CN109849576A - A kind of method of reference gray level figure auxiliary drawing - Google Patents
A kind of method of reference gray level figure auxiliary drawing Download PDFInfo
- Publication number
- CN109849576A CN109849576A CN201910149284.3A CN201910149284A CN109849576A CN 109849576 A CN109849576 A CN 109849576A CN 201910149284 A CN201910149284 A CN 201910149284A CN 109849576 A CN109849576 A CN 109849576A
- Authority
- CN
- China
- Prior art keywords
- image
- semantic segmentation
- grayscale image
- grayscale
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of methods of reference gray level figure auxiliary drawing, including the training stage: obtaining training image, and grayscale image is converted by training image, training image is subjected to semantic segmentation mark simultaneously and obtains semantic segmentation image, using semantic segmentation image as input, using corresponding grayscale image as exporting, the deep learning model that can be realized image and be converted into image is trained, it determines model parameter, obtains grayscale image and generate model;Application stage: semantic segmentation figure is drawn, and the semantic segmentation figure of drafting is input to grayscale image and is generated in model, it is computed and generates multiple grayscale images, according to the grayscale image that selection can refer to from grayscale image, texture rendering is carried out to semantic segmentation figure, improves paint.The drawing householder method will be converted into generating grayscale image by semantic segmentation figure, provide light and shade in the content detail on drawing, texture reference, support author to find inspiration from grayscale image, reduce time of the person of drawing a picture in design works details.
Description
Technical field
The invention belongs to field of auxiliary of painting, and in particular to a kind of method of reference gray level figure auxiliary drawing.
Background technique
With the development of computer industry, computer experienced during creative design execution, imitation, assistant turn
Become.In painting creation field, the tool that computer provides multiple functions from one is gradually transformed into the role of drawing auxiliary.Technology
Development allowed computer by learning the works of certain artist, capture its style, imitated and carry out drawing wound
Make.User and collaborative computer creation also may be implemented in the assistant role that computer is played the part of, and mentions during creation for user
For miscellaneous support, allow layman that can also carry out high-caliber painting creation.In the field of auxiliary drawing, calculate
Machine auxiliary still has great expansible space.
Application of the existing artificial intelligence technology generation technique in drawing can simply be divided into works and generate and assist letter
Breath generates two aspects.Works generation directly generates final as a result, being as a result usually fixed constant.Auxiliary information generates
Auxiliary information is provided during drawing a picture, helps the quick result obtained to the end of the person of drawing a picture.
The first draft that the application that works generate is drawn based on author, author first design simple works first draft, such as rough draft, line
Original text, layout etc., artificial intelligence technology are done directly next generation work, directly generate simple works first draft final
Works.Such method is although very convenient, but gives model sizable free space, and generation may be selected in model
As a result type is excessive, causes the image effect generated excessively poor, while depending on the content of first draft unduly and draw a picture level, only pole
The least a portion of first draft content for meeting model generation can obtain good result.Although current deep learning model has been able to
Learn the style of paintings out, and final works can be obtained by way of Style Transfer, effect is also preferable, but such
Works be all often it is abstract, fine distinction will not influence the experience of whole picture paintings.And on non-abstract visual pattern,
People is very sensitive to tiny distortion and entanglement, and the works that artificial intelligence technology generates often have the mistake for violating common sense
Difference.It is limited to current generation technique, in the result that non-abstract image domains directly generate artificial intelligence technology as works
Still there are many difficulties to need to solve.
The application that auxiliary information generates will generate result and be supplied to author during drawing a picture as reference, such reference
Mode type is more, such as provides the prediction lines of next record when setting-out original text, provides when drawing layout and generates effect by reference.
The reference single effect that these modes provide is typically based on lines and whole effect, and major part absolutely does not give particulars in content
Guidance.And the person of drawing a picture will enrich the details of paintings during common draw a picture, and need to obtain from daily life and memory
The inspiration of details, it is also very desirable to the guidance and reference of this respect.The related application also lacked in artificial intelligence technology at present is supported
Reference in content detail is provided.
Summary of the invention
The object of the present invention is to provide a kind of method of reference gray level figure auxiliary drawing, which will pass through language
Adopted segmentation figure is converted into generating grayscale image, provides light and shade in the content detail on drawing, texture reference, supports author from gray scale
Inspiration is found in figure, reduces time of the person of drawing a picture in design works details, and then can efficiently and conveniently assist painting.
For achieving the above object, the present invention the following technical schemes are provided:
A kind of method of the method reference gray level figure auxiliary drawing of reference gray level figure auxiliary drawing, comprising the following steps:
Training stage: training image is obtained, and converts grayscale image for training image, determines semantic type and color-language
Adopted corresponding relationship carries out semantic segmentation to training image according to color-semanteme corresponding relationship, semantic segmentation image is obtained, with language
Adopted segmented image is as input, using corresponding grayscale image as exporting, to the deep learning that can be realized image and be converted into image
Model carries out the training that iterates, the transformation model of acquisition semantic segmentation figure to grayscale image;
Application stage: according to color-semanteme corresponding relationship and semantic type, being intended to draw semantic segmentation figure according to drawing,
And the semantic segmentation figure of drafting is input to semantic segmentation figure into the transformation model of grayscale image, it is computed and generates multiple gray scales
Figure carries out texture rendering to semantic segmentation figure, enriches the interior of paint according to the grayscale image for selecting reference from grayscale image
Hold.
It in the present invention, assists painting using artificial intelligence technology, be arrived using deep learning model training semantic segmentation figure
Grayscale image obtains the model that semantic segmentation figure is converted to grayscale image.It is generated using trained model and meets the person's of drawing a picture intention
Grayscale image provide reference, can iterate generating process, support the person of drawing a picture constantly from model generate grayscale image in seek
Inspiration is looked for, reduces time of the person of drawing a picture in design works details, it is efficient, convenient, novel to have the characteristics that.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to do simply to introduce, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art, can be with root under the premise of not making the creative labor
Other accompanying drawings are obtained according to these attached drawings.
Fig. 1 is the flow diagram of the method for reference gray level figure auxiliary drawing of the present invention;
Fig. 2 is the example of semantic segmentation figure;
Fig. 3 is the example of grayscale image.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention more comprehensible, with reference to the accompanying drawings and embodiments to this
Invention is described in further detail.It should be appreciated that the specific embodiments described herein are only used to explain the present invention,
And the scope of protection of the present invention is not limited.
In order to promote the effect and efficiency of drawing, as shown in Figure 1, the present invention provides a kind of reference gray level figure auxiliary to paint
Method, including following procedure:
S101 obtains training image, converts grayscale image for training image by image gray processing algorithm.
In the present embodiment, training image can be from web search, artificial shooting or certain databases.General training image
Generally there is similar semantic composition, such as be all landscape painting, streetscape and high-rise building etc..In order to guarantee enough data instructions
Practice deep learning model, the training image of acquisition is not less than 5000.
Grayscale image is a kind of image that color and part details are eliminated on the basis of original image, but remains content
The information such as texture, light and shade in details, details richness is lower than original image but is above sketch.Grayscale image is as a kind of image class
There are two advantages for type tool, first is that can be generated by deep learning model, second is that having detailed information abundant, third is that can lead to
It crosses the execution algorithm on training image to obtain, features above determines that grayscale image is suitable as reference picture.
In the present embodiment, grayscale image is converted for training image (namely RGB color figure) using formula (1);
Gray (i, j)=0.299*R (i, j)+0.587*G (i, j)+0.144*B (i, j) (1)
Wherein, R (i, j) is R channel image, and G (i, j) is G channel image, and B (i, j) is channel B image, the gray scale of conversion
Figure is as shown in Figure 3.
S102 determines semantic type and color-semanteme corresponding relationship, according to color-semanteme corresponding relationship to training image
Semantic segmentation is carried out, semantic segmentation image is obtained.
Wherein, carrying out semantic segmentation to training image according to color-semanteme corresponding relationship includes:
Training image is split according to color corresponding to the object in image, to similar or identical in image
Object is labeled with same color, to obtain semantic segmentation image.This process can be it is artificial carry out semantic segmentation, can be with
It is to carry out semantic segmentation using the algorithm of setting, the semantic segmentation image after segmentation can be as shown in Figure 2.
In the present embodiment, classify to the object in training image, be always divided into several classes, each classification represents identical
Or similar object, its classification is indicated with fixed color.Wherein, similar to object refer to belong to of a sort object, example
Such as: birch, cypress all belong to trees this kind, can represent trees with green.In this manner it is possible to convert training image to only
The semantic segmentation image indicated with color in semantic segmentation image, indicates the area with identical semantic information with identical color
Domain.
S103, using corresponding grayscale image as output, is converted using semantic segmentation image as input to can be realized image
It is trained, iterates until model is restrained, acquisition semantic segmentation figure to grayscale image turns for the deep learning model of image
Change model.
In the present embodiment, selects pix2pixHD model as deep learning model, end is carried out to pix2pixHD model and is arrived
The training at end, the transformation model of acquisition semantic segmentation figure to grayscale image.Pix2pixHD model is a kind of condition generation confrontation net
Network is mainly used in image interpretation field.For task proposed by the present invention, the target of generator G is that semantic segmentation figure is raw
At corresponding grayscale image, and arbiter D is then: 1) distinguishing the grayscale image of true grayscale image and generation;2) judge grayscale image and language
Whether the mapping between adopted segmentation figure is correct.Pix2pixHD model is a kind of image interpretation model, is converted with powerful picture
Ability can learn from training picture to information such as texture, the light and shade rules of picture, it can convert semantic segmentation image
For grayscale image, thus abundant detailed information.
Specifically, the input of pix2pixHD model is that the one-hot vector of label figure indicates.Pix2pixHD model
Generator is made of two sub- generators, G1For global generator, G2For local enhancement generator, result figure is generated for expanding
Both the image size of picture is made of one group of convolutional network, residual error network and transposition convolutional network.In view of the gray scale of generation
Scheme figure for reference only to use, do not need excessively high display degree, therefore global generator G is used only1As generator, setting
Input and output size is all 256*512.Training process carries out on GPU, using Adam gradient descent method undated parameter, learning rate
Lr=0.0002, β1=0.5, β2=0.999.In 50 times initial iterative process, learning rate is remained unchanged, and is changed with subsequent
The increase linear attenuation of generation number.In 200 the number of iterations, model converges to stabilization.
S104 is intended to draw semantic segmentation figure according to drawing according to color-semanteme corresponding relationship and semantic type.
The usual mode of drawing is that first have a drawing to be intended to, and author has conceived the structure of entire paintings, then is drawn accurate
Content.It is relatively easy to conceive paintings structure, but be different space structure to mean that identical semantic chunk also can be different
Grain details, author can spend in the trial to details such as the specific texture of a certain partial content, light and shades a large amount of time and delete
On changing.Purpose is that the grayscale image obtained using the training stage generates the angle of model offer grayscale image to assist painting, first is that energy
The light and shade scheme of integral layout is provided, second is that the texture inspiration of content detail can be provided, third is that can be with by fuzzy grayscale image
It allows the person of drawing a picture to see the substantially effect of paintings rapidly to make corresponding modification, greatly shortens the time required to author draws a picture.
Author expresses drawing by way of drawing semantic segmentation figure and is intended to, and indicates to be laid out using semantic segmentation figure.It is semantic
Segmentation figure can indicate whole spatial structural form by the semantic information of the arrangement of color lump, the shape of color lump and representative.
The semantic segmentation figure of drafting is input to semantic segmentation figure into the transformation model of grayscale image, is computed life by S105
Meet the grayscale image of semantic segmentation content at multiple.
Specifically, multiple grayscale images of generation meet the layout of semantic segmentation figure on space structure, in detail content
Since the randomness that model generates has a little difference.The detail content of more Suresh Kumar can be provided by generating multiple grayscale images.
S106 chooses suitable grayscale image as the reference in detail content, to semantic segmentation figure from multiple grayscale images
Detail content carry out texture rendering, enrich paint content.
Specifically, the grayscale image that can be referred to is selected according to interest, includes light and shade, texture information in the grayscale image, presses
Light and shade, texture information according to grayscale image presentation, are filled semantic segmentation figure perfect.
Grayscale image has many characteristics, such as texture in details, light and shade, compares semantic segmentation figure, has details more abundant letter
Breath, these information are that model learns from training data, have certain reasonability, indicate that computer learning is arrived herein
Due texture, light and shade information.According to these grayscale images, author can therefrom find inspiration, enrich each semanteme of paintings
The details of block.
In order to preferably improve paint, the drawing householder method further include:
S107 repeats S104~S106, and each iteration further enriches the content of drawing, iterates, until improving whole
Open paintings.
Specifically, the paint after abundant details is repainted or is modified semantic segmentation image and be input to semantic segmentation
Figure is computed into the transformation model of grayscale image and generates multiple grayscale images, according to the gray scale that selection can refer to from grayscale image
Figure, i.e., repeatedly S104~S106 carry out texture rendering to perfect paint, further improve paint.Pass through
The renolation paint that constantly iterates can obtain final paint.
Above-mentioned drawing householder method generates grayscale image by the way that semantic segmentation figure to be converted into, and provides the content detail on drawing
On light and shade, texture reference, compensate for drawing auxiliary give particulars auxiliary on deficiency.Can iterate generating process,
Support the person of drawing a picture constantly to find inspiration from the grayscale image that model generates, reduce the person of drawing a picture conceive in works details when
Between, it is efficient, convenient, novel to have the characteristics that.
Technical solution of the present invention and beneficial effect is described in detail in above-described specific embodiment, Ying Li
Solution is not intended to restrict the invention the foregoing is merely presently most preferred embodiment of the invention, all in principle model of the invention
Interior done any modification, supplementary, and equivalent replacement etc. are enclosed, should all be included in the protection scope of the present invention.
Claims (7)
1. a kind of method of reference gray level figure auxiliary drawing, comprising the following steps:
Training stage: training image is obtained, and converts grayscale image for training image, determines that semantic type and color-semanteme are right
It should be related to, semantic segmentation is carried out to training image according to color-semanteme corresponding relationship, semantic segmentation image is obtained, with semanteme point
Image is cut as input, using corresponding grayscale image as exporting, to the deep learning model that can be realized image and be converted into image
Carry out the training that iterates, the transformation model of acquisition semantic segmentation figure to grayscale image;
Application stage: according to color-semanteme corresponding relationship and semantic type, it is intended to draw semantic segmentation figure according to drawing, and will
The semantic segmentation figure of drafting is input to semantic segmentation figure into the transformation model of grayscale image, is computed and generates multiple grayscale images, root
According to the grayscale image for selecting reference from grayscale image, texture rendering is carried out to semantic segmentation figure, enriches the content of paint.
2. the method for reference gray level figure auxiliary drawing as described in claim 1, which is characterized in that the training image of acquisition is not low
In 5000.
3. the method for reference gray level figure auxiliary drawing as described in claim 1, which is characterized in that will be trained using formula (1)
Image is converted into grayscale image;
Gray (i, j)=0.299*R (i, j)+0.587*G (i, j)+0.144*B (i, j) (1)
Wherein, R (i, j) is R channel image, and G (i, j) is G channel image, and B (i, j) is channel B image.
4. the method for reference gray level figure auxiliary drawing as described in claim 1, which is characterized in that corresponding according to color-semanteme
Relationship carries out semantic segmentation to training image
Training image is split according to color corresponding to the object in image, to the similar or identical object in image
It is labeled with same color, to obtain semantic segmentation image.
5. the method for reference gray level figure auxiliary drawing as described in claim 1, which is characterized in that selection pix2pixHD model
As deep learning model, pix2pixHD model is trained end to end, obtains the conversion of semantic segmentation figure to grayscale image
Model.
6. the method for reference gray level figure auxiliary drawing as described in claim 1, which is characterized in that can be joined according to interest selection
The grayscale image examined includes light and shade, texture information, the light and shade presented according to grayscale image, texture information, to semanteme in the grayscale image
Segmentation figure is filled perfect.
7. the method for reference gray level figure auxiliary drawing as described in any one of claims 1 to 6, which is characterized in that the drawing
Householder method further include:
Paint after abundant details is repainted or modified semantic segmentation image and is input to semantic segmentation figure to grayscale image
Transformation model in, be computed generate multiple grayscale images, according to selected from grayscale image reference grayscale image, to perfect
Paint carries out texture rendering, further improves paint.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910149284.3A CN109849576B (en) | 2019-02-28 | 2019-02-28 | Method for assisting drawing by referring to gray level diagram |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910149284.3A CN109849576B (en) | 2019-02-28 | 2019-02-28 | Method for assisting drawing by referring to gray level diagram |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109849576A true CN109849576A (en) | 2019-06-07 |
CN109849576B CN109849576B (en) | 2020-04-28 |
Family
ID=66899239
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910149284.3A Active CN109849576B (en) | 2019-02-28 | 2019-02-28 | Method for assisting drawing by referring to gray level diagram |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109849576B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110312139A (en) * | 2019-06-18 | 2019-10-08 | 深圳前海达闼云端智能科技有限公司 | The method and apparatus of image transmitting, storage medium |
CN110322529A (en) * | 2019-07-12 | 2019-10-11 | 电子科技大学 | A method of it is painted based on deep learning aided art |
CN111043953A (en) * | 2019-10-17 | 2020-04-21 | 杭州电子科技大学 | Two-dimensional phase unwrapping method based on deep learning semantic segmentation network |
CN111325212A (en) * | 2020-02-18 | 2020-06-23 | 北京奇艺世纪科技有限公司 | Model training method and device, electronic equipment and computer readable storage medium |
CN112001839A (en) * | 2020-07-23 | 2020-11-27 | 浙江大学 | Cross-domain image conversion method based on semantic feature transformation, computer device and storage medium |
CN113111906A (en) * | 2021-02-24 | 2021-07-13 | 浙江大学 | Method for generating confrontation network model based on condition of single pair image training |
CN116664773A (en) * | 2023-06-02 | 2023-08-29 | 北京元跃科技有限公司 | Method and system for generating 3D model by multiple paintings based on deep learning |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2905724A2 (en) * | 2014-02-07 | 2015-08-12 | Tata Consultancy Services Limited | Object detection system and method |
CN107944457A (en) * | 2017-11-23 | 2018-04-20 | 浙江清华长三角研究院 | Drawing object identification and extracting method under a kind of complex scene |
CN107945244A (en) * | 2017-12-29 | 2018-04-20 | 哈尔滨拓思科技有限公司 | A kind of simple picture generation method based on human face photo |
US10178982B2 (en) * | 2015-07-29 | 2019-01-15 | Perkinelmer Health Sciences, Inc. | System and methods for automated segmentation of individual skeletal bones in 3D anatomical images |
-
2019
- 2019-02-28 CN CN201910149284.3A patent/CN109849576B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2905724A2 (en) * | 2014-02-07 | 2015-08-12 | Tata Consultancy Services Limited | Object detection system and method |
US10178982B2 (en) * | 2015-07-29 | 2019-01-15 | Perkinelmer Health Sciences, Inc. | System and methods for automated segmentation of individual skeletal bones in 3D anatomical images |
CN107944457A (en) * | 2017-11-23 | 2018-04-20 | 浙江清华长三角研究院 | Drawing object identification and extracting method under a kind of complex scene |
CN107945244A (en) * | 2017-12-29 | 2018-04-20 | 哈尔滨拓思科技有限公司 | A kind of simple picture generation method based on human face photo |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110312139A (en) * | 2019-06-18 | 2019-10-08 | 深圳前海达闼云端智能科技有限公司 | The method and apparatus of image transmitting, storage medium |
CN110322529A (en) * | 2019-07-12 | 2019-10-11 | 电子科技大学 | A method of it is painted based on deep learning aided art |
CN111043953A (en) * | 2019-10-17 | 2020-04-21 | 杭州电子科技大学 | Two-dimensional phase unwrapping method based on deep learning semantic segmentation network |
CN111043953B (en) * | 2019-10-17 | 2021-07-27 | 杭州电子科技大学 | Two-dimensional phase unwrapping method based on deep learning semantic segmentation network |
CN111325212A (en) * | 2020-02-18 | 2020-06-23 | 北京奇艺世纪科技有限公司 | Model training method and device, electronic equipment and computer readable storage medium |
CN112001839A (en) * | 2020-07-23 | 2020-11-27 | 浙江大学 | Cross-domain image conversion method based on semantic feature transformation, computer device and storage medium |
CN112001839B (en) * | 2020-07-23 | 2022-09-13 | 浙江大学 | Cross-domain image conversion method based on semantic feature transformation, computer device and storage medium |
CN113111906A (en) * | 2021-02-24 | 2021-07-13 | 浙江大学 | Method for generating confrontation network model based on condition of single pair image training |
CN113111906B (en) * | 2021-02-24 | 2022-07-12 | 浙江大学 | Method for generating confrontation network model based on condition of single pair image training |
CN116664773A (en) * | 2023-06-02 | 2023-08-29 | 北京元跃科技有限公司 | Method and system for generating 3D model by multiple paintings based on deep learning |
CN116664773B (en) * | 2023-06-02 | 2024-01-16 | 北京元跃科技有限公司 | Method and system for generating 3D model by multiple paintings based on deep learning |
Also Published As
Publication number | Publication date |
---|---|
CN109849576B (en) | 2020-04-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109849576A (en) | A kind of method of reference gray level figure auxiliary drawing | |
CN110414519A (en) | A kind of recognition methods of picture character and its identification device | |
CN108830912A (en) | A kind of interactive grayscale image color method of depth characteristic confrontation type study | |
CN109190722A (en) | Font style based on language of the Manchus character picture migrates transform method | |
CN110458750A (en) | A kind of unsupervised image Style Transfer method based on paired-associate learning | |
CN110222722A (en) | Interactive image stylization processing method, calculates equipment and storage medium at system | |
CN109598279A (en) | Based on the zero sample learning method for generating network from coding confrontation | |
CN108986132A (en) | A method of certificate photo Trimap figure is generated using full convolutional neural networks | |
CN106023276A (en) | Pencil drawing making method and pencil drawing making device based on image processing | |
CN102289991A (en) | Visual-variable-based automatic classification and configuration method of map lettering | |
CN106960457A (en) | A kind of colored paintings creative method extracted and scribbled based on image, semantic | |
Xing et al. | Diffsketcher: Text guided vector sketch synthesis through latent diffusion models | |
CN109102457A (en) | A kind of intelligent color change system and method based on convolutional neural networks | |
CN115470365B (en) | Fine-grained cross-media retrieval method based on depth metric learning | |
CN113887737B (en) | Automatic sample set generation method based on machine learning | |
CN110298365A (en) | A kind of theme color extracting method based on human eye vision | |
CN114493997A (en) | Terrain wash painting generation method based on digital elevation model and style migration | |
Li et al. | Lbwgan: Label based shape synthesis from text with wgans | |
Han et al. | Innovative design of traditional calligraphy costume patterns based on deep learning | |
CN104268832B (en) | Gradient mesh recoloring method | |
Wu et al. | TEACHING KNOWLEDGE OF THE CHINESE NATIONAL ORCHESTRA AT THE SICHUAN CONSERVATORY OF MUSIC IN CHINA. | |
Yao et al. | Intelligent Data Collection and Aesthetic Activities-From the Aesthetic Perspective of Art Works | |
CN118097085B (en) | Method and system for automatically migrating topographic patterns and styles | |
Xu et al. | Comparing the design quality and efficiency between design intelligence and intermediate designers | |
Su et al. | Cartoon image colorization based on emotion recognition and superpixel color resolution |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |