CN110458117A - A kind of portraiture photography posture recommended method - Google Patents
A kind of portraiture photography posture recommended method Download PDFInfo
- Publication number
- CN110458117A CN110458117A CN201910749832.6A CN201910749832A CN110458117A CN 110458117 A CN110458117 A CN 110458117A CN 201910749832 A CN201910749832 A CN 201910749832A CN 110458117 A CN110458117 A CN 110458117A
- Authority
- CN
- China
- Prior art keywords
- pencil
- posture
- image
- divided
- recommended 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.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000004458 analytical method Methods 0.000 claims abstract description 14
- 238000002360 preparation method Methods 0.000 claims description 11
- 238000013135 deep learning Methods 0.000 claims description 7
- 238000012417 linear regression Methods 0.000 claims description 6
- 230000004044 response Effects 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 239000004744 fabric Substances 0.000 claims 1
- 238000004364 calculation method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000002087 whitening effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06T3/04—Context-preserving transformations, e.g. by using an importance map
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/60—Static or dynamic means for assisting the user to position a body part for biometric acquisition
- G06V40/67—Static or dynamic means for assisting the user to position a body part for biometric acquisition by interactive indications to the user
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to the technical fields of image procossing, more particularly to a kind of portraiture photography posture recommended method, it is designed to provide a kind of portraiture photography posture recommended method, before shooting portrait photo, the analysis that reference object number first is carried out to target image, shoots style, preset posture label is selected according to the result of analysis, after the recommendation posture element that reference object is provided according to system is posed, completes to shoot and forms picture;The following steps are included: (1) carries out reference object number to image data before shooting is completed, first and shoots the analysis of style;(2) preset posture label is selected according to the result of analysis;(3) the recommendation posture sketch provided in reference object according to system is completed to shoot and forms picture.
Description
Technical field
The present invention relates to the technical fields of image procossing, more particularly to a kind of portraiture photography posture recommended method.
Background technique
It is well known that with the development of traditional images processing method and deep learning image processing method, various intelligence hands
The machine APP that takes pictures joined various image algorithms during portraiture photography, such as U.S. face is magazine " mill skin ", " thin face ",
" whitening " function, this kind of APP is generally capable of user and improves " face value ", to like by user, still, when amateur photography
When teacher takes pictures to model, it tends to be difficult to provide guidance;When amateur model is facing to camera, also tends to be faced with and do not know this such as
The problem of what is posed.For example, schoolgirl wants to clap a lovely photo, boy student wants to clap a photo made fun, when party
Everybody wants to clap a tool group picture ingenious, but posture of suitably taking pictures of failing to give attention to again for the moment.
Summary of the invention
In order to solve the above technical problems, being clapped the purpose of the present invention is to provide a kind of portraiture photography posture recommended method
It takes the photograph before portrait photo, the analysis for first carrying out reference object number to target image, shooting style is selected according to the result of analysis
Preset posture label is completed to shoot and forms figure after the recommendation posture element that reference object is provided according to system is posed
Piece.
A kind of portraiture photography posture recommended method of the invention, comprising the following steps:
(1) before shooting is completed, reference object number first is carried out to image data and shoots the analysis of style;
(2) preset posture label is selected according to the result of analysis;
(3) the recommendation posture sketch provided in reference object according to system is completed to shoot and forms picture.
A kind of portraiture photography posture recommended method of the invention, the analysis of reference object number is according to bat in the step (1)
The number of taking the photograph is divided into single player mode and collective modes, wherein single player mode is divided into boy student and schoolgirl according to gender, and collective modes are divided into
Lovers, group builds and graduating photograph.
A kind of portraiture photography posture recommended method of the invention will recommend posture to be divided into aobvious height in single boy student's mode
Greatly, show strong, aobvious cruel, youth, it is sedate, make fun, be simple-minded, sunlight, madness;In single schoolgirl's mode, posture will be recommended to be divided into
Show thin in the face, aobvious stature, Gao Leng, it is lovely, show cruel, pure, charming, mature, youth, be sexuality, gracefulness, sunlight, not mature;In lovers' mould
In formula, will recommend posture be divided into tradition, prominent women, prominent male, stroll, kiss, leading along by hand, interest movement, sitting posture, shoulder simultaneously
Shoulder is kept for a flash, intention before some distances, kiss;It is built in group and posture in graduating photograph mode, will be recommended to be divided into unity, intimate, work
Power, love, youth, intention.
A kind of portraiture photography posture recommended method of the invention recommends the preparation method of posture sketch as follows:
A large amount of portrait picture is collected according to each quasi-mode posture label first;
Then the method by combining profile and tone generates the class pencil sketch of image on the basis of original image
Figure;
Then by the portrait parted pattern based on deep learning, corresponding portrait masking-out is generated on the basis of original image
(Mask);
It generates in conjunction with the sketch map of class pencil and portrait masking-out (Mask) finally by image AND operation and recommends posture
Sketch.
A kind of portraiture photography posture recommended method of the invention, the preparation method of the class pencil sketch map are as follows:
Firstly, generating the pencil stroke pattern (Pencil Stroke) for describing substantially space structure in image;
Then, it in conjunction with preset pencil sketch base map, generates for describing the shape in image, profile, shade, light and shade
Pencil tone illustration (Pencil Tone);
Finally, being generated final in conjunction with pencil stroke pattern (Pencil stroke) and pencil tone illustration (Pencil Tone)
Class pencil sketch map.
A kind of portraiture photography posture recommended method of the invention, the acquisition of the pencil stroke pattern (Pencil Stroke)
Method is as follows:
Firstly, original image is converted to grayscale image I;
Then, it on the basis of grayscale image I, calculates gradient G (Gradient);
Then, it on the basis of gradient G, calculates along east, the southeast, south, southwest, west, northwest, north, 8, northeast direction
Response diagram (Response Map);
Then, G is chosen in 8 directionsiMaximum value as map of magnitudes CiResult;
Finally, in map of magnitudes CiAnd line segmentOn the basis of, pencil stroke pattern S (Pencil Stroke) is calculated.
A kind of portraiture photography posture recommended method of the invention, the acquisition side of the pencil tone illustration (Pencil Tone)
Method is as follows:
Firstly, original image is converted to grayscale image I;
Then, on the basis of grayscale image I, pixel value is divided into three classes;
Then, it is fitted bright pixels collection P respectively using the parameter set A and presetting probability Distribution Model that learn in advance1, soft
Color pixel collection P2With dark-coloured set of pixels P3, fitting result is respectively P '1,P’2,P’3;
Then, in fitting result P '1,P’2,P’3On the basis of, general tint J is fitted using the parameter set B learnt in advance;
And then, on the basis of preset pencil sketch base map H, final fitting factor-beta is calculated*;
Finally, calculating pencil tone illustration T.
A kind of portraiture photography posture recommended method of the invention, the presetting probability Distribution Model are as follows:
General tint J is a linear regression model (LRM);Bright pixels collection p '1It (v) is a laplacian distribution;Soft colour
Element collection p'2(v) it is one to be uniformly distributed;Dark-coloured set of pixels p'3It (v) is a Gaussian Profile.
A kind of portraiture photography posture recommended method of the invention, the portrait parted pattern acquisition side based on deep learning
Method is as follows:
Firstly, building portrait partitioned data set;
Then, convolutional neural networks model is built;
Finally, start the training on training set, and when loss is almost without decline, deconditioning.
A kind of portraiture photography posture recommended method of the invention, the neural network model are broadly divided into following several moulds
Block: coding module (Encoder Module), Atrous spatial pyramid sampling module (Atrous Spatial Pyramid
Pooling Module), decoder module (Decoder Module).
Compared with prior art the invention has the benefit that can effectively be portrait painter by the above method
Model is provided and puts appearance reference, solves the problems, such as portrait pendulum appearance of taking pictures.
Detailed description of the invention
Fig. 1 is flow chart of the invention;
Fig. 2 is preset posture labeling figure;
Fig. 3 is the flow chart for obtaining posture sketch.
Specific embodiment
With reference to the accompanying drawings and examples, specific embodiments of the present invention will be described in further detail.Implement below
Example is not intended to limit the scope of the invention for illustrating the present invention.
Embodiment:
A kind of portraiture photography posture recommended method of the invention, comprising the following steps:
(1) before shooting is completed, reference object number first is carried out to image data and shoots the analysis of style, shooting pair
Analyzed as number and single player mode and collective modes be divided into according to shooting number, wherein single player mode according to gender be divided into boy student and
Schoolgirl, collective modes are divided into lovers, group builds and graduating photograph, in single boy student's mode, posture will be recommended to be divided into and show tall and big, aobvious strong
Strong, the aobvious cruel, youth, it is sedate, make fun, be simple-minded, sunlight, madness;In single schoolgirl's mode, by recommend posture be divided into show it is thin in the face,
Aobvious stature, Gao Leng, it is lovely, show cruel, pure, charming, mature, youth, be sexuality, gracefulness, sunlight, not mature;It, will in lovers' mode
Recommend posture be divided into tradition, prominent women, prominent male, stroll, kiss, lead along by hand, interest movement, sitting posture, shoulder to shoulder, holding
A flash, intention before some distances, kiss;It builds in group in graduating photograph mode, posture will be recommended to be divided into unity, intimate, vigor, love
Meaning, youth, intention;
(2) preset posture label is selected according to the result of analysis;
(3) the recommendation posture sketch provided in reference object according to system is completed to shoot and forms picture, wherein recommending appearance
The preparation method of state sketch is as follows:
A large amount of portrait picture is collected according to above-mentioned each quasi-mode posture label first;
Then the method by combining profile and tone generates the class pencil sketch of image on the basis of original image
Figure, the preparation method of class pencil sketch map are as follows: firstly, generating the pencil stroke pattern for describing substantially space structure in image
(Pencil Stroke) then in conjunction with preset pencil sketch base map, is generated for describing the shape in image, profile, yin
The pencil tone illustration (Pencil Tone) of shadow, light and shade, finally, in conjunction with pencil stroke pattern (Pencil stroke) and pencil tone
Scheme (Pencil Tone), generates final class pencil sketch map;
Then by the portrait parted pattern based on deep learning, corresponding portrait masking-out is generated on the basis of original image
(Mask);
It generates in conjunction with the sketch map of class pencil and portrait masking-out (Mask) finally by image AND operation and recommends posture
Sketch.
A kind of portraiture photography posture recommended method of the invention, the preparation method of pencil stroke pattern (Pencil Stroke)
It is as follows:
Firstly, original image is converted to grayscale image I;
Then, it on the basis of grayscale image I, calculates gradient G (Gradient), formula is as follows:
Wherein,WithIt is the gradient operator in x-axis direction and y-axis direction respectively;
Then, it on the basis of gradient G, calculates along east, the southeast, south, southwest, west, northwest, north, 8, northeast direction
Response diagram (Response Map), calculation formula is as follows:
Wherein,Indicate the line segment of a direction in 8 directions, length is taken as the 1/ of the smaller value of picture traverse and height
30, i.e. 1/30*min (h, w);* convolution algorithm is indicated;
Then, G is chosen in 8 directionsiMaximum value as map of magnitudes CiAs a result, formula is as follows:
Wherein, p indicates pixel, CiIndicate the map of magnitudes on the direction i;
Finally, in map of magnitudes CiAnd line segmentOn the basis of, it is calculated pencil stroke pattern S (Pencil Stroke), it is public
Formula is as follows:
A kind of portraiture photography posture recommended method of the invention, the preparation method of pencil tone illustration (Pencil Tone) is such as
Under:
Firstly, original image is converted to grayscale image I;
Then, on the basis of grayscale image I, pixel value is divided into three classes, classifying rules is as follows:
Bright pixels collection P1The pixel between 171 to 255 is in for gray value;
Soft color pixel collection P2The pixel between 86 to 170 is in for gray value;
Dark-coloured set of pixels P3The pixel between 0 to 85 is in for gray value
Then, it is fitted bright pixels collection P respectively using the parameter set A and presetting probability Distribution Model that learn in advance1, soft
Color pixel collection P2With dark-coloured set of pixels P3, fitting result is respectively P '1,P’2,P’3;
Then, in fitting result P '1,P’2,P’3On the basis of, general tint J is fitted using the parameter set B learnt in advance;
And then, on the basis of preset pencil sketch base map H, final fitting factor-beta is calculated*, calculation formula is as follows:
Wherein, J is general tint, and λ is fixed value 0.2
Finally, calculating pencil tone illustration T, calculation formula is as follows:
A kind of portraiture photography posture recommended method of the invention, presetting probability Distribution Model, the parameter set A learnt in advance
The parameter set B learnt in advance is respectively as follows:
Presetting probability Distribution Model:
General tint J is a linear regression model (LRM), and formula is as follows:
Wherein p 'iIt (v) is respectively to bright pixels collection P1, soft color pixel collection P2With dark-coloured set of pixels P3Fitting result, ωi
Parameter set B, v to learn in advance indicate pixel value;
Bright pixels collection p '1(v) it is a laplacian distribution, formula is as follows:
Wherein, σbThe parameter set A, v for belonging to pre- study indicate pixel value;
Soft color pixel collection p'2(v) it is one to be uniformly distributed, formula is as follows:
Wherein, ub, uaThe parameter set A, v for belonging to pre- study indicate pixel value;
Dark-coloured set of pixels p'3(v) it is a Gaussian Profile, formula is as follows:
Wherein, σd, μdThe parameter set A, v for belonging to pre- study indicate pixel value.
A kind of portraiture photography posture recommended method of the invention, the portrait parted pattern preparation method based on deep learning is such as
Under:
Firstly, building portrait partitioned data set, portrait partitioned data set by portrait picture and with its completely corresponding portrait
Mask (Mask) composition, wherein portrait masking-out (Mask) is by all two formed with portrait related pixel of the encirclement being accomplished manually
Value picture.Such as: all pixels for belonging to portrait are set as A class pixel, and all pixels for being not belonging to portrait are B class picture
Vegetarian refreshments, is set as 255 for A class pixel pixel value, and B class pixel pixel value is set as 0.
Then, convolutional neural networks model is built;
Finally, start the training on training set, and when loss is almost without decline, deconditioning.
Model detailed construction is as follows:
A kind of portraiture photography posture recommended method of the invention, neural network model are broadly divided into following several modules: compiling
Code module (Encoder Module), Atrous spatial pyramid sampling module (Atrous Spatial Pyramid
Pooling Module), decoder module (Decoder Module).
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, without departing from the technical principles of the invention, several improvements and modifications, these improvements and modifications can also be made
Also it should be regarded as protection scope of the present invention.
Claims (10)
1. a kind of portraiture photography posture recommended method, which comprises the following steps:
(1) before shooting is completed, reference object number first is carried out to image data and shoots the analysis of style;
(2) preset posture label is selected according to the result of analysis;
(3) the recommendation posture sketch provided in reference object according to system is completed to shoot and forms picture.
2. a kind of portraiture photography posture recommended method as described in claim 1, which is characterized in that shooting in the step (1)
The analysis of object number is divided into single player mode and collective modes according to shooting number, wherein single player mode is divided into boy student according to gender
And schoolgirl, collective modes are divided into lovers, group builds and graduating photograph.
3. a kind of portraiture photography posture recommended method as claimed in claim 2, which is characterized in that in single boy student's mode,
Posture will be recommended to be divided into show tall and big, show strong, the aobvious cruel, youth, is sedate, making fun, simple-minded, sunlight, madness;In single schoolgirl's mode
In, will recommend posture be divided into show thin in the face, aobvious stature, Gao Leng, it is lovely, show cruel, pure, charming, mature, youth, sexuality, gracefulness,
It is sunlight, not mature;In lovers' mode, posture will be recommended to be divided into tradition, prominent women, prominent male, stroll, kiss, leading along by hand, interest
Taste movement, sitting posture, shoulder to shoulder, kept for a flash, intention before some distances, kiss;It builds in group in graduating photograph mode, appearance will be recommended
State is divided into unity, intimate, vigor, love, youth, intention.
4. a kind of portraiture photography posture recommended method as described in claim 1, which is characterized in that recommend the acquisition of posture sketch
Method is as follows:
A large amount of portrait picture is collected according to each quasi-mode posture label described in claim 3 first;
Then the method by combining profile and tone generates the class pencil sketch map of image on the basis of original image;
Then by the portrait parted pattern based on deep learning, corresponding portrait masking-out is generated on the basis of original image
(Mask);
The sketch for recommending posture is generated in conjunction with the sketch map of class pencil and portrait masking-out (Mask) finally by image AND operation.
5. a kind of portraiture photography posture recommended method as claimed in claim 4, which is characterized in that the class pencil sketch map
Preparation method is as follows:
Firstly, generating the pencil stroke pattern (Pencil Stroke) for describing substantially space structure in image;
Then, in conjunction with preset pencil sketch base map, the pencil for describing the shape in image, profile, shade, light and shade is generated
Tone illustration (Pencil Tone);
Finally, generating final class in conjunction with pencil stroke pattern (Pencil stroke) and pencil tone illustration (Pencil Tone)
Pencil sketch map.
6. a kind of portraiture photography posture recommended method as claimed in claim 5, which is characterized in that the pencil stroke pattern
The preparation method of (Pencil Stroke) is as follows:
Firstly, original image is converted to grayscale image I;
Then, it on the basis of grayscale image I, calculates gradient G (Gradient);
Then, on the basis of gradient G, the response on east, the southeast, south, southwest, west, northwest, north, 8, northeast direction is calculated
Scheme (Response Map);
Then, G is chosen in 8 directionsiMaximum value as map of magnitudes CiResult;
Finally, in map of magnitudes CiWith line segment liOn the basis of, pencil stroke pattern S (Pencil Stroke) is calculated.
7. a kind of portraiture photography posture recommended method as claimed in claim 5, which is characterized in that the pencil tone illustration
The preparation method of (Pencil Tone) is as follows:
Firstly, original image is converted to grayscale image I;
Then, on the basis of grayscale image I, pixel value is divided into three classes;
Then, it is fitted bright pixels collection P respectively using the parameter set A and presetting probability Distribution Model that learn in advance1, soft colour
Element collection P2With dark-coloured set of pixels P3, fitting result is respectively P '1,P′2,P′3;
Then, in fitting result P '1,P′2,P′3On the basis of, general tint J is fitted using the parameter set B learnt in advance;
And then, on the basis of preset pencil sketch base map H, final fitting factor-beta is calculated*;
Finally, calculating pencil tone illustration T.
8. a kind of portraiture photography posture recommended method as claimed in claim 7, which is characterized in that the presetting probability point
Cloth model are as follows:
General tint J is a linear regression model (LRM);Bright pixels collection p '1It (v) is a laplacian distribution;Soft color pixel collection
p'2(v) it is one to be uniformly distributed;Dark-coloured set of pixels p'3It (v) is a Gaussian Profile.
9. a kind of portraiture photography posture recommended method as claimed in claim 4, which is characterized in that described based on deep learning
Portrait parted pattern preparation method is as follows:
Firstly, building portrait partitioned data set;
Then, convolutional neural networks model is built;
Finally, start the training on training set, and when loss is almost without decline, deconditioning.
10. a kind of portraiture photography posture recommended method as claimed in claim 9, which is characterized in that the neural network model
It is broadly divided into following several modules: coding module (Encoder Module), Atrous spatial pyramid sampling module (Atrous
Spatial Pyramid Pooling Module), decoder module (Decoder Module).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910749832.6A CN110458117A (en) | 2019-08-14 | 2019-08-14 | A kind of portraiture photography posture recommended method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910749832.6A CN110458117A (en) | 2019-08-14 | 2019-08-14 | A kind of portraiture photography posture recommended method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110458117A true CN110458117A (en) | 2019-11-15 |
Family
ID=68486478
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910749832.6A Withdrawn CN110458117A (en) | 2019-08-14 | 2019-08-14 | A kind of portraiture photography posture recommended method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110458117A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022077229A1 (en) * | 2020-10-13 | 2022-04-21 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Electric device, method of controlling electric device, and computer readable storage medium |
-
2019
- 2019-08-14 CN CN201910749832.6A patent/CN110458117A/en not_active Withdrawn
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022077229A1 (en) * | 2020-10-13 | 2022-04-21 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Electric device, method of controlling electric device, and computer readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108257084B (en) | Lightweight face automatic makeup method based on mobile terminal | |
Hold-Geoffroy et al. | A perceptual measure for deep single image camera calibration | |
Zuffi et al. | Lions and tigers and bears: Capturing non-rigid, 3d, articulated shape from images | |
CN108596024B (en) | Portrait generation method based on face structure information | |
CN107610209A (en) | Human face countenance synthesis method, device, storage medium and computer equipment | |
CN109815826B (en) | Method and device for generating face attribute model | |
CN109087379B (en) | Facial expression migration method and facial expression migration device | |
CN112037320B (en) | Image processing method, device, equipment and computer readable storage medium | |
CN112766160A (en) | Face replacement method based on multi-stage attribute encoder and attention mechanism | |
CN110738161A (en) | face image correction method based on improved generation type confrontation network | |
CN111932678B (en) | Multi-view real-time human motion, gesture, expression and texture reconstruction system | |
CN108537126B (en) | Face image processing method | |
CN107657664B (en) | Image optimization method and device after face expression synthesis, storage medium and computer equipment | |
CN110263768A (en) | A kind of face identification method based on depth residual error network | |
CN109472795A (en) | A kind of image edit method and device | |
CN104157001A (en) | Method and device for drawing head caricature | |
CN110135277B (en) | Human behavior recognition method based on convolutional neural network | |
CN109493403A (en) | A method of human face animation is realized based on moving cell Expression Mapping | |
CN102567716A (en) | Face synthetic system and implementation method | |
CN106228590B (en) | A kind of human body attitude edit methods in image | |
WO2023155533A1 (en) | Image driving method and apparatus, device and medium | |
CN112801064A (en) | Model training method, electronic device and storage medium | |
CN112669436A (en) | Deep learning sample generation method based on 3D point cloud | |
CN109472838A (en) | A kind of sketch generation method and device | |
CN113362422A (en) | Shadow robust makeup transfer system and method based on decoupling representation |
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 | ||
WW01 | Invention patent application withdrawn after publication | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20191115 |