CN109859096A - Image Style Transfer method, apparatus, electronic equipment and storage medium - Google Patents
Image Style Transfer method, apparatus, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN109859096A CN109859096A CN201811627179.8A CN201811627179A CN109859096A CN 109859096 A CN109859096 A CN 109859096A CN 201811627179 A CN201811627179 A CN 201811627179A CN 109859096 A CN109859096 A CN 109859096A
- Authority
- CN
- China
- Prior art keywords
- image
- style
- stylization
- stylized
- loss function
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
Abstract
Present application illustrates a kind of image Style Transfer method, apparatus, electronic equipment and storage mediums, and wherein image Style Transfer method includes: to obtain image to be processed;Using Style Transfer model, image to be processed is handled, obtains the first stylized image;Style Transfer model is to carry out model training previously according to content images, the second stylized image and third stylization image and obtain, second stylized image is that content images are carried out with stylization processing to obtain, and third stylization image is handled to obtain using neural network model to content images.The application carries out Style Transfer processing to image to be processed using the Style Transfer model that training obtains in advance, avoids carrying out the processes such as solving optimization to loss function during executing Style Transfer, to realize Fast Style Transfer;Also, Style Transfer model is to exercise supervision to train to neural network model using pairs of input and output sample to obtain, can be with the stylized effect of lift scheme.
Description
Technical field
This application involves field of computer technology more particularly to a kind of image Style Transfer method, apparatus, electronic equipment and
Storage medium.
Background technique
In image procossing and field of Computer Graphics, image Style Transfer (IST, Image Style Transfer) is
One extremely important and challenging project.According to the different problem scenes of actual treatment, IST is divided into artistic style again and moves
The sub- projects such as shifting, fidelity Style Transfer, semantic Style Transfer.IST has wide application scenarios, including picture editting, image
Synthesis, image special effect etc..
The method of research IST generally uses convolutional neural networks (CNN, Convolutional Neural at present
Networks it) extracts the feature of content (content) and style (style) figure and constructs Gram matrix, to define content damage
Style of becoming estranged loss, then Optimization Solution obtains stylized figure.This method although available good stylization as a result, but
Often solving optimization problem takes a long time, and cannot fast implement Style Transfer.
Summary of the invention
To overcome the problems in correlation technique, the application provides a kind of image Style Transfer method, apparatus, electronics is set
Standby and storage medium.
According to a first aspect of the present application, a kind of image Style Transfer method is provided, which comprises
Obtain image to be processed;
Using Style Transfer model, the image to be processed is handled, obtains the first stylized image;The style
Migration models are to carry out model training previously according to content images, the second stylized image and third stylization image and obtain
, wherein the described second stylized image is that the content images are carried out with stylization processing to obtain, the third stylization
Image is handled to obtain using neural network model to the content images.
In an optional implementation, Style Transfer model is used described, at the image to be processed
Before the step of managing, obtaining the first stylized image, further includes: obtain the Style Transfer model;
The step of acquisition Style Transfer model, comprising:
Obtain the content images and the second stylized image;
Using neural network model, the content images are handled, obtain third stylization image;
According to the linked character of the content images, the second stylized image and the third stylization image,
The neural network model is trained, the Style Transfer model is obtained;Wherein, the linked character includes content spy
At least one of sign, style and features, pixel point feature and semantic feature.
In an optional implementation, it is described according to the content images, the second stylized image and institute
The linked character for stating third stylization image is trained the neural network model, obtains the Style Transfer model
Step, comprising:
According to the third stylization image and the content characteristic difference of the content images, content loss letter is determined
Number;
According to the style and features difference of the third stylization image and the second stylized image, determine that style is damaged
Lose function;
According to the third stylization image and the pixel feature difference of the second stylized image, pixel is determined
Loss function;
According to the content loss function, the style loss function and the pixel loss function, the mind is determined
Parameter through network model obtains the Style Transfer model.
In an optional implementation, it is described according to the content images, the second stylized image and institute
The linked character for stating third stylization image is trained the neural network model, obtains the Style Transfer model
Step, comprising:
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image,
Obtain multiple semantic classes;
According to the content images, the second stylized image and the third stylization image, calculate each described
The corresponding loss function of semantic classes;
According to the corresponding loss function of each semantic classes, the parameter of the neural network model is determined, obtain described
Style Transfer model.
In an optional implementation, it is described according to the content images, the second stylized image and institute
The linked character for stating third stylization image is trained the neural network model, obtains the Style Transfer model
Step, comprising:
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image,
Obtain multiple semantic classes;
According to the corresponding content characteristic difference of semantic classes each in the third stylization image and the content images,
Determine content loss function;
It is special according to the corresponding style of semantic classes each in the third stylization image and the second stylized image
Difference is levied, determines style loss function;
According to the corresponding pixel of semantic classes each in the third stylization image and the second stylized image
Feature difference determines pixel loss function;
According to the content loss function, the style loss function and the pixel loss function, the mind is determined
Parameter through network model obtains the Style Transfer model.
According to a second aspect of the present application, a kind of image Style Transfer device is provided, described device includes:
Module is obtained, is configured as obtaining image to be processed;
Stylized module is configured as handling the image to be processed using Style Transfer model, being obtained first
Stylized image;The Style Transfer model is previously according to content images, the second stylized image and third stylization figure
It is obtained as carrying out model training, wherein the described second stylized image is to carry out stylization to the content images to handle
It arrives, the third stylization image is handled to obtain using neural network model to the content images.
In an optional implementation, described device further include: model training module is configured as obtaining the wind
Lattice migration models;
The model training module, comprising:
Acquiring unit is configured as obtaining the content images and the second stylized image;
Stylized unit is configured as handling the content images using neural network model, being obtained third wind
It formats image;
Training unit is configured as according to the content images, the second stylized image and the third style
The linked character for changing image, is trained the neural network model, obtains the Style Transfer model;Wherein, the pass
Joining feature includes at least one of content characteristic, style and features, pixel point feature and semantic feature.
In an optional implementation, the training unit is also configured to
According to the third stylization image and the content characteristic difference of the content images, content loss letter is determined
Number;
According to the style and features difference of the third stylization image and the second stylized image, determine that style is damaged
Lose function;
According to the third stylization image and the pixel feature difference of the second stylized image, pixel is determined
Loss function;
According to the content loss function, the style loss function and the pixel loss function, the mind is determined
Parameter through network model obtains the Style Transfer model.
In an optional implementation, the training unit is also configured to
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image,
Obtain multiple semantic classes;
According to the content images, the second stylized image and the third stylization image, calculate each described
The corresponding loss function of semantic classes;
According to the corresponding loss function of each semantic classes, the parameter of the neural network model is determined, obtain described
Style Transfer model.
In an optional implementation, the training unit is also configured to
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image,
Obtain multiple semantic classes;
According to the corresponding content characteristic difference of semantic classes each in the third stylization image and the content images,
Determine content loss function;
It is special according to the corresponding style of semantic classes each in the third stylization image and the second stylized image
Difference is levied, determines style loss function;
According to the corresponding pixel of semantic classes each in the third stylization image and the second stylized image
Feature difference determines pixel loss function;
According to the content loss function, the style loss function and the pixel loss function, the mind is determined
Parameter through network model obtains the Style Transfer model.
According to the third aspect of the application, a kind of electronic equipment is provided, the electronic equipment includes:
Processor;
For storing the memory of the processor-executable instruction;
Wherein, the processor is configured to executing image Style Transfer method as described in relation to the first aspect.
According to the fourth aspect of the application, a kind of non-transitorycomputer readable storage medium is provided, when the storage is situated between
When instruction in matter is executed by the processor of electronic equipment, so that the electronic equipment is able to carry out figure as described in relation to the first aspect
As Style Transfer method.
According to the 5th of the application the aspect, a kind of computer program product is provided, when in the computer program product
When instruction is executed by the processor of electronic equipment, so that the electronic equipment is able to carry out image style as described in relation to the first aspect
Moving method.
Technical solution provided by the present application can include the following benefits:
In this application, a kind of image Style Transfer method, apparatus, electronic equipment and storage medium are provided, wherein scheming
As Style Transfer method includes: to obtain image to be processed;Using Style Transfer model, image to be processed is handled, is obtained
First stylized image;Style Transfer model is previously according to content images, the second stylized image and third stylization figure
It is obtained as carrying out model training, wherein the second stylized image is that content images are carried out with stylization processing to obtain, third
Stylized image is handled to obtain using neural network model to content images.Technical scheme is using instruction in advance
The Style Transfer model got carries out Style Transfer processing to image to be processed, and it is right during executing Style Transfer to avoid
Loss function carries out the processes such as solving optimization, to realize Fast Style Transfer;Also, Style Transfer model is using in pairs
Input and output sample such as content images and the second stylized image, exercise supervision what training obtained to neural network model, pass through
There is the introducing of exemplar, it can be with the stylized effect of lift scheme.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not
The application can be limited.
Detailed description of the invention
The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the application
Example, and together with specification it is used to explain the principle of the application.
Fig. 1 is a kind of step flow chart of image Style Transfer method shown in the application.
Fig. 2 is a kind of step flow chart of acquisition Style Transfer model shown in the application.
Fig. 3 is a kind of functional block diagram of Style Transfer model shown in the application.
Fig. 4 is a kind of step flow chart of acquisition Style Transfer model realization mode shown in the application.
Fig. 5 is another step flow chart for obtaining Style Transfer model realization mode shown in the application.
Fig. 6 is a kind of block diagram of image Style Transfer device shown in the application.
Fig. 7 is the block diagram of a kind of electronic equipment shown in the application.
Fig. 8 is the block diagram of a kind of electronic equipment shown in the application.
Specific embodiment
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistent with the application.On the contrary, they be only with it is such as appended
The example of the consistent device and method of some aspects be described in detail in claims, the application.
What image Style Transfer (IST, Image Style Transfer) was studied is the style conversion between nature picture,
Its task is will to be got on reference to the Style Transfer of figure based on the content of original graph, obtained stylized figure.Fig. 1 is the application
A kind of flow chart of the image Style Transfer method shown, as shown in Figure 1, this method is applied in electronic equipment, this method packet
Include following steps.
In step s101, image to be processed is obtained.
Specifically, the image of pending stylization migration is obtained.
In step s 102, using Style Transfer model, image to be processed is handled, obtains the first stylization figure
Picture;Style Transfer model is to carry out model instruction previously according to content images, the second stylized image and third stylization image
It gets, wherein the second stylized image is that content images are carried out with stylization processing to obtain, and third stylization image is
Content images are handled using neural network model.
Specifically, using preparatory trained Style Transfer model, Style Transfer processing is carried out to image to be processed, is obtained
First stylized image.Since the present embodiment is using preparatory trained Style Transfer model, it is moved carrying out style
It is not necessary that image to be processed is directly input to Style Transfer model to loss function progress Gradient Iteration calculating during shifting,
I.e. exportable to obtain the first stylized image, thus the process existing " training " for generating stylized image becomes one
The process of " execution " preparatory trained Style Transfer model, therefore the speed for generating stylized image can be greatly speeded up.
Wherein, Style Transfer model can for example be obtained according to exemplar is trained in advance, and exemplar includes content graph
Picture and the second stylized image for carrying out stylized processing in advance according to content images, using neural network model to content graph
As being handled to obtain third stylization image, further according to content images, the second stylized image and third stylization image
Linked character neural network model is trained, obtain Style Transfer model.Wherein, content images are not carry out stylization
The original image samples of processing.
Image Style Transfer method provided in this embodiment, the Style Transfer model obtained using preparatory training is to be processed
Image carries out Style Transfer processing, without carrying out solving optimization to loss function during executing Style Transfer, thus real
Existing Fast Style Transfer;Also, Style Transfer model is using pairs of input and output sample such as content images and the second style
Change image, exercises supervision what training obtained to neural network model, it, can be with the wind of lift scheme by there is the introducing of exemplar
It formats effect.
It can also include: to obtain Style Transfer model before step S102 in a kind of implementation of the present embodiment
The step of, referring to Fig. 2, the step of obtaining Style Transfer model, be can specifically include:
In step s 201, content images and the second stylized image are obtained.
Specifically, the second stylized image (stylized figure) can carry out wind to content images (content figure) in advance
Processing of formatting obtains, such as the process of stylization processing can be completed by artist.
In practical applications, since we can only obtain a small amount of exemplar (pairs of content images and the second style
Change image), it is therefore possible to use single sample learning (One-Shot Learning) or few sample learning (Few-Shot
Learning method supervision machine study), few sample learning are to carry out model instruction using a pair of or several pairs of input and output samples
Practice, task is by completing model according to a small amount of exemplar to realize by a large amount of unlabeled exemplars or priori knowledge
Training.
According to the actually available quantity for having exemplar, such as largely helped without label substance image by no label data
It gives financial aid to students habit, i.e., carries out model training using the mode of semi-supervised learning algorithm.
In practical applications, such as model training can be carried out using 1000 content images Ic, wherein 1 Zhang Weiyou label
Content images Icos, it is corresponding with 1 second stylized image Isos, remaining 999 content images are the content images without label
Auxiliary carries out machine learning.It should be noted that there have the content images of label and corresponding second stylized image that can have to be more
It is right, in following model training process, it is only necessary to traverse each content images and each second stylized image.
In step S202, using neural network model, content images are handled, obtain third stylization image.
It specifically, (include the content images Ic of label by content images Icos) it is input to neural network model such as convolution
Neural network CNN, inference network as shown in Figure 3 (Inference Network) obtain third wind after carrying out stylized processing
Format image O.Third stylization image O is the corresponding inference network output figure of content images Ic, including IcosCorresponding inference net
Network output figure Oos。
In step S203, according to the linked character of content images, the second stylized image and third stylization image,
Neural network model is trained, Style Transfer model is obtained;Wherein, linked character include content characteristic, style and features, as
At least one of vegetarian refreshments feature and semantic feature.
It specifically, can be using Perceptual Loss network (such as VGG19) evaluation content image Ic shown in Fig. 3
It (include the content images Ic of labelos), the second stylized image IsosAnd third stylization image O (includes label
Content images IcosCorresponding inference network output figure Oos) between linked character difference, loss function is determined, by loss
Function optimizes, and determines the parameter in inference network, obtains Style Transfer model.
Wherein, linked character includes at least following at least one content characteristic, style and features, pixel point feature and semanteme
Feature etc..
In a kind of implementation, referring to Fig. 4, it may further include in step S203:
In step S401, according to third stylization image and the content characteristic difference of content images, determine that content is damaged
Lose function.
Specifically, content characteristic can be characterized using the characteristics of image of third stylization image O and content images Ic, because
This, content loss function can be calculated using following formula:
Wherein, L is the sum of CNN network convolutional layer, and i index indicates i-th layer of convolutional layer, shares N in each layeriGroup filter
Wave device, the corresponding characteristic pattern total pixel number of every group of filter are Di, Fi() corresponds to i-th layer of eigenmatrix of CNN, and (j, k) is used for rope
Draw, αiFor balancing the weight of each layer.
In step S402, according to the style and features difference of third stylization image and the second stylized image, determine
Style loss function.
It specifically, can be using the stylized image Is of third stylization image O and secondosGram matrix characterize style
Feature, therefore, style loss function can be calculated using following formula:
Wherein, L is the sum of CNN network convolutional layer, and i index indicates i-th layer of convolutional layer, shares N in each layeriGroup filter
Wave device, (j, k) are used to index, βiFor balancing the weight of each layer, GiWhat () corresponding i-th layer of feature calculation inner product of CNN obtained
Gram matrix, λsFor adjusting style loss ratio shared in total losses function.
In step S403, according to third stylization image and the pixel feature difference of the second stylized image, really
Determine pixel loss function.
It specifically, can be using the content images Ic for having label in third stylization image OosCorresponding inference network output
Scheme OosAnd the second stylized image IsosPixel number according to (such as gray scale) characterize pixel point feature, therefore, pixel loss
Function can be calculated using following formula:
Wherein, (j, k) is used to index, λeFor adjusting pixel loss ratio shared in total losses function, DosTo have
The content images Ic of labelosTotal pixel number.
In step s 404, according to content loss function, style loss function and pixel loss function, nerve net is determined
The parameter of network model obtains Style Transfer model.
Specifically, it can be lost according to the sum of content loss function, style loss function and pixel loss function
Function:
Then loss function is optimized, determines the parameter of neural network model, obtain Style Transfer model.
In a kind of implementation, referring to Fig. 5, it may further include in step S203:
In step S501, semantic segmentation is carried out to content images, the second stylized image and third stylization image,
Obtain multiple semantic classes.
Specifically, semantic segmentation, which refers to, is split image, and the classification (semantic category of object is determined according to segmentation result
?.In practical applications, it is real the modes such as can to divide automatically by manual or machine for semantic segmentation
It is existing.
In step S502, according to content images, the second stylized image and third stylization image, each semanteme is calculated
The corresponding loss function of classification.
For example, the method that can be provided using embodiment illustrated in fig. 4, calculates separately the corresponding content loss of each semantic classes
Function, style loss function and pixel loss function etc., can specifically include following steps:
According to the corresponding content characteristic difference of semantic classes each in third stylization image and content images, each language is determined
The corresponding content loss function of adopted classification;
According to the corresponding style and features difference of semantic classes each in third stylization image and the second stylized image, really
Determine the corresponding style loss function of each semantic classes;
According to the corresponding pixel feature difference of semantic classes each in third stylization image and the second stylized image,
Determine the corresponding pixel loss function of each semantic classes.
In step S503, according to the corresponding loss function of each semantic classes, determines the parameter of neural network model, obtain
Style Transfer model.
Specifically, the corresponding content loss function of each semantic classes can be summed, obtains content loss function;By each language
The corresponding style loss function summation of adopted classification, obtains style loss function;By the corresponding pixel loss function of each semantic classes
Summation, obtains pixel loss function;Then according to content loss function, style loss function and pixel loss function, such as
It sums to content loss function, style loss function and pixel loss function, obtains loss function, loss function is carried out excellent
Change and solve, can determine the parameter of neural network model, obtain Style Transfer model.
Image Style Transfer method provided in this embodiment can better discriminate between different languages by the way that semantic segmentation is added
Adopted classification avoids violating semantic information to carry out the calculating and optimization of loss function respectively for different semantic classes
Stylization further promotes whole style as a result, to which the Style Transfer model for obtaining training can be realized accurate migration
Effect.
Fig. 6 is a kind of block diagram of image Style Transfer device shown in the application.Reference Fig. 6, the apparatus may include:
Module 61 is obtained, is configured as obtaining image to be processed;
Stylized module 62 is configured as handling image to be processed using Style Transfer model, being obtained the first wind
It formats image;Style Transfer model is carried out previously according to content images, the second stylized image and third stylization image
What model training obtained, wherein the second stylized image is that content images are carried out with stylization processing to obtain, third stylization
Image is handled to obtain using neural network model to content images.
In an optional implementation, device can also include: model training module 63, be configured as obtaining style
Migration models.Model training module 63 can specifically include:
Acquiring unit is configured as obtaining content images and the second stylized image;
Stylized unit is configured as handling content images using neural network model, obtains third stylization
Image;
Training unit is configured as the association according to content images, the second stylized image and third stylization image
Feature is trained neural network model, obtains Style Transfer model;Wherein, linked character includes content characteristic, style spy
At least one of sign, pixel point feature and semantic feature.
In an optional implementation, training unit is also configured to
According to third stylization image and the content characteristic difference of content images, content loss function is determined;
According to the style and features difference of third stylization image and the second stylized image, style loss function is determined;
According to third stylization image and the pixel feature difference of the second stylized image, pixel loss letter is determined
Number;
According to content loss function, style loss function and pixel loss function, the parameter of neural network model is determined,
Obtain Style Transfer model.
Specifically, training unit can according to the sum of content loss function, style loss function and pixel loss function,
Obtain following loss function:
Wherein,
Then loss function is optimized, determines the parameter of neural network model, obtain Style Transfer model.
In an optional implementation, training unit is also configured to
Semantic segmentation is carried out to content images, the second stylized image and third stylization image, obtains multiple semantemes
Classification;
According to content images, the second stylized image and third stylization image, the corresponding damage of each semantic classes is calculated
Lose function;
According to the corresponding loss function of each semantic classes, the parameter of neural network model is determined, obtain Style Transfer model.
In an optional implementation, training unit is also configured to
Semantic segmentation is carried out to content images, the second stylized image and third stylization image, obtains multiple semantemes
Classification;
According to the corresponding content characteristic difference of semantic classes each in third stylization image and content images, content is determined
Loss function;
According to the corresponding style and features difference of semantic classes each in third stylization image and the second stylized image, really
Determine style loss function;
According to the corresponding pixel feature difference of semantic classes each in third stylization image and the second stylized image,
Determine pixel loss function;
According to content loss function, style loss function and pixel loss function, the parameter of neural network model is determined,
Obtain Style Transfer model.
About the device in above-described embodiment, the concrete mode and beneficial effect that wherein modules execute operation are
It is described in detail in the embodiment of the method, no detailed explanation will be given here.
Fig. 7 is the block diagram of a kind of electronic equipment 800 shown in the application.For example, electronic equipment 800 can be mobile electricity
Words, computer, digital broadcasting terminal, messaging device, game console, tablet device, Medical Devices, body-building equipment are a
Personal digital assistant etc..
Referring to Fig. 7, electronic equipment 800 may include following one or more components: processing component 802, memory 804,
Electric power assembly 806, multimedia component 808, audio component 810, the interface 812 of input/output (I/O), sensor module 814,
And communication component 816.
The integrated operation of the usual controlling electronic devices 800 of processing component 802, such as with display, call, data are logical
Letter, camera operation and record operate associated operation.Processing component 802 may include one or more processors 820 to hold
Row instruction, to complete all or part of the steps of the above method.In addition, processing component 802 may include one or more modules,
Convenient for the interaction between processing component 802 and other assemblies.For example, processing component 802 may include multi-media module, with convenient
Interaction between multimedia component 808 and processing component 802.
Memory 804 is configured as storing various types of data to support the operation in equipment 800.These data are shown
Example includes the instruction of any application or method for operating on electronic equipment 800, contact data, telephone directory number
According to, message, image, video etc..Memory 804 can by any kind of volatibility or non-volatile memory device or they
Combination realize, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM) is erasable
Programmable read only memory (EPROM), programmable read only memory (PROM), read-only memory (ROM), magnetic memory, quick flashing
Memory, disk or CD.
Power supply module 806 provides electric power for the various assemblies of electronic equipment 800.Power supply module 806 may include power supply pipe
Reason system, one or more power supplys and other with for electronic equipment 800 generate, manage, and distribute the associated component of electric power.
Multimedia component 808 includes the screen of one output interface of offer between the electronic equipment 800 and user.
In some embodiments, screen may include liquid crystal display (LCD) and touch panel (TP).If screen includes touch surface
Plate, screen may be implemented as touch screen, to receive input signal from the user.Touch panel includes one or more touches
Sensor is to sense the gesture on touch, slide, and touch panel.The touch sensor can not only sense touch or sliding
The boundary of movement, but also detect duration and pressure associated with the touch or slide operation.In some embodiments,
Multimedia component 808 includes a front camera and/or rear camera.When equipment 800 is in operation mode, as shot mould
When formula or video mode, front camera and/or rear camera can receive external multi-medium data.Each preposition camera shooting
Head and rear camera can be a fixed optical lens system or have focusing and optical zoom capabilities.
Audio component 810 is configured as output and/or input audio signal.For example, audio component 810 includes a Mike
Wind (MIC), when electronic equipment 800 is in operation mode, when such as call mode, recording mode, and voice recognition mode, microphone
It is configured as receiving external audio signal.The received audio signal can be further stored in memory 804 or via logical
Believe that component 816 is sent.In some embodiments, audio component 810 further includes a loudspeaker, is used for output audio signal.
I/O interface 812 provides interface between processing component 802 and peripheral interface module, and above-mentioned peripheral interface module can
To be keyboard, click wheel, button etc..These buttons may include, but are not limited to: home button, volume button, start button and lock
Determine button.
Sensor module 814 includes one or more sensors, for providing the state of various aspects for electronic equipment 800
Assessment.For example, sensor module 814 can detecte the state that opens/closes of equipment 800, the relative positioning of component, such as institute
The display and keypad that component is electronic equipment 800 are stated, sensor module 814 can also detect electronic equipment 800 or electronics
The position change of 800 1 components of equipment, the existence or non-existence that user contacts with electronic equipment 800,800 orientation of electronic equipment
Or the temperature change of acceleration/deceleration and electronic equipment 800.Sensor module 814 may include proximity sensor, be configured to
It detects the presence of nearby objects without any physical contact.Sensor module 814 can also include optical sensor, such as
CMOS or ccd image sensor, for being used in imaging applications.In some embodiments, which can be with
Including acceleration transducer, gyro sensor, Magnetic Sensor, pressure sensor or temperature sensor.
Communication component 816 is configured to facilitate the communication of wired or wireless way between electronic equipment 800 and other equipment.
Electronic equipment 800 can access the wireless network based on communication standard, such as WiFi, carrier network (such as 2G, 3G, 4G or 5G),
Or their combination.In one exemplary embodiment, communication component 816 receives via broadcast channel and comes from external broadcasting management
The broadcast singal or broadcast related information of system.In one exemplary embodiment, the communication component 816 further includes that near field is logical
(NFC) module is believed, to promote short range communication.For example, radio frequency identification (RFID) technology, infrared data association can be based in NFC module
Meeting (IrDA) technology, ultra wide band (UWB) technology, bluetooth (BT) technology and other technologies are realized.
In the exemplary embodiment, electronic equipment 800 can be by one or more application specific integrated circuit (ASIC), number
Word signal processor (DSP), programmable logic device (PLD), field programmable gate array (FPGA), controller, microcontroller,
Microprocessor or other electronic components are realized, for executing the above method.
In the exemplary embodiment, a kind of non-transitorycomputer readable storage medium including instruction, example are additionally provided
It such as include the memory 804 of instruction, above-metioned instruction can be executed by the processor 820 of electronic equipment 800 to complete the above method.Example
Such as, the non-transitorycomputer readable storage medium can be ROM, random access memory (RAM), CD-ROM, tape, soft
Disk and optical data storage devices etc..
Fig. 8 is the block diagram of a kind of electronic equipment 1900 shown in the application.For example, electronic equipment 1900 may be provided as
One server.
Referring to Fig. 8, it further comprises one or more processors that electronic equipment 1900, which includes processing component 1922, with
And memory resource represented by a memory 1932, it can be by the instruction of the execution of processing component 1922 for storing, such as answer
Use program.The application program stored in memory 1932 may include it is one or more each correspond to one group of instruction
Module.In addition, processing component 1922 is configured as executing instruction, to execute the above method.
Electronic equipment 1900 can also include that a power supply module 1926 is configured as executing the power supply of electronic equipment 1900
Management, a wired or wireless network interface 1950 is configured as electronic equipment 1900 being connected to network and an input is defeated
(I/O) interface 1958 out.Electronic equipment 1900 can be operated based on the operating system for being stored in memory 1932, such as
Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM or similar.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to its of the application
Its embodiment.This application is intended to cover any variations, uses, or adaptations of the application, these modifications, purposes or
Person's adaptive change follows the general principle of the application and including the undocumented common knowledge in the art of the application
Or conventional techniques.The description and examples are only to be considered as illustrative, and the true scope and spirit of the application are by following
Claim is pointed out.
It should be understood that the application is not limited to the precise structure that has been described above and shown in the drawings, and
And various modifications and changes may be made without departing from the scope thereof.Scope of the present application is only limited by the accompanying claims.
Claims (10)
1. a kind of image Style Transfer method, which is characterized in that the described method includes:
Obtain image to be processed;
Using Style Transfer model, the image to be processed is handled, obtains the first stylized image;The Style Transfer
Model is to carry out model training previously according to content images, the second stylized image and third stylization image and obtain,
In, the described second stylized image is that the content images are carried out with stylization processing to obtain, the third stylization image
The content images are handled to obtain using neural network model.
2. image Style Transfer method according to claim 1, which is characterized in that Style Transfer model is used described,
Before the step of handling the image to be processed, obtaining the first stylized image, further includes: obtain the Style Transfer
Model;
The step of acquisition Style Transfer model, comprising:
Obtain the content images and the second stylized image;
Using neural network model, the content images are handled, obtain third stylization image;
According to the linked character of the content images, the second stylized image and the third stylization image, to institute
It states neural network model to be trained, obtains the Style Transfer model;Wherein, the linked character includes content characteristic, wind
At least one of lattice feature, pixel point feature and semantic feature.
3. image Style Transfer method according to claim 2, which is characterized in that described according to the content images, institute
The linked character for stating the second stylized image and the third stylization image, is trained the neural network model,
The step of obtaining the Style Transfer model, comprising:
According to the third stylization image and the content characteristic difference of the content images, content loss function is determined;
According to the style and features difference of the third stylization image and the second stylized image, determine that style loses letter
Number;
According to the third stylization image and the pixel feature difference of the second stylized image, pixel loss is determined
Function;
According to the content loss function, the style loss function and the pixel loss function, the nerve net is determined
The parameter of network model obtains the Style Transfer model;
Or it is described according to the content images, the association of the second stylized image and the third stylization image it is special
Sign, the step of being trained to the neural network model, obtain the Style Transfer model, comprising:
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image, is obtained
Multiple semantic classes;
According to the content images, the second stylized image and the third stylization image, each semanteme is calculated
The corresponding loss function of classification;
According to the corresponding loss function of each semantic classes, determines the parameter of the neural network model, obtain the style
Migration models.
4. image Style Transfer method according to claim 2, which is characterized in that described according to the content images, institute
The linked character for stating the second stylized image and the third stylization image, is trained the neural network model,
The step of obtaining the Style Transfer model, comprising:
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image, is obtained
Multiple semantic classes;
According to the corresponding content characteristic difference of semantic classes each in the third stylization image and the content images, determine
Content loss function;
It is poor according to the corresponding style and features of semantic classes each in the third stylization image and the second stylized image
It is different, determine style loss function;
According to the corresponding pixel point feature of semantic classes each in the third stylization image and the second stylized image
Difference determines pixel loss function;
According to the content loss function, the style loss function and the pixel loss function, the nerve net is determined
The parameter of network model obtains the Style Transfer model.
5. a kind of image Style Transfer device, which is characterized in that described device includes:
Module is obtained, is configured as obtaining image to be processed;
Stylized module is configured as handling the image to be processed using Style Transfer model, being obtained the first style
Change image;The Style Transfer model be previously according to content images, the second stylized image and third stylization image into
Row model training obtains, wherein and the described second stylized image is that the content images are carried out with stylization processing to obtain,
The third stylization image is handled to obtain using neural network model to the content images.
6. image Style Transfer device according to claim 5, which is characterized in that described device further include: model training
Module is configured as obtaining the Style Transfer model;
The model training module, comprising:
Acquiring unit is configured as obtaining the content images and the second stylized image;
Stylized unit is configured as handling the content images using neural network model, obtains third stylization
Image;
Training unit is configured as according to the content images, the second stylized image and the third stylization figure
The linked character of picture is trained the neural network model, obtains the Style Transfer model;Wherein, the association is special
Sign includes at least one of content characteristic, style and features, pixel point feature and semantic feature.
7. image Style Transfer device according to claim 6, which is characterized in that the training unit is also configured to
According to the third stylization image and the content characteristic difference of the content images, content loss function is determined;
According to the style and features difference of the third stylization image and the second stylized image, determine that style loses letter
Number;
According to the third stylization image and the pixel feature difference of the second stylized image, pixel loss is determined
Function;
According to the content loss function, the style loss function and the pixel loss function, the nerve net is determined
The parameter of network model obtains the Style Transfer model;
Or the training unit is also configured to
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image, is obtained
Multiple semantic classes;
According to the content images, the second stylized image and the third stylization image, each semanteme is calculated
The corresponding loss function of classification;
According to the corresponding loss function of each semantic classes, determines the parameter of the neural network model, obtain the style
Migration models.
8. image Style Transfer device according to claim 6, which is characterized in that the training unit is also configured to
Semantic segmentation is carried out to the content images, the second stylized image and the third stylization image, is obtained
Multiple semantic classes;
According to the corresponding content characteristic difference of semantic classes each in the third stylization image and the content images, determine
Content loss function;
It is poor according to the corresponding style and features of semantic classes each in the third stylization image and the second stylized image
It is different, determine style loss function;
According to the corresponding pixel point feature of semantic classes each in the third stylization image and the second stylized image
Difference determines pixel loss function;
According to the content loss function, the style loss function and the pixel loss function, the nerve net is determined
The parameter of network model obtains the Style Transfer model.
9. a kind of electronic equipment, which is characterized in that the electronic equipment includes:
Processor;
For storing the memory of the processor-executable instruction;
Wherein, the processor is configured to executing image Style Transfer method according to any one of claims 1-4.
10. a kind of non-transitorycomputer readable storage medium, when the instruction in the storage medium is by the processing of electronic equipment
When device executes, so that the electronic equipment is able to carry out image Style Transfer method according to any one of claims 1-4.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811627179.8A CN109859096A (en) | 2018-12-28 | 2018-12-28 | Image Style Transfer method, apparatus, electronic equipment and storage medium |
PCT/CN2019/114793 WO2020134556A1 (en) | 2018-12-28 | 2019-10-31 | Image style transfer method, device, electronic apparatus, and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811627179.8A CN109859096A (en) | 2018-12-28 | 2018-12-28 | Image Style Transfer method, apparatus, electronic equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109859096A true CN109859096A (en) | 2019-06-07 |
Family
ID=66893099
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811627179.8A Pending CN109859096A (en) | 2018-12-28 | 2018-12-28 | Image Style Transfer method, apparatus, electronic equipment and storage medium |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN109859096A (en) |
WO (1) | WO2020134556A1 (en) |
Cited By (27)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110310222A (en) * | 2019-06-20 | 2019-10-08 | 北京奇艺世纪科技有限公司 | A kind of image Style Transfer method, apparatus, electronic equipment and storage medium |
CN110598781A (en) * | 2019-09-05 | 2019-12-20 | Oppo广东移动通信有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
CN110930295A (en) * | 2019-10-25 | 2020-03-27 | 广东开放大学(广东理工职业学院) | Image style migration method, system, device and storage medium |
CN111127309A (en) * | 2019-12-12 | 2020-05-08 | 杭州格像科技有限公司 | Portrait style transfer model training method, portrait style transfer method and device |
CN111325664A (en) * | 2020-02-27 | 2020-06-23 | Oppo广东移动通信有限公司 | Style migration method and device, storage medium and electronic equipment |
WO2020134556A1 (en) * | 2018-12-28 | 2020-07-02 | 北京达佳互联信息技术有限公司 | Image style transfer method, device, electronic apparatus, and storage medium |
CN111667399A (en) * | 2020-05-14 | 2020-09-15 | 华为技术有限公司 | Method for training style migration model, method and device for video style migration |
CN111710011A (en) * | 2020-06-10 | 2020-09-25 | 广州梦映动漫网络科技有限公司 | Cartoon generation method and system, electronic device and medium |
WO2020248767A1 (en) * | 2019-06-11 | 2020-12-17 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method, system, and computer-readable medium for stylizing video frames |
CN112258381A (en) * | 2020-09-29 | 2021-01-22 | 北京达佳互联信息技术有限公司 | Model training method, image processing method, device, equipment and storage medium |
CN112383765A (en) * | 2020-11-10 | 2021-02-19 | 中移雄安信息通信科技有限公司 | VR image transmission method and device |
CN112581358A (en) * | 2020-12-17 | 2021-03-30 | 北京达佳互联信息技术有限公司 | Training method of image processing model, image processing method and device |
CN112581361A (en) * | 2020-12-30 | 2021-03-30 | 北京达佳互联信息技术有限公司 | Training method of image style migration model, and image style migration method and device |
CN112734627A (en) * | 2020-12-24 | 2021-04-30 | 北京达佳互联信息技术有限公司 | Training method of image style migration model, and image style migration method and device |
CN112785493A (en) * | 2021-01-22 | 2021-05-11 | 北京百度网讯科技有限公司 | Model training method, style migration method, device, equipment and storage medium |
CN112862669A (en) * | 2021-02-02 | 2021-05-28 | 百果园技术(新加坡)有限公司 | Training method, generating method, device and equipment of image generation model |
CN112883806A (en) * | 2021-01-21 | 2021-06-01 | 杭州广电云网络科技有限公司 | Video style migration method and device based on neural network, computer equipment and storage medium |
CN113052757A (en) * | 2021-03-08 | 2021-06-29 | Oppo广东移动通信有限公司 | Image processing method, device, terminal and storage medium |
CN113139893A (en) * | 2020-01-20 | 2021-07-20 | 北京达佳互联信息技术有限公司 | Image translation model construction method and device and image translation method and device |
CN113240576A (en) * | 2021-05-12 | 2021-08-10 | 北京达佳互联信息技术有限公司 | Method and device for training style migration model, electronic equipment and storage medium |
CN113259583A (en) * | 2020-02-13 | 2021-08-13 | 北京小米移动软件有限公司 | Image processing method, device, terminal and storage medium |
CN113269213A (en) * | 2020-02-17 | 2021-08-17 | 百度在线网络技术(北京)有限公司 | Training set acquisition method and device and electronic equipment |
WO2021159781A1 (en) * | 2020-02-13 | 2021-08-19 | 腾讯科技(深圳)有限公司 | Image processing method, apparatus and device, and storage medium |
CN113656126A (en) * | 2021-07-30 | 2021-11-16 | 北京数码大方科技股份有限公司 | Menu display method and device of information system and electronic equipment |
CN113763230A (en) * | 2020-06-04 | 2021-12-07 | 北京达佳互联信息技术有限公司 | Image style migration model training method, style migration method and device |
CN113793253A (en) * | 2021-08-30 | 2021-12-14 | 易佳慧 | Style migration system for automatically generating stylized video |
CN114298895A (en) * | 2021-12-24 | 2022-04-08 | 成都索贝数码科技股份有限公司 | Image realistic style migration method, device, equipment and storage medium |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20230177662A1 (en) * | 2021-12-02 | 2023-06-08 | Robert Bosch Gmbh | System and Method for Augmenting Vision Transformers |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107392842A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | Image stylization processing method, device, computing device and computer-readable storage medium |
CN107977414A (en) * | 2017-11-22 | 2018-05-01 | 西安财经学院 | Image Style Transfer method and its system based on deep learning |
US20180144509A1 (en) * | 2016-09-02 | 2018-05-24 | Artomatix Ltd. | Systems and Methods for Providing Convolutional Neural Network Based Image Synthesis Using Stable and Controllable Parametric Models, a Multiscale Synthesis Framework and Novel Network Architectures |
CN108122264A (en) * | 2016-11-28 | 2018-06-05 | 奥多比公司 | Sketch is promoted to be converted to drawing |
CN108537776A (en) * | 2018-03-12 | 2018-09-14 | 维沃移动通信有限公司 | A kind of image Style Transfer model generating method and mobile terminal |
CN108596830A (en) * | 2018-04-28 | 2018-09-28 | 国信优易数据有限公司 | A kind of image Style Transfer model training method and image Style Transfer method |
CN108629747A (en) * | 2018-04-25 | 2018-10-09 | 腾讯科技(深圳)有限公司 | Image enchancing method, device, electronic equipment and storage medium |
CN108765278A (en) * | 2018-06-05 | 2018-11-06 | Oppo广东移动通信有限公司 | A kind of image processing method, mobile terminal and computer readable storage medium |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106651766A (en) * | 2016-12-30 | 2017-05-10 | 深圳市唯特视科技有限公司 | Image style migration method based on deep convolutional neural network |
US10152768B2 (en) * | 2017-04-14 | 2018-12-11 | Facebook, Inc. | Artifact reduction for image style transfer |
CN109859096A (en) * | 2018-12-28 | 2019-06-07 | 北京达佳互联信息技术有限公司 | Image Style Transfer method, apparatus, electronic equipment and storage medium |
-
2018
- 2018-12-28 CN CN201811627179.8A patent/CN109859096A/en active Pending
-
2019
- 2019-10-31 WO PCT/CN2019/114793 patent/WO2020134556A1/en active Application Filing
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180144509A1 (en) * | 2016-09-02 | 2018-05-24 | Artomatix Ltd. | Systems and Methods for Providing Convolutional Neural Network Based Image Synthesis Using Stable and Controllable Parametric Models, a Multiscale Synthesis Framework and Novel Network Architectures |
CN108122264A (en) * | 2016-11-28 | 2018-06-05 | 奥多比公司 | Sketch is promoted to be converted to drawing |
CN107392842A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | Image stylization processing method, device, computing device and computer-readable storage medium |
CN107977414A (en) * | 2017-11-22 | 2018-05-01 | 西安财经学院 | Image Style Transfer method and its system based on deep learning |
CN108537776A (en) * | 2018-03-12 | 2018-09-14 | 维沃移动通信有限公司 | A kind of image Style Transfer model generating method and mobile terminal |
CN108629747A (en) * | 2018-04-25 | 2018-10-09 | 腾讯科技(深圳)有限公司 | Image enchancing method, device, electronic equipment and storage medium |
CN108596830A (en) * | 2018-04-28 | 2018-09-28 | 国信优易数据有限公司 | A kind of image Style Transfer model training method and image Style Transfer method |
CN108765278A (en) * | 2018-06-05 | 2018-11-06 | Oppo广东移动通信有限公司 | A kind of image processing method, mobile terminal and computer readable storage medium |
Cited By (41)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020134556A1 (en) * | 2018-12-28 | 2020-07-02 | 北京达佳互联信息技术有限公司 | Image style transfer method, device, electronic apparatus, and storage medium |
WO2020248767A1 (en) * | 2019-06-11 | 2020-12-17 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method, system, and computer-readable medium for stylizing video frames |
CN110310222A (en) * | 2019-06-20 | 2019-10-08 | 北京奇艺世纪科技有限公司 | A kind of image Style Transfer method, apparatus, electronic equipment and storage medium |
CN110598781A (en) * | 2019-09-05 | 2019-12-20 | Oppo广东移动通信有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
CN110930295A (en) * | 2019-10-25 | 2020-03-27 | 广东开放大学(广东理工职业学院) | Image style migration method, system, device and storage medium |
CN111127309A (en) * | 2019-12-12 | 2020-05-08 | 杭州格像科技有限公司 | Portrait style transfer model training method, portrait style transfer method and device |
CN111127309B (en) * | 2019-12-12 | 2023-08-11 | 杭州格像科技有限公司 | Portrait style migration model training method, portrait style migration method and device |
CN113139893A (en) * | 2020-01-20 | 2021-07-20 | 北京达佳互联信息技术有限公司 | Image translation model construction method and device and image translation method and device |
CN113139893B (en) * | 2020-01-20 | 2023-10-03 | 北京达佳互联信息技术有限公司 | Image translation model construction method and device and image translation method and device |
CN113259583A (en) * | 2020-02-13 | 2021-08-13 | 北京小米移动软件有限公司 | Image processing method, device, terminal and storage medium |
CN113259583B (en) * | 2020-02-13 | 2023-05-12 | 北京小米移动软件有限公司 | Image processing method, device, terminal and storage medium |
WO2021159781A1 (en) * | 2020-02-13 | 2021-08-19 | 腾讯科技(深圳)有限公司 | Image processing method, apparatus and device, and storage medium |
CN113269213A (en) * | 2020-02-17 | 2021-08-17 | 百度在线网络技术(北京)有限公司 | Training set acquisition method and device and electronic equipment |
CN111325664A (en) * | 2020-02-27 | 2020-06-23 | Oppo广东移动通信有限公司 | Style migration method and device, storage medium and electronic equipment |
CN111325664B (en) * | 2020-02-27 | 2023-08-29 | Oppo广东移动通信有限公司 | Style migration method and device, storage medium and electronic equipment |
CN111667399B (en) * | 2020-05-14 | 2023-08-25 | 华为技术有限公司 | Training method of style migration model, video style migration method and device |
CN111667399A (en) * | 2020-05-14 | 2020-09-15 | 华为技术有限公司 | Method for training style migration model, method and device for video style migration |
CN113763230A (en) * | 2020-06-04 | 2021-12-07 | 北京达佳互联信息技术有限公司 | Image style migration model training method, style migration method and device |
CN111710011A (en) * | 2020-06-10 | 2020-09-25 | 广州梦映动漫网络科技有限公司 | Cartoon generation method and system, electronic device and medium |
CN112258381B (en) * | 2020-09-29 | 2024-02-09 | 北京达佳互联信息技术有限公司 | Model training method, image processing method, device, equipment and storage medium |
CN112258381A (en) * | 2020-09-29 | 2021-01-22 | 北京达佳互联信息技术有限公司 | Model training method, image processing method, device, equipment and storage medium |
CN112383765A (en) * | 2020-11-10 | 2021-02-19 | 中移雄安信息通信科技有限公司 | VR image transmission method and device |
CN112581358A (en) * | 2020-12-17 | 2021-03-30 | 北京达佳互联信息技术有限公司 | Training method of image processing model, image processing method and device |
CN112581358B (en) * | 2020-12-17 | 2023-09-26 | 北京达佳互联信息技术有限公司 | Training method of image processing model, image processing method and device |
CN112734627A (en) * | 2020-12-24 | 2021-04-30 | 北京达佳互联信息技术有限公司 | Training method of image style migration model, and image style migration method and device |
CN112734627B (en) * | 2020-12-24 | 2023-07-11 | 北京达佳互联信息技术有限公司 | Training method of image style migration model, image style migration method and device |
CN112581361A (en) * | 2020-12-30 | 2021-03-30 | 北京达佳互联信息技术有限公司 | Training method of image style migration model, and image style migration method and device |
CN112581361B (en) * | 2020-12-30 | 2023-07-11 | 北京达佳互联信息技术有限公司 | Training method of image style migration model, image style migration method and device |
CN112883806A (en) * | 2021-01-21 | 2021-06-01 | 杭州广电云网络科技有限公司 | Video style migration method and device based on neural network, computer equipment and storage medium |
CN112883806B (en) * | 2021-01-21 | 2024-03-22 | 杭州广电云网络科技有限公司 | Video style migration method and device based on neural network, computer equipment and storage medium |
CN112785493A (en) * | 2021-01-22 | 2021-05-11 | 北京百度网讯科技有限公司 | Model training method, style migration method, device, equipment and storage medium |
CN112785493B (en) * | 2021-01-22 | 2024-02-09 | 北京百度网讯科技有限公司 | Model training method, style migration method, device, equipment and storage medium |
CN112862669A (en) * | 2021-02-02 | 2021-05-28 | 百果园技术(新加坡)有限公司 | Training method, generating method, device and equipment of image generation model |
CN112862669B (en) * | 2021-02-02 | 2024-02-09 | 百果园技术(新加坡)有限公司 | Training method, generating method, device and equipment for image generating model |
CN113052757A (en) * | 2021-03-08 | 2021-06-29 | Oppo广东移动通信有限公司 | Image processing method, device, terminal and storage medium |
CN113240576A (en) * | 2021-05-12 | 2021-08-10 | 北京达佳互联信息技术有限公司 | Method and device for training style migration model, electronic equipment and storage medium |
CN113656126A (en) * | 2021-07-30 | 2021-11-16 | 北京数码大方科技股份有限公司 | Menu display method and device of information system and electronic equipment |
CN113656126B (en) * | 2021-07-30 | 2023-10-27 | 北京数码大方科技股份有限公司 | Menu display method and menu display device of information system and electronic equipment |
CN113793253A (en) * | 2021-08-30 | 2021-12-14 | 易佳慧 | Style migration system for automatically generating stylized video |
CN114298895B (en) * | 2021-12-24 | 2023-09-26 | 成都索贝数码科技股份有限公司 | Image realism style migration method, device, equipment and storage medium |
CN114298895A (en) * | 2021-12-24 | 2022-04-08 | 成都索贝数码科技股份有限公司 | Image realistic style migration method, device, equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
WO2020134556A1 (en) | 2020-07-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109859096A (en) | Image Style Transfer method, apparatus, electronic equipment and storage medium | |
CN108121952B (en) | Face key point positioning method, device, equipment and storage medium | |
CN111310616B (en) | Image processing method and device, electronic equipment and storage medium | |
CN109117862B (en) | Image tag recognition methods, device and server | |
WO2020224479A1 (en) | Method and apparatus for acquiring positions of target, and computer device and storage medium | |
CN104408402B (en) | Face identification method and device | |
CN109800325A (en) | Video recommendation method, device and computer readable storage medium | |
CN110909815B (en) | Neural network training method, neural network training device, neural network processing device, neural network training device, image processing device and electronic equipment | |
CN107133354B (en) | Method and device for acquiring image description information | |
CN109543066A (en) | Video recommendation method, device and computer readable storage medium | |
CN111552888A (en) | Content recommendation method, device, equipment and storage medium | |
US11556761B2 (en) | Method and device for compressing a neural network model for machine translation and storage medium | |
CN109670077A (en) | Video recommendation method, device and computer readable storage medium | |
CN109615006A (en) | Character recognition method and device, electronic equipment and storage medium | |
CN109360197A (en) | Processing method, device, electronic equipment and the storage medium of image | |
CN110188865B (en) | Information processing method and device, electronic equipment and storage medium | |
CN111368525A (en) | Information searching method, device, equipment and storage medium | |
JP2022522551A (en) | Image processing methods and devices, electronic devices and storage media | |
CN110781323A (en) | Method and device for determining label of multimedia resource, electronic equipment and storage medium | |
CN112148980B (en) | Article recommending method, device, equipment and storage medium based on user click | |
CN110399934A (en) | A kind of video classification methods, device and electronic equipment | |
CN109977860A (en) | Image processing method and device, electronic equipment and storage medium | |
CN109145876A (en) | Image classification method, device, electronic equipment and storage medium | |
CN107239758A (en) | The method and device of face key point location | |
CN115035596B (en) | Behavior detection method and device, electronic equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |