CN108460720A - A method of changing image style based on confrontation network model is generated - Google Patents
A method of changing image style based on confrontation network model is generated Download PDFInfo
- Publication number
- CN108460720A CN108460720A CN201810101716.9A CN201810101716A CN108460720A CN 108460720 A CN108460720 A CN 108460720A CN 201810101716 A CN201810101716 A CN 201810101716A CN 108460720 A CN108460720 A CN 108460720A
- Authority
- CN
- China
- Prior art keywords
- image
- style
- generator
- network model
- data distribution
- 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 30
- 230000008569 process Effects 0.000 claims abstract description 13
- 230000008859 change Effects 0.000 claims abstract description 11
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 7
- 238000013528 artificial neural network Methods 0.000 claims abstract description 5
- 238000010276 construction Methods 0.000 claims abstract description 5
- 230000004069 differentiation Effects 0.000 claims description 3
- 239000004744 fabric Substances 0.000 claims 1
- 230000006870 function Effects 0.000 abstract description 10
- 230000000694 effects Effects 0.000 abstract description 6
- 238000013135 deep learning Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G06T3/04—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Abstract
The invention discloses a kind of based on the method for generating confrontation network model change image style, belongs to deep learning field of neural networks, includes the following steps：S1, construction are originally generated confrontation network model, generate image by generator and are input to arbiter progress network training；S2, constructing neural network serve as the function of generator and arbiter；S3, initialization random noise, input in generator；S4, the style data distribution that image data is concentrated is obtained using convolutional neural networks；S5, style data distribution is added during generator each iteration, is trained.This method structure is originally generated confrontation network model based on change image style, by the style and features for learning image data set, the data distribution of image style is obtained, the data distribution of image style is added in the training process of generator, to achieve the effect that change image style information.
Description
Technical field
The present invention relates to deep learning nerual network technique fields, and in particular to one kind is changed based on confrontation network model is generated
Become the method for image style.
Background technology
It is by Goodfellow that production, which fights network (Generative Adversarial Network, abbreviation GAN),
The deep learning frame proposed in 2014, it is based on the thought of " game theory ", construction generator (generator) and arbiter
(discriminator) two kinds of models, the former generates image by the Uniform noise or gaussian random noise of input (0,1), after
Person differentiates the image of input, and determination is image from data set or the image generated by generator.
In traditional confrontation network model, arbiter only distinguishes the image data set of input and generation image,
There is no the features of study image, and also without the data distribution of acquisition image style, and generator is also in the training process
The data distribution of data set is trained, style is consistent.At present in technology, network mould is fought not over being originally generated
Type learns the style and features of image data set, realizes the application for changing image style.
Invention content
The purpose of the present invention is to solve drawbacks described above in the prior art, provide a kind of based on generation confrontation network mould
The method that type changes image style.
The purpose of the present invention can be reached by adopting the following technical scheme that：
A method of changing image style based on confrontation network model is generated, the method includes the following steps：
S1, construction are originally generated confrontation network model, generate image by generator and are input to arbiter progress network
Training；
S2, constructing neural network serve as the function of generator and arbiter；
S3, initialization random noise, input in generator；
S4, the data distribution that the style information that image data is concentrated is obtained using convolutional neural networks；
S5, the data distribution that style information is added during generator each iteration, update loss function, after progress
Continuous training.
Further, the step S4 processes are as follows：
S41, the style information that image data is concentrated is extracted by convolutional neural networks；
S42, the data distribution for obtaining style information.
Further, the step S5 processes are as follows：
During generator each iteration, style letter is added on the basis of generator generates the data distribution of image
The data distribution of breath updates loss function, is subsequently trained.
Further, the expression formula of the loss function is：
Wherein, D (x) indicates that differentiation of the arbiter to image, pr indicate that the distribution of data images, pg indicate to generate image
Distribution, λ is hyper parameter,For gradient.
The present invention has the following advantages and effects with respect to the prior art：
The style information that the present invention is concentrated according to the operating process for changing image style, by extracting image data, obtains
The data distribution of style information, is input in generator, can achieve the effect that change image style information intensity, has good
Specific aim.
Description of the drawings
Fig. 1 is the overall flow figure for being originally generated confrontation network in the present invention and being trained by changing image style.
Specific implementation mode
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
The every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
Embodiment
As shown in Fig. 1, the method that network model changes image style is fought based on generation present embodiment discloses a kind of,
Specifically include the following steps：
Step S1, construction is originally generated confrontation network model, generates image by generator and is input to arbiter progress
Network training.
Step S2, constructing neural network serves as the function of generator and arbiter；
Different convolution kernels is embodied in the difference of matrix numerical value, the difference of ranks number.
Multiple convolution kernels are constructed, during handling image, different convolution kernels is meant in network training
Different characteristic of the study to generation image in the process.
In traditional confrontation network model, arbiter only distinguishes the image data set of input and generation image,
There is no the features of study image, and also without the data distribution of acquisition image style, and generator is also in the training process
Data distribution toward data set is trained, and style is consistent.And the present invention is led to according to the operating process for changing image style
The style information that extraction image data is concentrated is crossed, the data distribution of style information is obtained, is input in generator, can reach and change
Become the effect of image style information intensity, there is good specific aim.
In practical applications, it should which, according to the complexity of data images feature, the number of convolution kernel is set.
Step S3, random noise is initialized, is inputted in generator.
Step S4, the data distribution for the style information that image data is concentrated is obtained using convolutional neural networks.
The specific method is as follows：
S41, the style information that image data is concentrated is extracted by convolutional neural networks；
S42, the data distribution for obtaining style information.
Step S5, style data distribution is added during generator each iteration, updates loss function, carries out follow-up
Training detailed process is as follows：
Here be added style information data distribution, be generator generate image data distribution on the basis of, in addition
The data distribution of style information, to achieve the purpose that change image style.
The effect of loss function is the ability weighed arbiter and judged generating image.The value of loss function is smaller, explanation
In current iteration, arbiter can have the generation image of preferable performance discrimination generator；Property that is on the contrary then illustrating arbiter
It can be poor.
The expression formula of loss function is：
Wherein, D (x) indicates that differentiation of the arbiter to image, pr indicate that the distribution of data images, pg indicate to generate image
Distribution, λ is hyper parameter,For gradient.
In conclusion present embodiment discloses a kind of method changing image style based on generation confrontation network model,
In traditional confrontation network model, arbiter only distinguishes that there is no learn to the image data set of input and generation image
The feature of image, also without the data distribution of acquisition image style, and generator is in the training process also only toward data set
Data distribution is trained, and style is consistent.And the present invention passes through extraction image according to the operating process for changing image style
Style information in data set obtains the data distribution of style information, is input in generator, can reach and change image style
The effect of information strength has good specific aim.
The above embodiment is a preferred embodiment of the present invention, but embodiments of the present invention are not by above-described embodiment
Limitation, it is other it is any without departing from the spirit and principles of the present invention made by changes, modifications, substitutions, combinations, simplifications,
Equivalent substitute mode is should be, is included within the scope of the present invention.
Claims (4)
1. a kind of based on the method for generating confrontation network model change image style, which is characterized in that under the method includes
Row step：
S1, construction are originally generated confrontation network model, generate image by generator and are input to arbiter progress network training；
S2, constructing neural network serve as the function of generator and arbiter；
S3, initialization random noise, input in generator；
S4, the data distribution that the style information that image data is concentrated is obtained using convolutional neural networks；
S5, the data distribution that style information is added during generator each iteration, update loss function, are subsequently instructed
Practice.
2. according to claim 1 a kind of based on the method for generating confrontation network model change image style, feature exists
In the step S4 processes are as follows：
S41, the style information that image data is concentrated is extracted by convolutional neural networks；
S42, the data distribution for obtaining style information.
3. according to claim 1 a kind of based on the method for generating confrontation network model change image style, feature exists
In the step S5 processes are as follows：
During generator each iteration, style information is added on the basis of generator generates the data distribution of image
Data distribution updates loss function, is subsequently trained.
4. according to claim 1 a kind of based on the method for generating confrontation network model change image style, feature exists
In the expression formula of the loss function is：
Wherein, D (x) indicates that differentiation of the arbiter to image, pr indicate that the distribution of data images, pg indicate to generate point of image
Cloth, λ are hyper parameter,For gradient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810101716.9A CN108460720A (en) | 2018-02-01 | 2018-02-01 | A method of changing image style based on confrontation network model is generated |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810101716.9A CN108460720A (en) | 2018-02-01 | 2018-02-01 | A method of changing image style based on confrontation network model is generated |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108460720A true CN108460720A (en) | 2018-08-28 |
Family
ID=63238456
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810101716.9A Pending CN108460720A (en) | 2018-02-01 | 2018-02-01 | A method of changing image style based on confrontation network model is generated |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108460720A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544442A (en) * | 2018-11-12 | 2019-03-29 | 南京邮电大学 | The image local Style Transfer method of production confrontation network based on dual confrontation |
| CN109639710A (en) * | 2018-12-29 | 2019-04-16 | 浙江工业大学 | A kind of network attack defence method based on dual training |
| CN110309889A (en) * | 2019-07-04 | 2019-10-08 | 西南大学 | A kind of Old-Yi character symbol restorative procedure of double arbiter GAN |
| CN110458216A (en) * | 2019-07-31 | 2019-11-15 | 中山大学 | The image Style Transfer method of confrontation network is generated based on condition |
| CN110490247A (en) * | 2019-03-08 | 2019-11-22 | 腾讯科技（深圳）有限公司 | Image processing model generation method, image processing method and device, electronic equipment |
| CN114549283A (en) * | 2022-01-14 | 2022-05-27 | 同济大学 | Training method of image generation model and image generation method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106886975A (en) * | 2016-11-29 | 2017-06-23 | 华南理工大学 | It is a kind of can real time execution image stylizing method |
| AU2017101166A4 (en) * | 2017-08-25 | 2017-11-02 | Lai, Haodong MR | A Method For Real-Time Image Style Transfer Based On Conditional Generative Adversarial Networks |
| CN107464210A (en) * | 2017-07-06 | 2017-12-12 | 浙江工业大学 | A kind of image Style Transfer method based on production confrontation network |
| CN107563510A (en) * | 2017-08-14 | 2018-01-09 | 华南理工大学 | A kind of WGAN model methods based on depth convolutional neural networks |
| CN107563995A (en) * | 2017-08-14 | 2018-01-09 | 华南理工大学 | A kind of confrontation network method of more arbiter error-duration models |
-
2018
- 2018-02-01 CN CN201810101716.9A patent/CN108460720A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106886975A (en) * | 2016-11-29 | 2017-06-23 | 华南理工大学 | It is a kind of can real time execution image stylizing method |
| CN107464210A (en) * | 2017-07-06 | 2017-12-12 | 浙江工业大学 | A kind of image Style Transfer method based on production confrontation network |
| CN107563510A (en) * | 2017-08-14 | 2018-01-09 | 华南理工大学 | A kind of WGAN model methods based on depth convolutional neural networks |
| CN107563995A (en) * | 2017-08-14 | 2018-01-09 | 华南理工大学 | A kind of confrontation network method of more arbiter error-duration models |
| AU2017101166A4 (en) * | 2017-08-25 | 2017-11-02 | Lai, Haodong MR | A Method For Real-Time Image Style Transfer Based On Conditional Generative Adversarial Networks |
Non-Patent Citations (2)
| Title |
|---|
| 徐一峰: "生成对抗网络理论模型和应用综述", 《金华职业技术学院学报》 * |
| 王坤峰 等: "平行图像: 图像生成的一个新型理论框架", 《模式识别与人工智能》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544442A (en) * | 2018-11-12 | 2019-03-29 | 南京邮电大学 | The image local Style Transfer method of production confrontation network based on dual confrontation |
| CN109544442B (en) * | 2018-11-12 | 2023-05-23 | 南京邮电大学 | Image local style migration method of double-countermeasure-based generation type countermeasure network |
| CN109639710A (en) * | 2018-12-29 | 2019-04-16 | 浙江工业大学 | A kind of network attack defence method based on dual training |
| CN109639710B (en) * | 2018-12-29 | 2021-02-26 | 浙江工业大学 | Network attack defense method based on countermeasure training |
| CN110490247A (en) * | 2019-03-08 | 2019-11-22 | 腾讯科技（深圳）有限公司 | Image processing model generation method, image processing method and device, electronic equipment |
| CN110309889A (en) * | 2019-07-04 | 2019-10-08 | 西南大学 | A kind of Old-Yi character symbol restorative procedure of double arbiter GAN |
| CN110458216A (en) * | 2019-07-31 | 2019-11-15 | 中山大学 | The image Style Transfer method of confrontation network is generated based on condition |
| CN110458216B (en) * | 2019-07-31 | 2022-04-12 | 中山大学 | Image style migration method for generating countermeasure network based on conditions |
| CN114549283A (en) * | 2022-01-14 | 2022-05-27 | 同济大学 | Training method of image generation model and image generation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108460720A (en) | A method of changing image style based on confrontation network model is generated | |
| Kobler et al. | Variational networks: connecting variational methods and deep learning | |
| CN107944358A (en) | A kind of human face generating method based on depth convolution confrontation network model | |
| CN107862377A (en) | A kind of packet convolution method that confrontation network model is generated based on text image | |
| CN108171320A (en) | A kind of image area switching network and conversion method based on production confrontation network | |
| CN109360159A (en) | A kind of image completion method based on generation confrontation network model | |
| CN108921123A (en) | A kind of face identification method based on double data enhancing | |
| CN107886169A (en) | A kind of multiple dimensioned convolution kernel method that confrontation network model is generated based on text image | |
| CN109978165A (en) | A kind of generation confrontation network method merged from attention mechanism | |
| CN108470196A (en) | A method of handwritten numeral is generated based on depth convolution confrontation network model | |
| CN109255364A (en) | A kind of scene recognition method generating confrontation network based on depth convolution | |
| JP2019510325A (en) | Method and system for generating multimodal digital images | |
| CN107437092A (en) | The sorting algorithm of retina OCT image based on Three dimensional convolution neutral net | |
| CN107871142A (en) | A kind of empty convolution method based on depth convolution confrontation network model | |
| CN105279519B (en) | Remote sensing image Clean water withdraw method and system based on coorinated training semi-supervised learning | |
| CN107590518A (en) | A kind of confrontation network training method of multiple features study | |
| CN109344879A (en) | A kind of decomposition convolution method fighting network model based on text-image | |
| CN109657612B (en) | Quality sorting system based on facial image features and application method thereof | |
| CN109859288A (en) | Based on the image painting methods and device for generating confrontation network | |
| CN107944546A (en) | It is a kind of based on be originally generated confrontation network model residual error network method | |
| CN108664922A (en) | A kind of infrared video Human bodys' response method based on personal safety | |
| CN107784630B (en) | Method, device and terminal for turning attributes of face image | |
| CN108985464A (en) | The continuous feature generation method of face for generating confrontation network is maximized based on information | |
| CN108537720A (en) | A kind of image processing method and device | |
| CN108021979A (en) | It is a kind of based on be originally generated confrontation network model feature recalibration convolution method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180828 |