CN110503601A - Face based on confrontation network generates picture replacement method and system - Google Patents
Face based on confrontation network generates picture replacement method and system Download PDFInfo
- Publication number
- CN110503601A CN110503601A CN201910802427.6A CN201910802427A CN110503601A CN 110503601 A CN110503601 A CN 110503601A CN 201910802427 A CN201910802427 A CN 201910802427A CN 110503601 A CN110503601 A CN 110503601A
- Authority
- CN
- China
- Prior art keywords
- face
- picture
- data
- module
- confrontation network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 18
- 239000000284 extract Substances 0.000 claims abstract description 10
- 238000012805 post-processing Methods 0.000 claims abstract description 8
- 230000004927 fusion Effects 0.000 claims description 15
- 230000009466 transformation Effects 0.000 claims description 11
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 9
- 238000002156 mixing Methods 0.000 claims description 9
- 238000012937 correction Methods 0.000 claims description 7
- 238000013480 data collection Methods 0.000 claims description 6
- 230000002708 enhancing effect Effects 0.000 claims description 6
- 241001269238 Data Species 0.000 claims description 3
- 238000002844 melting Methods 0.000 claims description 3
- 230000008018 melting Effects 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 3
- 238000004064 recycling Methods 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000008485 antagonism Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000001815 facial effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000002910 structure generation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/02—Affine transformations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of faces based on confrontation network to generate picture replacement method and system, comprising: the pictures for establishing personage A and B identify and extract face data, and in training process, A and B share an encoder, possess respective decoder and arbiter.The face data of A and B is compressed to by latent space by encoder first, then restores original face data by decoder, while constraining the reduction and generation of face using arbiter.Trained face characteristic model is finally obtained, includes five contents: encoder, decoder a, decoder b, arbiter a, arbiter b.To the face picture to be replaced of input, face data is intercepted, the translation for recycling the decoder of other side to carry out target face generates, and then realizes merging for face and original image by post-processing, finally obtains replacement result.The present invention has filled up the blank that face replacement is carried out in conjunction with self-encoding encoder and production confrontation network, and the face of multiresolution is supported to generate.
Description
Technical field
The present invention relates to Digital Image Processing and artificial intelligence technology crossing domains, and in particular, to one kind is based on confrontation
The face of network generates picture replacement method and system.
Background technique
Face replacement in picture refers to the human face region replaced in Target Photo with the human face region of source picture.It is mature
Face replacement technology there are a large amount of application scenarios, set in communication amusement, virtual reality, secret protection, production of film and TV, plane
Meter etc. has unique advantage.Traditional face replacement only carries out cutting physically simply by media editing tool
It cuts out and modifies, it is difficult to realize fusion naturally replacement effect, while learning cost is higher, is difficult to accomplish to popularize, therefore fail
To being widely applied；Later machine learning is developed, and be thus born face general key point detection model, to a certain extent
The automation of face replacement is realized, user adjusts without the details scratched figure and only focus on the colour of skin and edge；Nowadays due to depth
The development of study, and the rise of the technologies such as the autocoder based on deep learning and production confrontation network, further push away
The picture face replacement automation that self-supervisory is realized under Weakly supervised sample is moved.By the retrieval to existing face replacement technology
It was found that a kind of face replacement method of patent notes and device, publication date that China Patent Publication No. is CN106875329A are
On June 20th, 2017.The technology passes through 64 characteristic points of calibration source face and target face, is then based on Delaunay triangle
Net carries out region division and matching.Finally according to matching result, face image data to be replaced is mapped to target face and defeated
Out.This method may be implemented preferably to replace effect in the situation similar in source face and target facial angle, expression, illumination
Fruit, but when difference is excessive, then hardly result in ideal replacement effect.Generalization ability is poor simultaneously, not for same personage
Same face picture, when replacement, will carry out matching and map operation every time.China Patent Publication No. is the special of CN107316020A
Benefit describes a kind of face replacement method based on threedimensional model, and publication date is on November 3rd, 2017.The technology initially sets up mesh
The three-dimensional head model for marking personage, is rendered into the background image of removal source personage's head zone, guarantees the complete of head
Replacement.The disadvantage is that requirement of the foundation of threedimensional model for sample data is very stringent, the training stage will need to expend more
Calculation amount, while being the replacement to entire head, without retaining original features such as hair, face contour.It is same currently without discovery
The explanation or report of similar techniques of the present invention are also not yet collected into data similar both at home and abroad.
Summary of the invention
For the defects in the prior art, the object of the present invention is to provide a kind of faces based on confrontation network to generate picture
Replacement method and system.
A kind of face based on confrontation network provided according to the present invention generates picture replacement method, includes the following steps:
Step 1, the image data collection comprising face of personage A and B are obtained；
Step 2, data set is cleaned by MTCNN, extracts face data；
Step 3, after carrying out data enhancing at random, the sample data that network is fought when generating for training is obtained；
Step 4, confrontation network model is generated with pretreated face picture sample training, obtained raw for realizing face
At generation fight network N；
Step 5, the extraction that face is carried out to the picture to be replaced inputted in real time carries out face alignment by affine transformation,
Obtain face picture to be replaced；
Step 6, it is loaded into the trained network N of step 3, face picture to be replaced obtained in step 4 is input to the net
Network generates the face picture of target face；
Step 7, for the face picture of target face obtained in step 5, by the picture blending algorithm of post-processing into
The fusion of row face picture and original image；
Step 8, picture to be replaced is by step 4 to step 6, so that it may complete melting between the generation and picture of target face
It closes, the face picture that final output has been replaced.
Preferably, the step 2 includes the following steps:
Step 2.1, with the image data in MTCNN identification step 1, if returning to zero or two or more faces are waited
Frame is selected, then leaves out such data；
Step 2.2, the individual human face candidate frame that MTCNN is returned, wide and Gao Ruo have one less than 64, then leave out；
Step 2.3, for other satisfactory image datas, the extraction of face is carried out by individual human face candidate frame,
Only retain the face portion of picture.
Preferably, the step 3 includes the following steps:
Step 3.1, the probability for having 30% will do it three kinds before entering model training by the face picture data of acquisition
Data enhancement method；
Step 3.2, for the picture chosen, rotation is realized by the cv2.getRotationMatrix2D function of OpenCV
Transformation is changed, while having the probability of half to will do it doubling；
Step 3.3, for the picture chosen, the color data of an other picture in the picture and pictures is calculated, with
One random ratio carries out color mixing；
Step 3.4, for the picture chosen, a motion blur core is generated in range limiting, then linearly roll up by core
Product adds motion blur to picture.
Preferably, in the step 4, personage A and B share one and same coding device, there is oneself decoder and differentiation respectively
Device, wherein encoder and decoder collectively constitute the generator for generating confrontation network, for generating face and mask.
Preferably, the step 5 includes the following steps:
Step 5.1, MTCNN identifies face and extracts face data, five characteristic points of label；
Step 5.2, affine transformation is carried out by the position of five characteristic points, realizes the alignment of face.
Preferably, the step 6 includes the following steps:
Step 6.1, it is loaded into trained generation and fights network；
Step 6.2, the picture after face being aligned is input to trained network, calls the decoder of other side, output life
At target face and mask.
Preferably, the step 7 includes the following steps:
Step 7.1, the face of generation and mask data are merged, so that five features is more prominent；
Step 7.2, fused face may have larger color difference with original image, by histogram matching to fusion
Face afterwards carries out colour correction, so that the colour of skin is close；
Step 7.3, the face picture after color correction is covered into the identical position of original image；
Step 7.4, Gaussian Blur is carried out to the edge of face, so that edge transition is more smooth, merged more natural.
A kind of face based on confrontation network provided according to the present invention generates picture replacement system, including following module:
Module M1 obtains the image data collection comprising face of personage A and B；
Module M2 cleans data set by MTCNN, extracts face data；
Module M3 obtains the sample data of confrontation network when generating for training after random progress data enhancing；
Module M4 generates confrontation network model with pretreated face picture sample training, obtains for realizing face
The generation of generation fights network N；
Module M5 carries out the extraction of face to the picture to be replaced inputted in real time, carries out face alignment by affine transformation,
Obtain face picture to be replaced；
The trained network N of module M6, the M3 that insmods, is input to this for face picture to be replaced obtained in module M4
Network generates the face picture of target face；
Module M7 passes through the picture blending algorithm of post-processing for the face picture of target face obtained in module M5
Carry out the fusion of face picture and original image；
Module M8, picture to be replaced pass through module M4 to module M6, so that it may between the generation and picture of completing target face
Fusion, the face picture that final output has been replaced.
Compared with prior art, the present invention have it is following the utility model has the advantages that
1, it is raw in image to have incorporated generation confrontation network on the basis of the self-encoding encoder of deep learning generates model by the present invention
At the good characteristic of aspect, proposes the face based on confrontation network and generate picture replacement method.In human face structure generation side
Face introduces antagonism loss, multiresolution generates aspect introducing dynamic network structural support, color is added in picture later period fusion aspect
Color correction and mask fusion have larger compared with face replacement method before in terms of face is generated with picture fusion
Promotion.
2, the present invention has filled up the blank for carrying out the relevant patent of face replacement using confrontation network is generated, and face generates
Ability is strong, and picture syncretizing effect is naturally, support the face of more kinds of resolution ratio of 64*64,128*128,256*256 to generate.
Detailed description of the invention
Upon reading the detailed description of non-limiting embodiments with reference to the following drawings, other feature of the invention,
Objects and advantages will become more apparent upon:
Fig. 1 is that the present invention is based on the model framework figures that the face of confrontation network generates picture replacement method and system.
Fig. 2 is that generation constructed by this method fights network overall schematic.
Fig. 3 is multiresolution generating principle schematic diagram.
Fig. 4 is the post-processing schematic diagram that face generates
Specific embodiment
The present invention is described in detail combined with specific embodiments below.Following embodiment will be helpful to the technology of this field
Personnel further understand the present invention, but the invention is not limited in any way.It should be pointed out that the ordinary skill of this field
For personnel, without departing from the inventive concept of the premise, several changes and improvements can also be made.These belong to the present invention
Protection scope.
As shown in Figure 1 to Figure 3, a kind of face based on confrontation network provided according to the present invention generates picture replacement side
Method includes the following steps:
Step 1, the image data collection comprising face of personage A and B are obtained；
Step 2, data set is cleaned by MTCNN, extracts face data；
Step 3, after carrying out data enhancing at random, the sample data that network is fought when generating for training is obtained；
Step 4, confrontation network model is generated with pretreated face picture sample training, obtained raw for realizing face
At generation fight network N；
Step 5, the extraction that face is carried out to the picture to be replaced inputted in real time carries out face alignment by affine transformation,
Obtain face picture to be replaced；
Step 6, it is loaded into the trained network N of step 3, face picture to be replaced obtained in step 4 is input to the net
Network generates the face picture of target face；
Step 7, for the face picture of target face obtained in step 5, by the picture blending algorithm of post-processing into
The fusion of row face picture and original image；
Step 8, picture to be replaced is by step 4 to step 6, so that it may complete melting between the generation and picture of target face
It closes, the face picture that final output has been replaced.
The step 2 includes the following steps:
Step 2.1, with the image data in MTCNN identification step 1, if returning to zero or two or more faces are waited
Frame is selected, then leaves out such data；
Step 2.2, the individual human face candidate frame that MTCNN is returned, wide and Gao Ruo have one less than 64, then leave out；
Step 2.3, for other satisfactory image datas, the extraction of face is carried out by individual human face candidate frame,
Only retain the face portion of picture.
The step 3 includes the following steps:
Step 3.1, the probability for having 30% will do it three kinds before entering model training by the face picture data of acquisition
Data enhancement method；
Step 3.2, for the picture chosen, rotation is realized by the cv2.getRotationMatrix2D function of OpenCV
Transformation is changed, while having the probability of half to will do it doubling；
Step 3.3, for the picture chosen, the color data of an other picture in the picture and pictures is calculated,
Color mixing is carried out with a random ratio；
Step 3.4, for the picture chosen, a motion blur core is generated in range limiting, then linearly roll up by core
Product adds motion blur to picture.
As shown in Fig. 2, personage A and B share one and same coding device in the step 4, there is the decoder of oneself respectively and sentence
Other device, wherein encoder and decoder collectively constitute the generator for generating confrontation network, for generating face and mask.Such as Fig. 3
Shown, this algorithm realizes training front and back using the combination of convolutional layer and up-sampling according to the size difference of input and output picture
Dimension and size alignment, therefore the image of the higher resolutions such as 128*128,256*256 can be generated, has more wide
Application space.
The step 5 includes the following steps:
Step 5.1, MTCNN identifies face and extracts face data, five characteristic points of label；
Step 5.2, affine transformation is carried out by the position of five characteristic points, realizes the alignment of face.
The step 6 includes the following steps:
Step 6.1, it is loaded into trained generation and fights network；
Step 6.2, the picture after face being aligned is input to trained network, calls the decoder of other side, output life
At target face and mask.
As shown in figure 4, the step 7 includes the following steps:
Step 7.1, the face of generation and mask data are merged, so that five features is more prominent；
Step 7.2, fused face may have larger color difference with original image, by histogram matching to fusion
Face afterwards carries out colour correction, so that the colour of skin is close；
Step 7.3, the face picture after color correction is covered into the identical position of original image；
Step 7.4, Gaussian Blur is carried out to the edge of face, so that edge transition is more smooth, merged more natural.
A kind of face based on confrontation network provided according to the present invention generates picture replacement system, including following module:
Module M1 obtains the image data collection comprising face of personage A and B；
Module M2 cleans data set by MTCNN, extracts face data；
Module M3 obtains the sample data of confrontation network when generating for training after random progress data enhancing；
Module M4 generates confrontation network model with pretreated face picture sample training, obtains for realizing face
The generation of generation fights network N；
Module M5 carries out the extraction of face to the picture to be replaced inputted in real time, carries out face alignment by affine transformation,
Obtain face picture to be replaced；
The trained network N of module M6, the M3 that insmods, is input to this for face picture to be replaced obtained in module M4
Network generates the face picture of target face；
Module M7 passes through the picture blending algorithm of post-processing for the face picture of target face obtained in module M5
Carry out the fusion of face picture and original image；
Module M8, picture to be replaced pass through module M4 to module M6, so that it may between the generation and picture of completing target face
Fusion, the face picture that final output has been replaced.
One skilled in the art will appreciate that in addition to realizing system provided by the invention in a manner of pure computer readable program code
It, completely can be by the way that method and step be carried out programming in logic come so that the present invention provides and its other than each device, module, unit
System and its each device, module, unit with logic gate, switch, specific integrated circuit, programmable logic controller (PLC) and embedding
Enter the form of the controller that declines etc. to realize identical function.So system provided by the invention and its every device, module, list
Member is considered a kind of hardware component, and to include in it can also for realizing the device of various functions, module, unit
To be considered as the structure in hardware component；It can also will be considered as realizing the device of various functions, module, unit either real
The software module of existing method can be the structure in hardware component again.
In the description of the present application, it is to be understood that term " on ", "front", "rear", "left", "right", " is erected at "lower"
Directly ", the orientation or positional relationship of the instructions such as "horizontal", "top", "bottom", "inner", "outside" is orientation based on the figure or position
Relationship is set, description the application is merely for convenience of and simplifies description, rather than the device or element of indication or suggestion meaning are necessary
It with specific orientation, is constructed and operated in a specific orientation, therefore should not be understood as the limitation to the application.
Specific embodiments of the present invention are described above.It is to be appreciated that the invention is not limited to above-mentioned
Particular implementation, those skilled in the art can make a variety of changes or modify within the scope of the claims, this not shadow
Ring substantive content of the invention.In the absence of conflict, the feature in embodiments herein and embodiment can any phase
Mutually combination.
Claims (8)
1. a kind of face based on confrontation network generates picture replacement method, which comprises the steps of:
Step 1, the image data collection comprising face of personage A and B are obtained；
Step 2, data set is cleaned by MTCNN, extracts face data；
Step 3, after carrying out data enhancing at random, the sample data that network is fought when generating for training is obtained；
Step 4, confrontation network model is generated with pretreated face picture sample training, obtained for realizing face generation
Generate confrontation network N；
Step 5, the extraction that face is carried out to the picture to be replaced inputted in real time carries out face alignment by affine transformation, obtains
Face picture to be replaced；
Step 6, it is loaded into the trained network N of step 3, face picture to be replaced obtained in step 4 is input to the network, it is raw
At the face picture of target face；
Step 7, for the face picture of target face obtained in step 5, face is carried out by the picture blending algorithm of post-processing
The fusion of portion's picture and original image；
Step 8, picture to be replaced is by step 4 to step 6, so that it may the fusion between the generation and picture of target face is completed,
The face picture that final output has been replaced.
2. the face according to claim 1 based on confrontation network generates picture replacement method, which is characterized in that the step
Rapid 2 include the following steps:
Step 2.1, with the image data in MTCNN identification step 1, if returning to zero or two or more face candidates
Frame then leaves out such data；
Step 2.2, the individual human face candidate frame that MTCNN is returned, wide and Gao Ruo have one less than 64, then leave out；
Step 2.3, for other satisfactory image datas, the extraction of face is carried out by individual human face candidate frame, is only protected
Stay the face portion of picture.
3. the face according to claim 1 based on confrontation network generates picture replacement method, which is characterized in that the step
Rapid 3 include the following steps:
Step 3.1, the probability for having 30% will do it three kinds of data before entering model training by the face picture data of acquisition
Enhancement method；
Step 3.2, for the picture chosen, realize that rotation becomes by the cv2.getRotationMatrix2D function of OpenCV
It changes, while thering is the probability of half to will do it doubling；
Step 3.3, for the picture chosen, the color data of an other picture in the picture and pictures is calculated, with one
Random ratio carries out color mixing；
Step 3.4, for the picture chosen, a motion blur core is generated in range limiting, then pass through core linear convolution pair
Picture adds motion blur.
4. the face according to claim 1 based on confrontation network generates picture replacement method, which is characterized in that the step
In rapid 4, personage A and B share one and same coding device, have oneself decoder and arbiter respectively, wherein encoder and decoder
The generator for generating confrontation network is collectively constituted, for generating face and mask.
5. the face according to claim 1 based on confrontation network generates picture replacement method, which is characterized in that the step
Rapid 5 include the following steps:
Step 5.1, MTCNN identifies face and extracts face data, five characteristic points of label；
Step 5.2, affine transformation is carried out by the position of five characteristic points, realizes the alignment of face.
6. the face according to claim 1 based on confrontation network generates picture replacement method, which is characterized in that the step
Rapid 6 include the following steps:
Step 6.1, it is loaded into trained generation and fights network；
Step 6.2, the picture after face being aligned is input to trained network, calls the decoder of other side, exports generation
Target face and mask.
7. the face according to claim 1 based on confrontation network generates picture replacement method, which is characterized in that the step
Rapid 7 include the following steps:
Step 7.1, the face of generation and mask data are merged, so that five features is more prominent；
Step 7.2, fused face may have larger color difference with original image, by histogram matching to fused
Face carries out colour correction, so that the colour of skin is close；
Step 7.3, the face picture after color correction is covered into the identical position of original image；
Step 7.4, Gaussian Blur is carried out to the edge of face, so that edge transition is more smooth, merged more natural.
8. a kind of face based on confrontation network generates picture replacement system, which is characterized in that including following module:
Module M1 obtains the image data collection comprising face of personage A and B；
Module M2 cleans data set by MTCNN, extracts face data；
Module M3 obtains the sample data of confrontation network when generating for training after random progress data enhancing；
Module M4 generates confrontation network model with pretreated face picture sample training, obtains for realizing face generation
Generation fight network N；
Module M5 carries out the extraction of face to the picture to be replaced inputted in real time, carries out face alignment by affine transformation, obtains
Face picture to be replaced；
The trained network N of module M6, the M3 that insmods, is input to the net for face picture to be replaced obtained in module M4
Network generates the face picture of target face；
Module M7 carries out the face picture of target face obtained in module M5 by the picture blending algorithm of post-processing
The fusion of face picture and original image；
Module M8, picture to be replaced pass through module M4 to module M6, so that it may complete melting between the generation and picture of target face
It closes, the face picture that final output has been replaced.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910802427.6A CN110503601A (en) | 2019-08-28 | 2019-08-28 | Face based on confrontation network generates picture replacement method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910802427.6A CN110503601A (en) | 2019-08-28 | 2019-08-28 | Face based on confrontation network generates picture replacement method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110503601A true CN110503601A (en) | 2019-11-26 |
Family
ID=68590006
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910802427.6A Pending CN110503601A (en) | 2019-08-28 | 2019-08-28 | Face based on confrontation network generates picture replacement method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110503601A (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110889381A (en) * | 2019-11-29 | 2020-03-17 | 广州华多网络科技有限公司 | Face changing method and device, electronic equipment and storage medium |
| CN111242837A (en) * | 2020-01-03 | 2020-06-05 | 杭州电子科技大学 | Face anonymous privacy protection method based on generation of countermeasure network |
| CN111260545A (en) * | 2020-01-20 | 2020-06-09 | 北京百度网讯科技有限公司 | Method and device for generating image |
| CN111275784A (en) * | 2020-01-20 | 2020-06-12 | 北京百度网讯科技有限公司 | Method and device for generating image |
| CN111368796A (en) * | 2020-03-20 | 2020-07-03 | 北京达佳互联信息技术有限公司 | Face image processing method and device, electronic equipment and storage medium |
| CN111402118A (en) * | 2020-03-17 | 2020-07-10 | 腾讯科技（深圳）有限公司 | Image replacement method and device, computer equipment and storage medium |
| CN111563868A (en) * | 2020-07-20 | 2020-08-21 | 腾讯科技（深圳）有限公司 | Artificial intelligence-based head decoration processing method and device |
| CN111598818A (en) * | 2020-04-17 | 2020-08-28 | 北京百度网讯科技有限公司 | Face fusion model training method and device and electronic equipment |
| CN111783647A (en) * | 2020-06-30 | 2020-10-16 | 北京百度网讯科技有限公司 | Training method of face fusion model, face fusion method, device and equipment |
| CN111931148A (en) * | 2020-07-31 | 2020-11-13 | 支付宝(杭州)信息技术有限公司 | Image processing method and device and electronic equipment |
| CN112258388A (en) * | 2020-11-02 | 2021-01-22 | 公安部第三研究所 | Public security view desensitization test data generation method, system and storage medium |
| CN112712138A (en) * | 2021-01-19 | 2021-04-27 | 腾讯科技（深圳）有限公司 | Image processing method, device, equipment and storage medium |
| CN113962845A (en) * | 2021-08-25 | 2022-01-21 | 北京百度网讯科技有限公司 | Image processing method, image processing apparatus, electronic device, and storage medium |
| CN113989868A (en) * | 2020-07-10 | 2022-01-28 | 广州数字空间网络科技有限公司 | Face data exchange method |
| TWI755768B (en) * | 2019-11-29 | 2022-02-21 | 大陸商北京市商湯科技開發有限公司 | Image processing method, image processing device and storage medium thereof |
| CN114187165A (en) * | 2021-11-09 | 2022-03-15 | 阿里巴巴云计算(北京)有限公司 | Image processing method and device |
| CN114973382A (en) * | 2022-06-15 | 2022-08-30 | 平安科技（深圳）有限公司 | Human face replacement method, device, equipment and storage medium based on artificial intelligence |
| CN117611460A (en) * | 2023-11-10 | 2024-02-27 | 深圳市鹏中科技有限公司 | Face image fusion method, device, equipment and storage medium |
| CN117710502A (en) * | 2023-12-12 | 2024-03-15 | 摩尔线程智能科技(北京)有限责任公司 | Rendering method, rendering device and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107577985A (en) * | 2017-07-18 | 2018-01-12 | 南京邮电大学 | The implementation method of the face head portrait cartooning of confrontation network is generated based on circulation |
| CN110084121A (en) * | 2019-03-27 | 2019-08-02 | 南京邮电大学 | Implementation method based on the human face expression migration for composing normalized circulation production confrontation network |
| CN110148081A (en) * | 2019-03-25 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Training method, image processing method, device and the storage medium of image processing model |
-
2019
- 2019-08-28 CN CN201910802427.6A patent/CN110503601A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107577985A (en) * | 2017-07-18 | 2018-01-12 | 南京邮电大学 | The implementation method of the face head portrait cartooning of confrontation network is generated based on circulation |
| CN110148081A (en) * | 2019-03-25 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Training method, image processing method, device and the storage medium of image processing model |
| CN110084121A (en) * | 2019-03-27 | 2019-08-02 | 南京邮电大学 | Implementation method based on the human face expression migration for composing normalized circulation production confrontation network |
Non-Patent Citations (2)
| Title |
|---|
| GUO YH，KE X AND MA J: "A face replacement netural network for image and video", 《ACM》 * |
| 汪亚楠: "基于卷积神经网络的电商图像识别研究", 《中国优秀硕士学位论文全文数据库 社会科学II辑》 * |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110889381A (en) * | 2019-11-29 | 2020-03-17 | 广州华多网络科技有限公司 | Face changing method and device, electronic equipment and storage medium |
| TWI755768B (en) * | 2019-11-29 | 2022-02-21 | 大陸商北京市商湯科技開發有限公司 | Image processing method, image processing device and storage medium thereof |
| CN110889381B (en) * | 2019-11-29 | 2022-12-02 | 广州方硅信息技术有限公司 | Face changing method and device, electronic equipment and storage medium |
| CN111242837A (en) * | 2020-01-03 | 2020-06-05 | 杭州电子科技大学 | Face anonymous privacy protection method based on generation of countermeasure network |
| CN111242837B (en) * | 2020-01-03 | 2023-05-12 | 杭州电子科技大学 | Face anonymity privacy protection method based on generation countermeasure network |
| CN111275784A (en) * | 2020-01-20 | 2020-06-12 | 北京百度网讯科技有限公司 | Method and device for generating image |
| CN111260545A (en) * | 2020-01-20 | 2020-06-09 | 北京百度网讯科技有限公司 | Method and device for generating image |
| CN111402118A (en) * | 2020-03-17 | 2020-07-10 | 腾讯科技（深圳）有限公司 | Image replacement method and device, computer equipment and storage medium |
| CN111368796B (en) * | 2020-03-20 | 2024-03-08 | 北京达佳互联信息技术有限公司 | Face image processing method and device, electronic equipment and storage medium |
| CN111368796A (en) * | 2020-03-20 | 2020-07-03 | 北京达佳互联信息技术有限公司 | Face image processing method and device, electronic equipment and storage medium |
| CN111598818A (en) * | 2020-04-17 | 2020-08-28 | 北京百度网讯科技有限公司 | Face fusion model training method and device and electronic equipment |
| US11830288B2 (en) | 2020-04-17 | 2023-11-28 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for training face fusion model and electronic device |
| CN111783647A (en) * | 2020-06-30 | 2020-10-16 | 北京百度网讯科技有限公司 | Training method of face fusion model, face fusion method, device and equipment |
| CN111783647B (en) * | 2020-06-30 | 2023-11-03 | 北京百度网讯科技有限公司 | Training method of face fusion model, face fusion method, device and equipment |
| CN113989868A (en) * | 2020-07-10 | 2022-01-28 | 广州数字空间网络科技有限公司 | Face data exchange method |
| CN111563868A (en) * | 2020-07-20 | 2020-08-21 | 腾讯科技（深圳）有限公司 | Artificial intelligence-based head decoration processing method and device |
| CN111931148A (en) * | 2020-07-31 | 2020-11-13 | 支付宝(杭州)信息技术有限公司 | Image processing method and device and electronic equipment |
| CN112258388A (en) * | 2020-11-02 | 2021-01-22 | 公安部第三研究所 | Public security view desensitization test data generation method, system and storage medium |
| CN112712138B (en) * | 2021-01-19 | 2022-05-20 | 腾讯科技（深圳）有限公司 | Image processing method, device, equipment and storage medium |
| CN112712138A (en) * | 2021-01-19 | 2021-04-27 | 腾讯科技（深圳）有限公司 | Image processing method, device, equipment and storage medium |
| CN113962845A (en) * | 2021-08-25 | 2022-01-21 | 北京百度网讯科技有限公司 | Image processing method, image processing apparatus, electronic device, and storage medium |
| CN113962845B (en) * | 2021-08-25 | 2023-08-29 | 北京百度网讯科技有限公司 | Image processing method, image processing apparatus, electronic device, and storage medium |
| CN114187165A (en) * | 2021-11-09 | 2022-03-15 | 阿里巴巴云计算(北京)有限公司 | Image processing method and device |
| CN114973382A (en) * | 2022-06-15 | 2022-08-30 | 平安科技（深圳）有限公司 | Human face replacement method, device, equipment and storage medium based on artificial intelligence |
| CN114973382B (en) * | 2022-06-15 | 2024-07-19 | 平安科技（深圳）有限公司 | Artificial intelligence-based face replacement method, device, equipment and storage medium |
| CN117611460A (en) * | 2023-11-10 | 2024-02-27 | 深圳市鹏中科技有限公司 | Face image fusion method, device, equipment and storage medium |
| CN117710502A (en) * | 2023-12-12 | 2024-03-15 | 摩尔线程智能科技(北京)有限责任公司 | Rendering method, rendering device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110503601A (en) | Face based on confrontation network generates picture replacement method and system | |
| CN105184253B (en) | Face recognition method and face recognition system | |
| KR101221451B1 (en) | Methodlogy of animatable digital clone creation from multi-view images capturing dynamic performance | |
| Liu et al. | Psgan++: Robust detail-preserving makeup transfer and removal | |
| CN108573222A (en) | The pedestrian image occlusion detection method for generating network is fought based on cycle | |
| Rabby et al. | BeyondPixels: A comprehensive review of the evolution of neural radiance fields | |
| CN116385606A (en) | Speech signal driven personalized three-dimensional face animation generation method and application thereof | |
| Liu et al. | Spatial-aware texture transformer for high-fidelity garment transfer | |
| Zha et al. | Towards compact 3d representations via point feature enhancement masked autoencoders | |
| Wang et al. | PACCDU: Pyramid attention cross-convolutional dual UNet for infrared and visible image fusion | |
| Huang et al. | Object-occluded human shape and pose estimation with probabilistic latent consistency | |
| Yan et al. | Video face swap based on autoencoder generation network | |
| CN114782460B (en) | Image segmentation model generation method, image segmentation method and computer equipment | |
| CN110889854B (en) | Sketch part segmentation method, system, device and storage medium based on multi-scale deep learning | |
| CN117218246A (en) | Training method and device for image generation model, electronic equipment and storage medium | |
| CN114120068A (en) | Image processing method, image processing device, electronic equipment, storage medium and computer product | |
| Singh | Future of Animated Narrative and the Effects of Ai on Conventional Animation Techniques | |
| Singh et al. | Deepfake as an Artificial Intelligence tool for VFX Films | |
| CN110473276A (en) | A kind of high efficiency three-dimensional cartoon production method | |
| CN116681579A (en) | Real-time video face replacement method, medium and system | |
| Chen et al. | Cantonese porcelain image generation using user-guided generative adversarial networks | |
| Yang et al. | Shapeediter: a stylegan encoder for face swapping | |
| Vasiliu et al. | Coherent rendering of virtual smile previews with fast neural style transfer | |
| Zhong et al. | Implicit epipolar geometric function based light field continuous angular representation | |
| Wang et al. | Expression-aware neural radiance fields for high-fidelity talking portrait synthesis |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191126 |