reference to face is obtained；
Step S132: each information with reference to face is compared with the information of the first face, thus to obtain reference man
In face database with most like at least one of the first face with reference to face and this at least one with reference to face information.
For example, can carry out Face datection before step S131 to each reference face and obtain multiple key points, so
Trigonometric ratio processing is executed to each key point with reference to face afterwards, to obtain multiple gores corresponding with each reference face
Piece.The second side length of each of each triangle surface is normalized, i.e., by the second side length of each divided by correspondence
Reference face face area, and using the second side length of each after normalized as the corresponding letter with reference to face
Breath.For example, the information with reference to face can represent the face relative positional relationship with reference to face.For example, by the letter of reference face
Breath storage is into reference face database, in order to inquire and obtain.Here, " the second side length " refers to carrying out reference face
The each side length of multiple triangle surfaces obtained after trigonometric ratio processing, " second " and be not specific to certain triangle surfaces or certain
A little side lengths, and be only intended to distinguish with the first side length of the triangle surface in above-described first face.
For example, in step S131, it is available with reference to each of face database with reference to the information of face.Certainly,
Embodiment of the disclosure is without being limited thereto, for example, it is also possible to which obtaining has identical gender with the first face with reference in face database
Reference face information, to reach preferably U.S. face effect by the processing of subsequent step.For example, can be in step
The gender that the first face is detected before S131, can also obtain the gender of the first face, the disclosure by way of being manually entered
Implementation to this with no restriction.For example, with reference to each gender with reference to face is stored in face database.
It should be noted that in embodiment of the disclosure the image processing method can be being executed with reference to face database
Preceding building, in order to be used when executing the image processing method.For example, may include multiple reference men with reference to face database
Face and each information with reference to face also may further include each gender with reference to face and each with reference to face pair
The multiple key points answered, can also include other desired data and information, such as can also be at artificial as needed or machine
Reason addition keyword etc., embodiment of the disclosure to this with no restriction.This can use data appropriate with reference to face database
Library form, such as relevant database or non-relational database, such as can be with the image procossing of the realization embodiment of the present disclosure
The program of method operate in same computer or isolated operation database server in a local network on, or operate in mutually
Database server (such as Cloud Server) in networking is first-class.For example, referring to the information of face in order to obtain, need to reference
Face carries out Face datection and trigonometric ratio processing, and used algorithm carries out Face datection and trigonometric ratio processing with to the first face
Algorithm it is identical.Quantity with reference to the reference face in face database is unrestricted, can be according to memory space and operation energy
Depending on power.For example, may include 100 males with higher face value with reference to face and 100 women with reference to face database
With reference to face.For example, this with reference to face database can be carried out at any time after the completion of building supplement and it is perfect so that the disclosure is real
The U.S. face effect for applying the image processing method of example offer is more preferable.
For example, in one example, as shown in fig. 6, step S132 may comprise steps of:
Step S1321: according to the information of the first face and each information with reference to face, the first face and each ginseng are obtained
Examine the corresponding multiple differences of face；
Step S1322: according to the first face with each with reference to the corresponding multiple differences of face, the first face and each is calculated
With reference to the corresponding fiducial value of face；
Step S1323: select at least one the smallest fiducial value it is corresponding at least one with reference to face as with it is the first
Most like at least one of face refers to face.
For example, trigonometric ratio processing can be executed to the first face and each reference face respectively in step S1321, with
Respectively obtain multiple triangle surfaces.Then by the first side length of each of multiple triangle surfaces of the first face divided by first
The face area of face is to be normalized, by the second side length of each of each multiple triangle surfaces with reference to face
Divided by each face area with reference to face to be normalized, thus the first side length and normalization after being normalized
The second side length afterwards.For example, the first side length of each in the first face corresponds to each each second with reference in face
Side length.The first side length of each and each the second side length with reference to after normalization corresponding in face after calculating normalization it
Thus difference obtains the first face multiple differences corresponding with each reference face.For example, multiple differences indicate the first face and every
A difference with reference between face.
For example, the first side length after normalization has been obtained in step S122 before, therefore in step S1321
In can be omitted the process of the first side length after calculating normalization, the first side length after directlying adopt obtained normalization into
The calculating of row difference.For example, with reference to each information with reference to face is stored in face database, with reference to the information of face
Including the second side length after normalization, therefore it can be omitted the mistake for calculating the second side length after normalizing in step S1321
Journey directlys adopt the calculating for being stored in and carrying out difference with reference to the second side length after the normalization in face database.
For example, as shown in fig. 7, the first triangle surface A1 is a triangle surface in the first face, the second triangle
Shape dough sheet A2 is with reference to a triangle surface in face.First triangle surface A1 is opposite with the second triangle surface A2
It answers, i.e. the first triangle surface A1 and the second triangle surface A2 correspond to the first face and with reference to identical face in face
Position, key point qi, qj, qk of the first face and key point qi ', qj ', qk ' with reference to face also correspond respectively to the first face
With face identical in reference face and profile position.In step S1321, for example, calculating separately the first triangle surface A1
Normalization after first side length di, dj, dk and the second triangle surface A2 normalization after the second side length di ', dj ', dk '
Difference, that is, calculate separately di and di ' difference, the difference of dj and dj ', the difference of dk and dk ', to obtain the first face and each reference
The corresponding multiple differences of face.For example, each difference is the first side length and the second side after corresponding normalization after normalization
The absolute value of the difference of length.It should be noted that Fig. 7 schematically shows only a triangle surface and ginseng in the first face
A triangle surface in face is examined, when carrying out the calculating of difference, needs to calculate whole gores in the first face
The difference of piece and the correspondence side length with reference to whole triangle surfaces in face.
In step S1322, calculating the first face fiducial value corresponding with each reference face can be there are many mode.
For example, in one example, can calculate the first face with it is each with reference to the corresponding multiple differences of face and, and
Should and as the first face with this with reference to the corresponding fiducial value of face.For example, the fiducial value can indicate are as follows:
Wherein, R indicate the first face with this with reference to the corresponding fiducial value of face, i indicates the serial number of side length, di expression first
The first side length after normalization in face, di ' are indicated with reference to the second side length after the normalization in face, and di and di ' phase
It is corresponding.
For example, in another example, weight can be distributed to each difference, then calculate the first face and each reference
The weighted sum of the corresponding multiple differences of face, and using the weighted sum as the first face with this with reference to the corresponding fiducial value of face.
For example, the fiducial value can indicate are as follows:
Wherein, R indicate the first face with this with reference to the corresponding fiducial value of face, i indicates the serial number of side length, di expression first
The first side length after normalization in face, di ' are indicated with reference to the second side length after the normalization in face, and di and di ' phase
Corresponding, n indicates weight.For example, can according to need the significance level of the different parts of the face of adjustment as corresponding difference point
With weight.
It should be noted that the mode for calculating fiducial value is not limited to manner described above in embodiment of the disclosure,
Can be other applicable modes, embodiment of the disclosure to this with no restriction.
In step S1323, select at least one the smallest fiducial value it is corresponding at least one with reference to face as with the
Most like at least one of one face refers to face.For example, S1322 has obtained the first face and each reference through the above steps
The corresponding fiducial value of face, these fiducial values represent the first face and each comprehensive differences with reference to face.Therefore, the smallest
One or more fiducial values are corresponding with reference to the comprehensive differences of face and the first face minimum, so as to select these reference men
Face is as the reference face most like with the first face.For example, the selected and most like number with reference to face of the first face
Measure it is unrestricted, such as can be 3,5 etc., or any other number, embodiment of the disclosure do not limit this
System.Selection can reduce Facial metamorphosis with the most like reference face of the first face with the processing and calculating for subsequent step
The change degree of first face when processing, to improve U.S. face effect.
For example, as shown in Fig. 2, the information of average face can be obtained according to the information of reference face in step S210.
For example, in one example, when with the most like reference face of the first face being multiple, the face characteristic of average face is represented
The average level of multiple face characteristics with reference to face.For example, can handle to obtain to multiple reference faces execution trigonometric ratios more
Then a triangle surface carries out average computation to triangle surface corresponding with each reference face, to obtain average people
The information of face.For example, the information of average face include multiple triangle surfaces of average face normalization after each side
It is long.Can be with reference to conventional face fusion algorithm about the method for obtaining average face, details are not described herein again.For example, another
In a example, when the reference face most like with the first face is only one, then the information using this with reference to face is as average people
The information of face.
For example, as shown in Fig. 2, in step S220, according to the information of average face to the first in image to be processed
Face executes face deformation process, to obtain the output image after U.S. face.For example, in one example, as shown in figure 8, step
S220 may comprise steps of:
Step S221: multiple key points of average face are obtained；
Step S222: using multiple key points of the first face, multiple key points of average face, target face is obtained
Multiple key points；
Step S223: first face executor's face in image to be processed is deformed according to multiple key points of target face
Processing is to obtain output image.
For example, obtaining multiple key points of average face in step S221.For example, can be in aforementioned step S210
The key point of average face is extracted when the middle information for calculating average face, therefore can be directly acquired averagely in step S221
The key point of face no longer needs to detect and extract by respective algorithms.
It, can be with using multiple key points of the first face, multiple key points of average face for example, in step S222
Using following expression, to obtain multiple key points of target face:
Wherein, pi,...,pNIndicate multiple key points of target face, eijIndicate each side of triangle surface, q'i
Indicate multiple key points of average face, λ indicates adjustment factor, qiIndicate multiple key points of the first face, i and j indicate more
The serial number of a key point, and i is not equal to j.
For example, target face is to export the face for including in image, that is, target face is to first face executor's face
Obtained face after deformation process.By above-mentioned expression formula, multiple key points of target face are obtained, consequently facilitating subsequent step
Face deformation process is executed in rapid to obtain the image of target face.Illustratively, multiple key points and first of target face
Multiple key points of face correspond.Multiple key points of average face and multiple key points of the first face are preceding
It states and is obtained in each step, therefore be used directly for the calculating of above-mentioned expression formula in step S222.
By multiple key points of the available target face of above-mentioned expression formula, multiple key points of target face both with it is flat
Multiple key point coordinates of equal face are close and close with the coordinate of multiple key points of the first face.λ is adjustment factor,
Numerical values recited illustrates the similarity degree of target face and the first face, such as can be being executed at the image every time according to demand
Sets itself when reason method.By adjusting the size of λ, the similarity degree of target face and the first face, namely control can control
The adjustment degree of face relative position processed.For example, the adjustment degree of face relative position is smaller when the numerical value of λ is larger, so as to
So that target face is more like with the first face while having certain U.S. face effect；When the numerical value of λ is smaller, face are with respect to position
The adjustment degree set is larger, to keep target face and the higher average face of face value more like, to have preferably U.S. face effect
Fruit.This mode makes the adjustment degree of face relative position controllable, to meet a variety of application demands.It should be noted that this
In disclosed embodiment, λ can be when executing the image processing method by user's sets itself every time, or a fixation
Preset value, the set-up mode of λ can according to actual needs depending on, embodiment of the disclosure to this with no restriction.
For example, in step S223, it can be using multiple key points of target face as target control point, by figure to be processed
Image uniform to be processed is divided into grid, is calculated using distortion of the mesh by multiple key points of the first face as source control point as in
Method obtains output image to first face executor's face deformation process in image to be processed.For example, Facial metamorphosis processing can
To be realized using conventional mesh torsion algorithm, embodiment of the disclosure to this with no restriction.It exports and is contained in image as a result,
Target face, and the face relative position of target face is optimized and is adjusted, therefore the face value of target face is higher than first
Face, to realize the effect of U.S. face.
It should be noted that the execution sequence of each step of image processing method is unrestricted in embodiment of the disclosure
System, although describing the implementation procedure of each step above with particular order, this does not constitute the limit to the embodiment of the present disclosure
System.Each step in the image processing method can be executed serially or be executed parallel, this can according to actual needs depending on.It should
Image processing method can also include the steps that it is more or less, for example, increasing to reach preferably U.S. face effect
Pre-treatment step, or the data of some pilot process are stored and are used for subsequent processing and calculating, it is some similar to omit
Step.
A disclosure at least embodiment also provides a kind of image processing apparatus, which can be in image
The relative position of human face five-sense-organ is adjusted, and it is controllable to adjust degree, has U.S. face effect.
Fig. 9 is a kind of schematic block diagram for image processing apparatus that one embodiment of the disclosure provides.With reference to Fig. 9, image procossing
Device 300 may include with reference to face acquiring unit 310 and anamorphose unit 320.Image processing apparatus 300 for example with image
The connection of 330 signal of acquisition device can obtain the image to be processed comprising the first face from image collecting device 330, should be wait locate
Managing image can be static images or video etc..Image processing apparatus 300 is also connect with 340 signal of image output device, is used for
The output image that processing obtains is exported by image output device 340, for showing or storing.
It is configured to obtain and most like at least one of the first face in image to be processed with reference to face acquiring unit 310
With reference to the information of face.Anamorphose unit 320 is configured at least one information for referring to face, in image to be processed
First face executor's face deformation process with obtain output image.For example, in one example, with reference to face acquiring unit
310 may include with reference to face database.For example, in another example, as shown in figure 9, can be with reference to face database 400
It is provided by the server being separately arranged, and signal connects therewith by wired, wireless or similar mode, image processing apparatus
300 server communications that can be separately arranged with this, to carry out inquiry and acquisition of information to reference face database.
For example, image processing apparatus 300 can be applied in any electronic equipment with photograph or camera function.Electronics
Equipment can be for example smart phone, tablet computer, laptop, desktop computer, digital camera etc..Certainly, the disclosure
Embodiment is without being limited thereto, and image processing apparatus 300 also can be applied to other electronic equipments for not having photograph or camera function
In.For example, image processing apparatus 300 or an independent electronic equipment.
For example, with reference to face acquiring unit 310 and anamorphose unit 320 can for hardware, software, firmware and they
Any feasible combination.For example, can be dedicated or general with reference to face acquiring unit 310 and anamorphose unit 320
Circuit, chip or device etc., or the combination of processor and memory.For example, in one example, being obtained with reference to face
Unit 310 can also include signal receiving device (such as antenna), modulation-demodulation device, storage device etc..About reference face
The specific implementation form of acquiring unit 310 and anamorphose unit 320, embodiment of the disclosure to this with no restriction.
For example, image to be processed can be obtained by image collecting device 330, and it is transmitted in image processing apparatus 300.
Image collecting device 330 may include the camera of smart phone, the camera of tablet computer, personal computer camera,
Digital camera or network shooting head etc., such as can be used for providing arbitrary image resource, such as the image locally prestored
File, the image file for interconnecting online storage etc..Image output device 340 for example can be display screen, printer, modulation /demodulation
Device etc..
When above-mentioned image processing apparatus 300, image collecting device 330 and image output device 340 are arranged in same equipment
Among when, between them can by system bus, signal is connected each other, quickly to transmit data.
It should be noted that in embodiment of the disclosure, at each unit of image processing apparatus 300 and image above-mentioned
Each step of reason method is corresponding, and the concrete function about image processing apparatus 300 can be with reference to about image processing method
Associated description, details are not described herein again.The component and structure of image processing apparatus 300 shown in Fig. 9 are only exemplary, rather than are limited
Property processed, as needed, which can also include other assemblies and structure.
Figure 10 is the schematic block diagram for another image processing apparatus that one embodiment of the disclosure provides.With reference to Figure 10, image
Processing unit 300 may include processor 350 and memory 360.Memory 360 is for storing the computer-readable finger of non-transitory
Enable (such as one or more computer program modules).Processor 350 is non-temporary for running non-transitory computer-readable instruction
When property computer-readable instruction can execute one or more in image processing method described above when being run by processor 350
A step.Memory 360 and processor 350 can be mutual by bindiny mechanism's (not shown) of bus system and/or other forms
Even.
For example, memory 360 and processor 350 etc. can be set in server end (or cloud).Certainly, the reality of the disclosure
It is without being limited thereto to apply example, memory 360 and processor 350 etc. also can be set at Image Acquisition end.
For example, processor 350 can be central processing unit (CPU), digital signal processor (DSP) or have data
Processing unit of the other forms of processing capacity and/or program executive capability, such as field programmable gate array (FPGA) etc.；Example
Such as, central processing unit (CPU) can be X86 or ARM framework etc..Processor 350 can be general processor or dedicated processes
Device can control other components in image processing apparatus 300 to execute desired function.
For example, memory 360 may include any combination of one or more computer program products, computer program is produced
Product may include various forms of computer readable storage mediums, such as volatile memory and/or nonvolatile memory.Easily
The property lost memory for example may include random access memory (RAM) and/or cache memory (cache) etc..It is non-volatile
Property memory for example may include read-only memory (ROM), hard disk, Erasable Programmable Read Only Memory EPROM (EPROM), portable
Aacompactadisk read onlyamemory (CD-ROM), USB storage, flash memory etc..It can store one on computer readable storage medium
Or multiple computer program modules, processor 350 can run one or more computer program modules, to realize image procossing
The various functions of device 300.Can also store in a computer-readable storage medium various application programs and various data and
Application program use and/or the various data generated etc..
Figure 11 is the schematic block diagram for another image processing apparatus that one embodiment of the disclosure provides.With reference to Figure 11, image
Processing unit 300 includes processor 350 and memory 360.The image processing apparatus 300 and image processing apparatus shown in Fig. 10
300 is essentially identical, and details are not described herein again.In this embodiment, it is stored in other devices with reference to face database 400, such as
Server, cloud storage etc..When image processing apparatus 300 works, communication network (such as cable LAN, nothing can be passed through
Line local area network, 3G/4G/5G communication network etc.) it is read based on corresponding communication protocol with reference to the number stored in face database 400
According to and information.For example, communication protocol can be Bluetooth communication protocol, Ethernet, serial interface communication agreement, parallel interface communication
Any suitable communication protocol such as agreement, embodiment of the disclosure to this with no restriction.Image processing apparatus 300 can be by having
Line or the server communication that face database 400 is wirelessly referred to storage.It should be noted that embodiment of the disclosure
In, the concrete function and technical effect of image processing apparatus 300 can be with reference to the descriptions above in connection with image processing method, this
Place repeats no more.
A disclosure at least embodiment also provides a kind of storage medium, for storing non-transitory computer-readable instruction,
It can execute and be realized described in disclosure any embodiment when the non-transitory computer-readable instruction is executed as computer
The instruction of image processing method.The image processing method can be executed using the storage medium, it can be to the face five in image
The relative position of official is adjusted, and it is controllable to adjust degree, has U.S. face effect.
Figure 12 is a kind of schematic diagram for storage medium that one embodiment of the disclosure provides.With reference to Figure 12, storage medium 500 is used
In storage non-transitory computer-readable instruction 510.For example, when non-transitory computer-readable instruction 510 is executed by computer
When can execute according to one or more steps in image processing method described above.
For example, the storage medium 500 can be applied in above-mentioned image processing apparatus 300.For example, storage medium 500 can
Think the memory 360 in image processing apparatus 300 shown in Figure 10 or Figure 11.For example, mutually speaking on somebody's behalf about storage medium 500
The bright corresponding description that can refer to the memory 360 in image processing apparatus 300 shown in Fig. 10, details are not described herein again.
There is the following to need to illustrate:
(1) embodiment of the present disclosure attached drawing relates only to the structure that the embodiment of the present disclosure is related to, and other structures can refer to logical
Standing meter.
(2) in the absence of conflict, the feature in embodiment of the disclosure and embodiment can be combined with each other to obtain
New embodiment.
The above, the only specific embodiment of the disclosure, but the protection scope of the disclosure is not limited thereto, this public affairs
The protection scope opened should be based on the protection scope of the described claims.
Claims (18)
1. a kind of image processing method, comprising:
Obtain the information that face is referred to most like at least one of the first face in image to be processed；
Based on it is described at least one refer to face information, at first face executor's face deformation in the image to be processed
Reason is to obtain output image.
2. image processing method according to claim 1, wherein it is described based on it is described at least one refer to face letter
Breath, to first face executor's face deformation process in the image to be processed to obtain output image, comprising:
The information of average face is obtained based at least one described information for referring to face；
According to the information of the average face to first face executor's face deformation process in the image to be processed to obtain
The output image.
3. image processing method according to claim 1 or 2, wherein it is described acquisition with it is the first in image to be processed
Most like at least one of face refers to the information of face, comprising:
Face datection is carried out to the image to be processed, to obtain multiple key points of first face；
Multiple key points based on first face obtain the information of first face；
Based on the information of first face, face database is referred to from scheduled, is obtained most like with first face
At least one with reference to face and it is described at least one refer to face information, wherein it is described with reference to face database include multiple
With reference to face and each information with reference to face.
4. image processing method according to claim 3, wherein multiple key points based on first face obtain
To the information of first face, comprising:
Trigonometric ratio processing is executed to multiple key points of first face, obtains multiple triangle surfaces；
Based on the multiple triangle surface, the information of first face is obtained.
5. image processing method according to claim 4, wherein it is described to be based on the multiple triangle surface, obtain institute
State the information of the first face, comprising:
By the first side length of each of the multiple triangle surface divided by the face area of first face to carry out normalizing
Change processing, and using the first side length of each after normalization as the information of first face.
6. image processing method according to claim 3, wherein the information based on first face, from predetermined
Reference face database, obtain with most like at least one of first face with reference to face and at least one described reference
The information of face, comprising:
Obtain each information with reference to face；
Each information with reference to face is compared with the information of first face, refers to face thus to obtain described
In database with first face it is most like described at least one with reference to face and it is described at least one refer to face letter
Breath.
7. image processing method according to claim 6, wherein it is described by each information with reference to face with it is described
The information of first face is compared, thus to obtain it is described with reference in face database with first face it is most like described in
At least one with reference to face and it is described at least one refer to face information, comprising:
According to the information of first face and each information with reference to face, first face and each described is obtained
With reference to the corresponding multiple differences of face, wherein the multiple difference indicates first face and each described with reference to face
Difference；
According to first face with it is each described with reference to the corresponding the multiple difference of face, calculate first face and every
It is a described with reference to the corresponding fiducial value of face；
Select at least one the smallest fiducial value it is corresponding at least one with reference to face as most like with first face
At least one refers to face.
8. image processing method according to claim 7, wherein the information according to first face and each institute
The information with reference to face is stated, first face is obtained with each and described refers to the corresponding multiple differences of face, comprising:
Trigonometric ratio processing is executed to first face and each reference face respectively, to respectively obtain multiple gores
Piece；
By the first side length of each of multiple triangle surfaces of first face divided by the face area of first face
To be normalized, by the second side length of each of each multiple triangle surfaces with reference to face divided by each institute
The second side after the first side length and normalization that the face area with reference to face is stated to be normalized, after being normalized
It is long, wherein the first side length of each in first face corresponds to each second side of each with reference in face
It is long；
Calculate the first side length of each after normalizing and each second side length with reference to after normalization corresponding in face
Difference, thus obtain first face with it is each described with reference to the corresponding multiple differences of face.
9. image processing method according to claim 7, wherein described according to first face and each reference
The corresponding the multiple difference of face calculates first face with each and described refers to the corresponding fiducial value of face, comprising:
The sum of the multiple difference is calculated, it will be described and as the fiducial value.
10. image processing method according to claim 7, wherein described according to first face and each ginseng
It examines the corresponding the multiple difference of face, calculates first face with each and described refer to the corresponding fiducial value of face, comprising:
Weight is distributed to each difference；
The weighted sum for calculating the multiple difference, using the weighted sum as the fiducial value.
11. image processing method according to claim 7, wherein the information based on first face, from predetermined
Reference face database, obtain with most like at least one of first face with reference to face and at least one described reference
The information of face, further includes:
Before calculating the multiple difference, select that there is identical gender with first face in the reference face database
Reference face, and the information of the reference face with identical gender to be used for the calculating of the multiple difference.
12. image processing method according to claim 2, wherein it is described based on it is described at least one refer to face letter
Breath obtains the information of average face, comprising:
To it is described at least one with reference to face execute trigonometric ratio handle to obtain multiple triangle surfaces；
To with it is each it is described carry out average computation with reference to the corresponding triangle surface of face, obtain the information of the average face.
13. image processing method according to claim 2, wherein the information according to the average face is to described
First face executor's face deformation process in image to be processed is to obtain the output image, comprising:
Obtain multiple key points of the average face；
Using multiple key points of first face, multiple key points of the average face, the multiple of target face are obtained
Key point；
According to multiple key points of the target face to first face executor's face deformation process in the image to be processed
To obtain output image, wherein the output image includes the target face.
14. image processing method according to claim 13, wherein utilize multiple key points of first face, institute
The multiple key points for stating average face obtain multiple key points of the target face using following formula:
Wherein, pi,...,pNIndicate multiple key points of the target face, eijIndicate each side of triangle surface, q'i
Indicate multiple key points of the average face, λ indicates adjustment factor, qiIndicate multiple key points of first face, i and
J indicates the serial number of multiple key points, and i is not equal to j.
15. image processing method according to claim 13, wherein multiple key points according to the target face
To first face executor's face deformation process in the image to be processed to obtain output image, comprising:
Using multiple key points of the target face as control point, using distortion of the mesh algorithm in the image to be processed
First face executor's face deformation process is to obtain output image.
16. a kind of image processing apparatus, comprising:
With reference to face acquiring unit, it is configured to obtain at least one reference man most like with the first face in image to be processed
The information of face；
Anamorphose unit, be configured to it is described at least one refer to the information of face, to the in the image to be processed
One face executes face deformation process to obtain output image.
17. a kind of image processing apparatus, comprising:
Processor；
Memory；
One or more computer program modules, one or more of computer program modules are stored in the memory
And be configured as being executed by the processor, one or more of computer program modules include for realizing claim 1-
The instruction of 15 any image processing methods.
18. a kind of storage medium, for storing non-transitory computer-readable instruction, when the computer-readable finger of the non-transitory
The instruction for realizing any image processing method of claim 1-15 can be executed by enabling when being executed by computer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810955579.5A CN109087240B (en) | 2018-08-21 | 2018-08-21 | Image processing method, image processing apparatus, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810955579.5A CN109087240B (en) | 2018-08-21 | 2018-08-21 | Image processing method, image processing apparatus, and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109087240A true CN109087240A (en) | 2018-12-25 |
| CN109087240B CN109087240B (en) | 2023-06-06 |
Family
ID=64794148
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810955579.5A Active CN109087240B (en) | 2018-08-21 | 2018-08-21 | Image processing method, image processing apparatus, and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109087240B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110688948A (en) * | 2019-09-26 | 2020-01-14 | 北京百度网讯科技有限公司 | Method and device for transforming gender of human face in video, electronic equipment and storage medium |
| CN112037271A (en) * | 2020-07-16 | 2020-12-04 | 北京旷视科技有限公司 | Image processing method, apparatus, device and medium |
| CN113706369A (en) * | 2020-05-21 | 2021-11-26 | 北京达佳互联信息技术有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1932842A (en) * | 2006-08-10 | 2007-03-21 | 中山大学 | Three-dimensional human face identification method based on grid |
| CN104156993A (en) * | 2014-07-18 | 2014-11-19 | 小米科技有限责任公司 | Method and device for switching face image in picture |
| CN106919884A (en) * | 2015-12-24 | 2017-07-04 | 北京汉王智远科技有限公司 | Human facial expression recognition method and device |
| CN107122727A (en) * | 2017-04-20 | 2017-09-01 | 北京旷视科技有限公司 | Method, apparatus and system for face's shaping |
| CN107274354A (en) * | 2017-05-22 | 2017-10-20 | 奇酷互联网络科技（深圳）有限公司 | image processing method, device and mobile terminal |
| CN107506693A (en) * | 2017-07-24 | 2017-12-22 | 深圳市智美达科技股份有限公司 | Distort face image correcting method, device, computer equipment and storage medium |
| CN107730445A (en) * | 2017-10-31 | 2018-02-23 | 广东欧珀移动通信有限公司 | Image processing method, device, storage medium and electronic equipment |
| CN108269272A (en) * | 2018-01-31 | 2018-07-10 | 北京青燕祥云科技有限公司 | Liver's CT method for registering and system |
-
2018
- 2018-08-21 CN CN201810955579.5A patent/CN109087240B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1932842A (en) * | 2006-08-10 | 2007-03-21 | 中山大学 | Three-dimensional human face identification method based on grid |
| CN104156993A (en) * | 2014-07-18 | 2014-11-19 | 小米科技有限责任公司 | Method and device for switching face image in picture |
| CN106919884A (en) * | 2015-12-24 | 2017-07-04 | 北京汉王智远科技有限公司 | Human facial expression recognition method and device |
| CN107122727A (en) * | 2017-04-20 | 2017-09-01 | 北京旷视科技有限公司 | Method, apparatus and system for face's shaping |
| CN107274354A (en) * | 2017-05-22 | 2017-10-20 | 奇酷互联网络科技（深圳）有限公司 | image processing method, device and mobile terminal |
| CN107506693A (en) * | 2017-07-24 | 2017-12-22 | 深圳市智美达科技股份有限公司 | Distort face image correcting method, device, computer equipment and storage medium |
| CN107730445A (en) * | 2017-10-31 | 2018-02-23 | 广东欧珀移动通信有限公司 | Image processing method, device, storage medium and electronic equipment |
| CN108269272A (en) * | 2018-01-31 | 2018-07-10 | 北京青燕祥云科技有限公司 | Liver's CT method for registering and system |
Non-Patent Citations (2)
| Title |
|---|
| BYUNG K.JUNG等: "Magnetic Resonance Image Retrieval Based on Contour to Centroid Ttiangulation with Shape Feature Similarity", 《ACM》 * |
| 陈英鹏: "人脸特征点提取与匹配软件的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110688948A (en) * | 2019-09-26 | 2020-01-14 | 北京百度网讯科技有限公司 | Method and device for transforming gender of human face in video, electronic equipment and storage medium |
| CN113706369A (en) * | 2020-05-21 | 2021-11-26 | 北京达佳互联信息技术有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
| CN112037271A (en) * | 2020-07-16 | 2020-12-04 | 北京旷视科技有限公司 | Image processing method, apparatus, device and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109087240B (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109712080A (en) | Image processing method, image processing apparatus and storage medium | |
| US11922356B1 (en) | Emotion recognition for workforce analytics | |
| US11893828B2 (en) | System and method for image de-identification | |
| CN104680121B (en) | Method and device for processing face image | |
| US10799010B2 (en) | Makeup application assist device and makeup application assist method | |
| US11163978B2 (en) | Method and device for face image processing, storage medium, and electronic device | |
| KR20210110533A (en) | Method and apparatus for providing virtual plastic surgery sns service | |
| EP3399467A1 (en) | Emotion recognition in video conferencing | |
| CN106920277A (en) | Simulation beauty and shaping effect visualizes the method and system of online scope of freedom carving | |
| WO2019100282A1 (en) | Face skin color recognition method, device and intelligent terminal | |
| EP2884414A1 (en) | Health state determining method and apparatus using facial image | |
| CN109087240A (en) | Image processing method, image processing apparatus and storage medium | |
| CN104657974A (en) | Image processing method and device | |
| CN108537126A (en) | A kind of face image processing system and method | |
| US20200355945A1 (en) | Equipment to obtain 3d image data of a face and automatic method for customized modeling and manufacturing of eyeglass frames | |
| CN113327191A (en) | Face image synthesis method and device | |
| Kela et al. | Illumination invariant elastic bunch graph matching for efficient face recognition | |
| WO2022121886A1 (en) | Method and apparatus for identifying dress code for electric power operations | |
| CN109376661A (en) | Method and apparatus for output information | |
| Ubaid et al. | Human age prediction and classification using facial image | |
| TW201629907A (en) | System and method for generating three-dimensional facial image and device thereof | |
| Verma et al. | Efficient face recognition method using RBF kernel and genetic algorithm | |
| JP2021010652A (en) | Information processing device, evaluation method, and information processing program | |
| Lai et al. | Novel approach to neural network based caricature generation | |
| CN107944406A (en) | Automatic brightness-regulating system and method based on Eye states recognition technology |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |