CN116342739B - Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence - Google Patents

Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence Download PDF

Info

Publication number
CN116342739B
CN116342739B CN202310148885.9A CN202310148885A CN116342739B CN 116342739 B CN116342739 B CN 116342739B CN 202310148885 A CN202310148885 A CN 202310148885A CN 116342739 B CN116342739 B CN 116342739B
Authority
CN
China
Prior art keywords
image
artificial intelligence
painting
pictorial
generating
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.)
Active
Application number
CN202310148885.9A
Other languages
Chinese (zh)
Other versions
CN116342739A (en
Inventor
农长霖
张文晶
洪峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Qianhai Shenlei Semiconductor Co ltd
Original Assignee
Shenzhen Qianhai Shenlei Semiconductor Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Qianhai Shenlei Semiconductor Co ltd filed Critical Shenzhen Qianhai Shenlei Semiconductor Co ltd
Priority to CN202310148885.9A priority Critical patent/CN116342739B/en
Publication of CN116342739A publication Critical patent/CN116342739A/en
Application granted granted Critical
Publication of CN116342739B publication Critical patent/CN116342739B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/53—Querying
    • G06F16/535—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—2D [Two Dimensional] image generation
    • G06T11/20—Drawing from basic elements, e.g. lines or circles
    • G06T11/206—Drawing of charts or graphs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L15/18—Speech classification or search using natural language modelling
    • G10L15/1815—Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
    • G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
    • G10L25/63—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for estimating an emotional state
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G10L2015/225—Feedback of the input speech
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Computational Linguistics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Acoustics & Sound (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Computer Interaction (AREA)
  • Data Mining & Analysis (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Hospice & Palliative Care (AREA)
  • Child & Adolescent Psychology (AREA)
  • Signal Processing (AREA)
  • Medical Informatics (AREA)
  • Psychiatry (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Mathematical Physics (AREA)
  • Processing Or Creating Images (AREA)

Abstract

The application relates to an artificial intelligence drawing technology, and discloses a method for generating a plurality of drawing images based on artificial intelligence, which comprises the following steps: when any one of the display devices receives a voice drawing instruction, drawing elements are extracted from the voice drawing instruction; generating a first pictorial image with pictorial elements based on the first artificial intelligence model; generating a second pictorial image using the first pictorial image based on the second artificial intelligence model; the first pictorial image is displayed on a master device of the plurality of display devices and the second pictorial image is displayed on a slave device of the plurality of display devices. The application also discloses an electronic device and a computer readable storage medium. The application aims to facilitate a user to give drawing instructions so as to generate a plurality of AI drawing images with the same drawing type for a plurality of display devices to display.

Description

Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence
Technical Field
The present application relates to the field of artificial intelligence painting, and in particular, to a method, an electronic device, and a computer readable storage medium for generating multiple painting images based on artificial intelligence.
Background
The digital photo frame (Digital Photo Frame) is a display device for displaying digital photos rather than paper photos. The digital photo frame can acquire the pictures from the storage and display the pictures in a circulating mode, compared with a common photo frame, the digital photo frame is more convenient to display the pictures, and the display mode is flexible and changeable. With the rapid development of AI (Artificial Intelligence ) drawing technology nowadays, it is also possible to give drawing instructions to the AI by a user, and to use a display device for generating images by AI drawing and displaying the images on a digital photo frame.
Currently, in some medium-large-sized scenes (such as art houses, shops, etc.), there is a need to display images by using multiple display devices simultaneously to decorate the scene, and for the multiple display devices with similar positions, it is preferable to display images with the same painting type (such as painting theme and style) together, so that the appearance brought by people can be more aesthetic and artistic. If a plurality of display devices are required to generate AI painting images at the same time, users are required to issue corresponding painting instructions to each display device one by one, and the process is complicated; in addition, the generation of the AI drawing images has certain randomness, so that if the drawing instructions issued by the user do not make specific restrictions on the drawing subject or style, even if the AI drawing images displayed by the plurality of display devices are generated by the same artificial intelligence model aiming at the same drawing instruction, the drawing types among the AI drawing images are not necessarily the same or similar, and the requirement of decorating the same scene by using a plurality of images of the same drawing type is difficult to meet.
The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present application and is not intended to represent an admission that the foregoing is prior art.
Disclosure of Invention
The application mainly aims to provide a method for generating a plurality of painting images based on artificial intelligence, display equipment and a computer readable storage medium, which are used for facilitating a user to give painting instructions so as to generate a plurality of AI painting images with the same painting type for a plurality of display equipment to display.
In order to achieve the above object, the present application provides a method for generating a plurality of painting images based on artificial intelligence, comprising the steps of:
extracting drawing elements from voice drawing instructions when any one of the display devices receives the voice drawing instructions among a plurality of display devices associated with each other;
generating a first pictorial image using the pictorial element based on a first artificial intelligence model;
generating a second painting image by using the first painting image based on a second artificial intelligence model, wherein the second artificial intelligence model is trained in advance based on a plurality of groups of main body images and local images of the same type so as to learn the capability of generating the local images of the same type based on the main body images;
The first pictorial image is displayed on a master device of the plurality of display devices and the second pictorial image is displayed on a slave device of the plurality of display devices.
Optionally, the step of extracting the drawing element from the voice drawing instruction includes:
identifying semantic information and voice emotion information of the voice painting instruction;
inquiring a preset drawing element matched with the semantic information and inquiring a preset drawing element matched with the voice emotion information;
and taking the preset drawing elements obtained by inquiry as the drawing elements.
Optionally, before the step of generating the first drawing image by using the drawing element based on the first artificial intelligence model, the method further includes:
inquiring a preset drawing element matched with the display size of the main equipment, and taking the preset drawing element as the drawing element.
Optionally, before the step of generating the first drawing image by using the drawing element based on the first artificial intelligence model, the method further includes:
extracting audio features from the voice painting instruction, and determining a user type according to the audio features;
inquiring a preset drawing element matched with the user type as the drawing element.
Optionally, the drawing element corresponding to the first drawing image is a first drawing element; the step of generating a second pictorial image using the first pictorial image based on a second artificial intelligence model includes:
acquiring a second painting element according to the first painting image and the association characteristics between the main image and the local image learned by a second artificial intelligence model; the method comprises the steps of,
inquiring a preset drawing element matched with the display size of the slave device to serve as a second drawing element;
a second pictorial image is generated based on the second pictorial element.
Optionally, the number of the slave devices is plural; the step of generating a second pictorial image based on the second pictorial element includes:
and generating a second painting image corresponding to each slave device according to the second painting element matched with the display size of each slave device and the second painting element acquired based on the first painting image.
Optionally, before the step of generating the second drawing image using the first drawing image based on the second artificial intelligence model, the method further includes:
acquiring a plurality of painting images, and marking a main body image and a local image in each painting image by using a third artificial intelligent model;
Generating a training sample according to the main body image and the local image corresponding to each drawing image;
the second artificial intelligence model is trained based on a plurality of the training samples.
Optionally, the method for generating multiple painting images based on artificial intelligence further includes:
taking a display device which receives the voice painting instruction as the master device and taking other display devices except the master device as the slave devices;
or, the display device with the largest display size among the plurality of display devices is used as the master device, and other display devices except the master device are used as the slave devices.
To achieve the above object, the present application also provides an electronic device including: the system comprises a memory, a processor and a program which is stored in the memory and can run on the processor and is used for generating a plurality of drawing images based on artificial intelligence, wherein the program for generating the plurality of drawing images based on the artificial intelligence realizes the steps of the method for generating the plurality of drawing images based on the artificial intelligence when being executed by the processor.
To achieve the above object, the present application also provides a computer-readable storage medium having stored thereon a program for generating a plurality of drawing images based on artificial intelligence, which when executed by a processor, implements the steps of the method for generating a plurality of drawing images based on artificial intelligence as described above.
According to the method, the electronic device and the computer readable storage medium for generating the plurality of drawing images based on the artificial intelligence, provided by the application, a user can generate the plurality of drawing images with the same drawing type by only giving the voice drawing instruction to any one of the plurality of display devices so as to display the plurality of drawing images respectively, so that the user can give the drawing instruction conveniently, and the efficiency of generating the plurality of drawing images with the same type based on the artificial intelligence model can be improved.
Drawings
FIG. 1 is a schematic diagram of steps of a method for generating multiple pictorial images based on artificial intelligence in accordance with an embodiment of the present application;
fig. 2 is a schematic block diagram of an internal structure of an electronic device according to an embodiment of the present application.
The achievement of the objects, functional features and advantages of the present application will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below are exemplary and intended to illustrate the present application and should not be construed as limiting the application, and all other embodiments, based on the embodiments of the present application, which may be obtained by persons of ordinary skill in the art without inventive effort, are within the scope of the present application.
Referring to FIG. 1, in one embodiment, the method of generating a plurality of pictorial images based on artificial intelligence includes:
step S10, extracting drawing elements from voice drawing instructions when any one of a plurality of display devices associated with each other receives the voice drawing instructions;
step S20, generating a first painting image by using the painting elements based on the first artificial intelligent model;
step S30, generating a second painting image by using the first painting image based on a second artificial intelligence model, wherein the second artificial intelligence model is trained in advance based on a plurality of groups of main body images and local images of the same type so as to learn the capability of generating the local images of the same type based on the main body images;
step S40, displaying the first drawing image on a master device among the plurality of display devices, and displaying the second drawing image on a slave device among the plurality of display devices.
In this embodiment, a plurality of display devices disposed in the same scene may be connected to the same central control device or the same communication network, so as to be associated with each other; or multiple display devices may establish communication connections with each other to associate with each other.
Alternatively, the execution terminal of the embodiment may be any one of a plurality of display devices associated with each other, or may be a central control device that controls a plurality of display devices. The display device can be electronic equipment which is mainly used for displaying digital images, such as digital photo frames; the central control device can be deployed at a local end or at a cloud end (i.e. cloud central control).
As described in step S10, the display device is provided with a microphone module in addition to basic function modules such as display, storage and communication, and the microphone module can be used for receiving voice painting instructions sent by a user.
Optionally, when the user wants to send out a corresponding drawing instruction, so as to draw by using an artificial intelligence technology through the display device, and generate a plurality of drawing images of the same type to be displayed on a plurality of display devices respectively, the user can speak a corresponding voice drawing instruction to the display device in a voice acquisition range of any display device. The voice painting instruction taught by the user can comprise some key words for instructing the display device to execute artificial intelligent painting and some painting elements of a desired painting image; the same type of drawing image refers to the drawing image having the same theme or style.
The basic elements of the pictorial representation are: the basic principle of combining the elements into a complete work comprises diversity, unification, proportion, symmetry, balance, rhythm, comparison, harmony and the like; in addition, the painting elements can be painting styles, such as abstract paintings, oil paintings, cartoon paintings, ink paintings and the like; alternatively, the drawing element may be a drawing main content such as a person (e.g., a public person, a famous person, a specific person, etc.), a scene (e.g., a war scene, a sports scene, a four seasons scene, etc.), a landscape (e.g., mountain, waterfall, desert, etc.), etc. That is, the drawing element can be determined as long as it is an element contributing to the formation of the drawing content.
For example, the voice painting instruction taught by the user may be "please generate a picture of the oil painting style, the picture including blue sky and white cloud, mountain, river. The terminal can refine the instruction of generating the photo and identify the instruction as an artificial intelligent drawing instruction, and keywords such as oil painting style, blue sky and white cloud, mountain and river can be used as drawing elements.
Optionally, when any one of the plurality of display devices associated with each other receives a voice drawing instruction sent by a user, if the display device is the embodiment terminal, the display device is responsible for extracting drawing elements according to the voice drawing instruction; if the display device is not the embodiment terminal, the display device sends the received voice painting instruction to the embodiment terminal, and the embodiment terminal is responsible for extracting the painting elements from the voice painting instruction.
If the terminal is any one of a plurality of display devices associated with each other, and the computing capability of the terminal configuration is insufficient, the drawing element extraction can be completed by using the computing power of the local end device or the cloud end device which is in communication connection with the display device.
Optionally, when the terminal extracts the drawing element from the voice drawing instruction, the terminal may identify semantic information and voice emotion information of the voice drawing instruction, and generate the drawing element according to the semantic information and the voice emotion information. The terminal can identify semantic information and voice emotion information of the voice painting instruction, inquire preset painting elements matched with the semantic information, screen the preset painting elements obtained by inquiry by utilizing the voice emotion information, and take the remained preset painting elements after screening as the painting elements; or, the terminal may identify the semantic information and the voice emotion information of the voice painting instruction, then query the preset painting elements matched with the semantic information and query the preset painting elements matched with the voice emotion information respectively, and then use all the preset painting elements obtained by query together as the painting elements.
It should be noted that, the terminal may identify text information in the voice command by using a voice recognition technology, and then identify semantic information and voice emotion information from the text information based on the semantic recognition technology and the voice emotion recognition technology. The semantic information obtained by terminal identification at least comprises a keyword, the voice emotion information at least comprises one emotion, the emotion can be roughly divided into positive emotion, negative emotion and common emotion (neutral emotion), the positive emotion is emotion generated by increasing positive value or decreasing negative value of human beings such as pleasure, high excitement, feeling, celebration and the like, the negative emotion is emotion generated by decreasing positive value or increasing negative value of human beings such as pain, difficulty, loss and the like, and the common emotion is emotion expressed by no strong positive or negative emotion of human beings.
Optionally, the terminal presets a plurality of preset drawing elements, and each preset drawing element is associated with at least one keyword. Since a word may have a plurality of paraphrasing, the same keyword may be associated with a plurality of preset drawing elements.
Optionally, after the terminal obtains the semantic information and the voice emotion information, each keyword in the semantic information can be screened, and then a database is queried for preset drawing elements associated with the extracted keywords. In view of the fact that when emotion of a user is different for the same word, different definitions may exist, therefore when a plurality of preset drawing elements are associated with the same keyword, the terminal can assign corresponding emotion labels to the preset drawing elements in advance, then after the preset drawing elements associated with each keyword in semantic information are queried, the preset drawing elements are further screened by utilizing voice emotion information, so that preset drawing elements in which emotion labels accord with emotion in the voice emotion information are screened, and finally the screened residual preset drawing elements are used as the drawing elements. Therefore, the painting elements which are more suitable for the mind of the user can be extracted according to the user semantics, and the painting image generated by the artificial intelligence technology based on the extracted painting elements in the follow-up process is more suitable for the mind of the user (namely, the painting image which is suitable for the mind of the user is obtained).
Or after the terminal obtains the semantic information and the voice emotion information, each keyword in the semantic information can be screened, then the database is queried for preset drawing elements associated with the extracted keywords, keywords corresponding to emotion described by the voice emotion information are queried, and the database is queried for preset drawing elements associated with the keywords corresponding to the voice emotion information. And finally, the terminal uses the preset drawing elements matched with the semantic information and the preset drawing elements matched with the voice emotion information together as the drawing elements. The subsequent drawing image generated by utilizing the artificial intelligence technology based on the extracted drawing elements can be integrated into the emotion elements of the user, and the emotion of the user can be expressed in a drawing artistic form.
It should be noted that, the terminal may extract one or more drawing elements based on a voice drawing instruction.
As described in step S20, the present embodiment constructs an artificial intelligence model (labeled as a first artificial intelligence model) dedicated to performing artificial intelligence drawing in advance based on an artificial intelligence technique and a machine learning model (may be a diffusion probability model), and the first artificial intelligence model may be deployed in a terminal. Of course, in order to save storage and calculation power of the terminal, the first artificial intelligent model may be deployed in a local end device or a cloud end device which is in communication connection with the terminal, and the terminal may invoke the first artificial intelligent model deployed in the local end device or the cloud end device through data interaction with the first artificial intelligent model; the following description will take the deployment of the first artificial intelligence model in the terminal as an example.
It should be understood that, because the open source technology exists in the artificial intelligence painting, the logic and mode of the training of the related model are not described herein, and the related engineer may perform optimization and improvement on the basis of the original trained artificial intelligence painting model, adjust some related parameters, and then obtain the first artificial intelligence model suitable for the embodiment; or building a model assembly related to artificial intelligence drawing based on a deep learning framework which is gradually improved, so that a first artificial intelligence model is obtained. Of course, the relevant engineers may write and train the first artificial intelligence model from scratch, as conditions allow.
Optionally, after the terminal extracts the drawing elements, the drawing elements can be input into a first artificial intelligent model which is deployed and trained in advance, so that the first artificial intelligent model can generate corresponding image data by matching drawing features related to the drawing elements (i.e. image features related to drawing) and fitting the distribution of the drawing features, and finally, a generated drawing image is obtained and output (the drawing image generated by the first artificial intelligent model can be marked as a first drawing image).
As described in step S30, the present embodiment constructs an artificial intelligence model (labeled as a second artificial intelligence model) usable for image feature learning and artificial intelligence drawing based on the artificial intelligence technology and the machine learning model in advance, and the second artificial intelligence model may be deployed in the terminal. Of course, in order to save storage and calculation power of the terminal, the second artificial intelligent model may be deployed in a local end device or a cloud end device which is in communication connection with the terminal, and the terminal may invoke the second artificial intelligent model deployed in the local end device or the cloud end device through data interaction with the second artificial intelligent model; the deployment of the second artificial intelligence model at the terminal is described below as an example.
Alternatively, the second artificial intelligence model may be an image feature generation unit and a pictorial image generation unit, respectively, wherein the image feature generation unit may be constructed based on a convolutional neural network; the drawing image generating unit may be a drawing image generating unit constructed by using a diffusion probability model as in constructing and training the first artificial intelligence model, or the drawing image generating unit may be a model calling unit (having a function of calling other artificial intelligence models), that is, a function of realizing artificial intelligence drawing by calling the first artificial intelligence model.
Optionally, each training sample used in the training image feature generating unit includes a main image and at least one local image, and the drawing types of the main image and the local image in the same training sample are the same.
Alternatively, the subject image and the partial image in the same training sample may be derived from the same pictorial image. When collecting a plurality of drawing images for generating a plurality of training samples (such as collecting one thousand drawing images and one ten thousand drawing images), the relevant engineers responsible for establishing the second artificial intelligence model can mark the whole image in the same drawing image as a main body image and mark each local area in the same drawing image as a local image (the mode can be applied to drawing images with unobtrusive drawing targets, such as some abstract drawings without a so-called primary-secondary relationship and mountain-water drawings); and/or, the related engineer distinguishes the main body region and the partial region from one drawing image based on drawing criteria provided by a professional painter, and marks the main body image and at least one partial image accordingly (this manner may be applied to drawing target highlighting or explicit drawing images such as explicit drawing characters (e.g., characters, animals, plants, etc.); taking a single character drawing as an example, dividing the display region of the character in the drawing into the main body region, and dividing the other region except the main body region into at least one partial region).
Alternatively, the subject image and the partial image in the same training sample may be derived from different pictorial images, so long as the pictorial images are the same or similar in type (e.g., two pictorial images of different sunflowers); when the main image and the partial image in the same training sample are derived from different drawing images, a drawing image with a large image size may be used as the main image, and a drawing image with a small image size may be used as the partial image.
Optionally, after the relevant engineers generate a plurality of training samples, the training samples may be input to a terminal, the terminal inputs the training samples to the second artificial intelligence model for performing iterative training for a plurality of times, and the second artificial intelligence model continuously learns the correlation characteristics between the main body image and the local image in each training sample from the second artificial intelligence model (i.e. for each training sample, the second artificial intelligence model may extract global image characteristics from the main body image and extract local image characteristics from the local image, and establish the correlation between the global image characteristics and the local image characteristics), so that when the number of training samples and the training times are enough, the image characteristic generating unit in the second artificial intelligence model gradually converges, and after the training of the image characteristic generating unit is completed, the second artificial intelligence model learns to obtain the capability of generating the local image of the same type based on the main body image.
It should be noted that, whether the neural network model reaches the convergence standard is generally detected whether the key value of the model reaches a threshold preset by an engineer, and if the key value meets the relevant threshold (i.e. exceeds the relevant threshold), the consistency of the input result and the output result of the model is higher. However, in view of the fact that the first and second painting images only need to be generated in the same manner, but the painting elements other than the painting elements related to the painting types do not need to be identical, the painting types between the first and second painting images only need to be set to satisfy a higher similarity (such as between 85% and 95%), in addition, the other painting elements between the first and second painting images can be set to satisfy a lower similarity (such as between 30% and 50%), so that the second and first painting images obtained later can satisfy the same painting type, and different painting contents exist (such as both painting types can be blue subjects, but one painting content is blue sea, and the other is blue sky), so that a plurality of painting images of the same type can be obtained later to decorate the same scene, and a plurality of painting images can also have diversity between the contents, so that a plurality of thousands of painting images can not be generated.
Optionally, after the first artificial intelligence model generates the first drawing image based on the drawing element, the terminal may add the label of the main image to the first drawing image, then input the first drawing image into a second artificial intelligence model that has been trained in advance, use the second artificial intelligence model to take the first drawing image as the main image, and use the correlation feature between the main image and the local image learned previously to match the image feature of the local image corresponding to the first drawing image, and then use the second artificial intelligence model to generate the second drawing image according to the matched image feature.
Therefore, the generated first and second painting images are identical in painting type, and the painting contents of the first and second painting images are associated with each other to a certain extent (namely, the association between the parts and the main body) because the second painting image is more important to reflect the relevant characteristics of the partial image corresponding to the first painting image rather than the relevant characteristics of the whole first painting image (the second painting image and the first painting image are identical in painting type, but the specific painting contents of the second painting image and the first painting image are different from each other), namely, the painting style or the painting content of each second painting image can be identical to or similar to the partial image in the first painting image, and when the first painting image and the second painting image are displayed on different display devices, the second painting image can be equivalent to the partial image of the first painting image displayed on another display device, and a scene is decorated together with the first painting image.
For example, when the main image of the first drawing image is a shepherd who swings the whips to repel the flocks of sheep, and the partial image is a flocked sheep, another flocked sheep may be drawn in the generated second drawing image, so that the second drawing image and the first drawing image are set off mutually, and the drawing images displayed on a plurality of different display devices may be combined into the same scroll.
Wherein, in a plurality of display devices that are associated with each other, a distinction is made between one master device and at least one slave device.
Optionally, if the number of slave devices is only one, the second artificial intelligence model only needs to generate a second painting image; if the number of the slave devices is plural, the second artificial intelligence model needs to generate a plurality of second painting images corresponding to the number of the slave devices.
Alternatively, when the second artificial intelligence model needs to generate a plurality of second painting images, the second painting images may be generated successively based on the first painting images. It should be noted that, because the machine learning has a certain randomness, and the internal network can be optimized once more every time the artificial intelligent model runs, even if the input of the model is the same, the plurality of second drawing images sequentially output may not be completely the same.
Or when the second artificial intelligent model needs to generate a plurality of second painting images, a first second painting image can be generated based on the first painting image, then the second painting image output by the second artificial intelligent model each time is fed back to the input end of the second artificial intelligent model, a new second painting image is generated by the second artificial intelligent model based on the second painting image generated last time (for example, the first second painting image is used as the input of the second artificial intelligent model, the second painting image is generated by the second artificial intelligent model based on the first second painting image, and a third second painting image can be generated based on the second painting image, and N different second painting images can be obtained by cycling for N-1 times).
In step S40, after the terminal obtains the first drawing image and the second drawing image, the first drawing image may be output to a master device of the multiple display devices for display, and the second drawing image may be output to a slave device of the multiple display devices for display (i.e., the master device is used for displaying the first drawing image and the slave device is used for displaying the second drawing image).
In an embodiment, a user only needs to issue a voice drawing instruction to any one of the multiple display devices, so that multiple drawing images with the same drawing type (the drawing images are different in drawing content except for the drawing type) can be generated, and the multiple display devices can display the drawing images respectively, so that the user can issue the drawing instruction conveniently, the efficiency of generating multiple drawing images with the same type based on the artificial intelligence model can be improved (namely, the user does not need to issue corresponding drawing instructions to each display device respectively, and multiple drawing images with the same type and different drawing content can be generated only by issuing the drawing instruction once.
In addition, because traditional AI painting generally utilizes intelligent equipment (such as a computer) with relatively complete functions to input painting elements to AI in a text input mode so as to generate AI painting images, for some display equipment (such as a digital photo frame) with relatively single functions and mainly used for displaying images, the embodiment extracts corresponding painting elements by realizing voice painting instructions received based on the display equipment so as to generate corresponding painting images, and therefore, when a user wants to respectively display a plurality of painting images with the same painting type and different painting contents on a plurality of display equipment, the user can realize the requirement by only teaching one voice painting instruction to any display equipment, and the user can conveniently issue the painting instructions.
In an embodiment, based on the foregoing embodiment, before the step of generating the first drawing image using the drawing element based on the first artificial intelligence model, the method further includes:
inquiring a preset drawing element matched with the display size of the main equipment, and taking the preset drawing element as the drawing element.
In this embodiment, a plurality of preset size intervals (the numerical ranges of the preset size intervals are different) are divided in advance, and each preset size interval is associated with a corresponding preset drawing element. For example, for some preset size intervals with large values, preset drawing elements related to keywords such as ambitious, broad and the like can be associated; for some preset size intervals with small values, preset drawing elements related to warm, fine, exquisite and other keywords can be associated.
Optionally, when the terminal extracts the first drawing element from the voice drawing instruction, the terminal matches the corresponding preset drawing element according to a preset size interval where the display size of the main device is located, and the corresponding preset drawing element is used as the first drawing element together. And then the terminal inputs all the first drawing elements into the first artificial intelligent model together to generate a first drawing image.
It should be noted that, the existing artificial intelligence drawing technology does not pay attention to the content in the AI drawing image, and the inherent relationship between the display size of the device for displaying the AI drawing image; even though the existing artificial intelligence painting technology supports a user to set the size of the output AI painting image, the setting of the image size does not change the painting content in the image, and the setting of different sizes only cuts out different sizes for the image with the same painting content; compared with the prior art, the present embodiment fully captures the relevance between the generated AI painting content and the display size of the display device (for example, for large display sizes, some preset painting elements related to ambitious painting scenes can be generally associated, the preset painting elements can be represented by adding more characters and scenes, arranging larger scene scenes and designing historical famous scenes, for small display sizes, painting settings of some exquisite or single characters and objects can be associated as preset painting elements to highlight main characters or objects in a limited space), and corresponding preset painting elements are allocated for different display sizes, so that the artificial intelligent model can learn painting features corresponding to different display sizes, and when a painting image is generated, the painting features corresponding to the painting elements extracted in a voice command can be fitted, so that the finally generated painting image can reflect the painting features related to the display size (even if the content of the generated painting image is influenced by the display size of the display device, and when the display size is changed, the painting features can influence the final painting image with the same voice command can be generated, and the first artificial intelligent model can also be displayed under the condition that the same painting image is different from the first artificial intelligent model.
In view of the fact that the user's painting intention is sometimes also affected by the display size of the display device (if the display size of the display device is large, the user may want to obtain a more magnificent and aerodynamic drawing in the scene, and if the display size of the display device is small, the user may want to make a more exquisite drawing with emphasis on a single figure or object), for example, the user stands in front of the display device with different display sizes, may have different painting intentions, and the experience is easy to express for the user with a certain painting skill, but is difficult to express for a general user or a user with a lack of painting foundation, often thinks about how to express the corresponding painting intention, even if the experience is more, the user only stays in the subconscious layer, and the embodiment can make a certain supplement to the user's painting intention in the aspect by fully capturing the relevance between the content of the AI drawing and the display size of the display device, further improves the artificial painting intention of the instruction to be expressed in the current environment, and the intelligent model is more convenient for the user to generate the user's painting intention to display the voice image.
Compared with the existing artificial intelligence painting model, the improvement made by the first artificial intelligence model provided by the embodiment comprises the steps of distributing corresponding weights for various painting elements, namely, the painting elements corresponding to voice painting instructions, and the painting elements corresponding to the display size of the main equipment are respectively matched with the corresponding weights (the existing artificial intelligence painting does not distinguish the weights of the painting elements in this way, and more common conditions are that the weights of all the painting elements are consistent).
The weight rule set by the first artificial intelligence model provided in this embodiment includes: the weight (which may be marked as a first weight) allocated to the drawing element corresponding to the voice drawing instruction is greater than the weight (which may be marked as a second weight) allocated to the drawing element corresponding to the display size of the main device, for example, the ratio of the first weight to the second weight may be 0.7:0.3, 0.6:0.4, 0.8:0.2, or the like. And even if the related engineer sets the initial ratio of the first weight and the second weight in the first artificial intelligent model, the first artificial intelligent model can automatically adjust and optimize the ratio between the first weight and the second weight according to the feedback of training or learning results in the subsequent operation, but the adjusted ratio cannot violate the weight rule, such as setting the first weight to be greater than or equal to 0.51.
In the logic for training and learning the object by the artificial intelligence model, if the weight assigned to the object is larger, the relevant feature of the object is more focused and learned, so that the relevant feature of the object is more reflected in the result output subsequently.
In view of the fact that the voice drawing instruction is the most subjective drawing intention expression of the user, more weight is allocated to the drawing elements corresponding to the voice drawing instruction, so that the first drawing image generated by the first artificial intelligent model is focused on expressing the characteristics associated with the voice drawing instruction, and the drawing image more conforming to the mind of the user can be obtained.
In an embodiment, based on the foregoing embodiment, before the step of generating the first drawing image using the drawing element based on the first artificial intelligence model, the method further includes:
extracting audio features from the voice painting instruction, and determining a user type according to the audio features;
inquiring a preset drawing element matched with the user type as the drawing element.
In this embodiment, the first drawing element may further include, in addition to the drawing element extracted from the voice drawing instruction (in some embodiments, the first drawing element may further include a drawing element corresponding to a display size of the main device), a drawing element corresponding to a user type related to the voice drawing instruction.
Optionally, the terminal may extract the audio feature of the voice drawing instruction by performing audio analysis on the voice drawing instruction, and then perform user portrait according to the extracted audio feature, so as to obtain a user type corresponding to the user who sends the voice drawing instruction.
It should be understood that, because people of different ages and sexes generally have corresponding audio characteristics, and the combination of the big data analysis technology can obtain representative audio characteristics corresponding to people of different ages and sexes. After the preset audio features corresponding to the people of all ages and sexes are stored in the terminal, when the user type identification is needed, after the audio features corresponding to the user to be identified are extracted, the user type corresponding to the user to be identified can be obtained by inquiring the preset audio features matched with the user to be identified in the database and further according to the crowd type associated with the voice audio features obtained by inquiry.
Because different types of users generally have corresponding drawing preferences, corresponding drawing styles can be configured for various user types in advance to serve as preset drawing elements, and after the user type of the user corresponding to the voice drawing instruction is determined, the preset drawing elements matched with the user type can be queried in a database and added into the first drawing elements. For example, when the user type is the elderly, the associated preset drawing element may be a wash painting or a landscape painting style; when the user type is adult female, the associated preset drawing element can be oil painting and impression painting style; when the user type is child, the associated preset drawing element may be a cartoon, a simple drawing style.
Optionally, after obtaining the first drawing elements of various types, the terminal inputs the first drawing elements together into the first artificial intelligence model to generate the first drawing image.
It should be noted that, the existing artificial intelligence drawing technology does not pay attention to the internal connection between the drawing instruction issuing person and the drawing content, so that the situation that different types of users want to obtain different drawing contents even if the users issue the same drawing instruction is ignored; in this embodiment, through fully capturing the relevance between the user type and the AI painting content, and allocating corresponding preset painting elements for different user types, the first artificial intelligence model may learn the painting features corresponding to different user types, and fit the painting features with the painting features extracted from the voice painting instruction (in some embodiments, the painting elements corresponding to the display size of the main device may further be included) when generating the first painting image, so that the finally generated painting image may reflect the painting features related to the user types (that is, the content of the generated painting image is affected by the user types, and when the user types are changed, the content finally generated by AI may also be affected), so that the AI painting image more accords with the mind of the user of the corresponding type.
Of course, since the voice painting instruction is after all the most subjective painting intention expression of the user, the weight (which may be marked as the third weight) assigned by the first artificial intelligence model to the painting element related to the user type should also satisfy the set weight rule: the third weight is less than the first weight.
As for the third weight and the second weight, both may be set equal, or the third weight may be set smaller or larger than the second weight.
In an embodiment, based on the foregoing embodiment, the drawing element corresponding to the first drawing image is a first drawing element; the step of generating a second pictorial image using the first pictorial image based on a second artificial intelligence model includes:
acquiring a second painting element according to the first painting image and the association characteristics between the main image and the local image learned by a second artificial intelligence model; the method comprises the steps of,
inquiring a preset drawing element matched with the display size of the slave device to serve as a second drawing element;
a second pictorial image is generated based on the second pictorial element.
In the present embodiment, the drawing elements for generating the first drawing image are collectively referred to as a first drawing element, and the drawing elements for generating the second drawing image are collectively referred to as a second drawing element.
Optionally, after obtaining the first painting image based on the first painting element, the terminal may add the label of the main image to the first painting image, then input the first painting image into a second artificial intelligence model that has been trained in advance, use the first painting image as the main image by the second artificial intelligence model, and utilize the association feature between the main image and the local image learned previously to match out the image feature of the local image corresponding to the first painting image, and then obtain the painting element corresponding to the image feature of the local image by the second artificial intelligence model according to the association between the preset image feature and the painting element as the second painting element; meanwhile, the terminal also matches corresponding preset drawing elements according to a preset size interval where the display size of the slave device is located, and the corresponding preset drawing elements are used as two drawing elements together.
It should be noted that, one of the training emphasis of the artificial intelligence drawing model is to train the relevance between various drawing elements and various image features continuously, where the relevance is a key that the artificial intelligence drawing model can create a corresponding AI drawing image according to the combination of various drawing elements, and after all, the AI drawing image is fitted by a plurality of drawing-related image features. Thus, the pictorial image generation unit in either the pre-trained first or second artificial intelligence model may have the ability to invoke the association between pictorial elements and image features.
Optionally, after the second drawing element is obtained from the first drawing image and the second drawing element matching the display size of the slave device is obtained, the second artificial intelligence model generates a second drawing image based on all the second drawing elements.
Like this, the second drawing image that the second artificial intelligence model finally produced, except possessing the drawing type the same with first drawing image, can also embody the drawing characteristic correlated with display size from equipment, richened the drawing content of second drawing image, realize based on a speech painting instruction, can produce first drawing image and second drawing image that the drawing type is the same, specific drawing content has certain difference, like this when the user wants to demonstrate the drawing image that a plurality of drawing types are the same, specific drawing content has certain difference respectively on a plurality of display devices, only need to give out a speech painting instruction (need not to design corresponding drawing instruction for every display device respectively), greatly made things convenient for the user to give off to the drawing instruction.
Optionally, if there are a plurality of slave devices, the second artificial intelligence model needs to generate a plurality of second drawing images corresponding to the number of slave devices. The second artificial intelligence model can generate a second painting image corresponding to each slave device according to a second painting element matched with the display size of each slave device and according to the second painting element acquired based on the first painting image.
Optionally, when the second artificial intelligence model needs to generate a plurality of second drawing images, the second drawing images may be sequentially generated based on the first drawing images, so as to obtain a second drawing image corresponding to each slave device.
When the terminal generates corresponding second drawing images for each slave device by using the second artificial intelligent model, the display sizes of the slave devices are respectively obtained, and preset drawing elements corresponding to the slave devices are matched according to the display sizes corresponding to the slave devices to serve as the second drawing elements corresponding to the slave devices. And the second artificial intelligence model may also obtain a second drawing element based on the first drawing image and associate such second drawing element with each slave device.
Optionally, when the second artificial intelligence model generates the corresponding second drawing image for each slave device, the second artificial intelligence model generates the second drawing image corresponding to each slave device according to the second drawing element specifically associated with each slave device.
Like this, the final produced many second painting images of second artificial intelligence model except possess the same painting type with first painting image, still can embody the drawing characteristic that the slave device of different display sizes is relevant, obtain various second painting images and demonstrate respectively on a plurality of slave devices, both compromise the demand of producing many painting images that the painting type is the same, the painting content is different, the while still satisfied the variety demand to the second painting image that artificial intelligence model produced, make the user that has relevant demand only need to give down one and to obtain many painting images that the painting type is the same, the painting content is different and demonstrate on a plurality of display devices, need not the user to give down corresponding painting instruction to each display device one by one, greatly made things convenient for the user to the giving down of painting instruction, and the quantity of display device is more, this kind of high-efficient advantage of producing many AI painting images is more obvious.
In an embodiment, based on the foregoing embodiment, before the step of generating the second drawing image using the first drawing image based on the second artificial intelligence model, the method further includes:
acquiring a plurality of painting images, and marking a main body image and a local image in each painting image by using a third artificial intelligent model;
generating a training sample according to the main body image and the local image corresponding to each drawing image;
the second artificial intelligence model is trained based on a plurality of the training samples.
In this embodiment, an artificial intelligence model (labeled as a third artificial intelligence model) that can be used to generate training samples for the second artificial intelligence model is constructed in advance based on the artificial intelligence technique and the machine learning model, and the third artificial intelligence model can be deployed in the terminal. Of course, in order to save storage and calculation power of the terminal, the third artificial intelligent model may be deployed in a local end device or a cloud end device which is in communication connection with the terminal, and the terminal may invoke the third artificial intelligent model deployed in the local end device or the cloud end device through data interaction with the third artificial intelligent model.
Alternatively, when constructing the third artificial intelligence model, the relevant engineer may set the labeling rules of the main image and the partial image in the drawing image in the third artificial intelligence model, and manually construct a number (the number is denoted as Z1) of training samples (each training sample includes one drawing image and the main image and at least one partial image are labeled in the drawing image) of the second artificial intelligence model, which is less than the total number of samples (denoted as Z2) for training the second artificial intelligence model, and may be 5% -30% of the total number of samples (i.e., Z1 may be Z2 equal to 5% -30%), and input the training samples into the third artificial intelligence model for multiple iterative training so that the third artificial intelligence model can learn the ability to label the main image and the partial image in the drawing image.
Alternatively, after the training of the third artificial intelligence model is completed, the terminal may automatically acquire a plurality of drawing images (the number of the drawing images is greater than or equal to Z2-Z1, and the contents of the drawing images may be various) from the internet, and input the drawing images one by one into the pre-trained third artificial intelligence model, and the third artificial intelligence model identifies the subject image and the partial image in each drawing image one by one.
Optionally, after the terminal invokes the third artificial intelligence model to mark the main image and the local image in each drawing image, each marked drawing image may be updated to a training sample of the second artificial intelligence model.
Optionally, the terminal inputs the automatically generated training samples and the training samples manually created by the related engineers into the second artificial intelligence model together to train the image feature generating unit in the second artificial intelligence model.
In this way, the cost of manually creating the training sample of the second artificial intelligence model can be saved, the efficiency of generating the training sample of the second artificial intelligence model is improved, and the efficiency of training the second artificial intelligence model is further improved.
In an embodiment, based on the foregoing embodiment, the method for generating a plurality of drawing images based on artificial intelligence further includes:
taking a display device which receives the voice painting instruction as the master device and taking other display devices except the master device as the slave devices;
or, the display device with the largest display size among the plurality of display devices is used as the master device, and other display devices except the master device are used as the slave devices.
In this embodiment, among the plurality of display devices associated with each other, one master device and at least one slave device are distinguished.
Alternatively, when any one of the plurality of display devices associated with each other receives a language drawing instruction issued by a user, the terminal may be a display device that receives the voice drawing instruction as a master device and other display devices other than the master device as slave devices.
Alternatively, the terminal may acquire the display size of each display device, and then use the display device in which the display size is largest as the master device, and use other display devices than the master device as the slave devices.
Referring to fig. 2, an embodiment of the present application further provides an electronic device, where an internal structure of the electronic device may be as shown in fig. 2. The electronic device includes a processor, a memory, a communication interface, and a database connected by a system bus. Wherein the processor is configured to provide computing and control capabilities. The memory of the electronic device includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the electronic device is used for storing a program for generating a plurality of painting images based on artificial intelligence. The communication interface of the electronic device is used for carrying out data communication with an external terminal. The input device of the electronic device is used for receiving signals input by external equipment. The computer program is executed by a processor to implement a method of generating a plurality of pictorial images based on artificial intelligence as described in the embodiments above.
It will be appreciated by those skilled in the art that the structure shown in fig. 2 is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation of the display device to which the present application is applied.
Alternatively, the electronic device may be a display device, or a central control device that controls the display device.
Furthermore, the present application also proposes a computer-readable storage medium including a program for generating a plurality of drawing images based on artificial intelligence, which when executed by a processor, implements the steps of the method for generating a plurality of drawing images based on artificial intelligence as described in the above embodiments. It is understood that the computer readable storage medium in this embodiment may be a volatile readable storage medium or a nonvolatile readable storage medium.
In summary, in the method, the electronic device and the computer readable storage medium for generating a plurality of drawing images based on artificial intelligence provided by the embodiments of the present application, a user may generate a plurality of drawing images with the same drawing type by only sending a voice drawing instruction to any one of a plurality of display devices, so that the plurality of display devices may display the drawing images respectively, which is convenient for the user to send the drawing instruction, and may improve efficiency of generating a plurality of drawing images with the same type based on an artificial intelligence model.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium provided by the present application and used in embodiments may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, apparatus, article or method that comprises the element.
The foregoing description is only of the preferred embodiments of the present application and is not intended to limit the scope of the application, and all equivalent structures or equivalent processes using the descriptions and drawings of the present application or direct or indirect application in other related technical fields are included in the scope of the present application.

Claims (10)

1. A method for generating a plurality of pictorial images based on artificial intelligence, comprising:
extracting drawing elements from voice drawing instructions when any one of the display devices receives the voice drawing instructions among a plurality of display devices associated with each other;
Generating a first pictorial image using the pictorial element based on a first artificial intelligence model;
generating a second painting image by using the first painting image based on a second artificial intelligence model, wherein the second artificial intelligence model is trained in advance based on a plurality of groups of main body images and local images of the same type so as to learn the capability of generating the local images of the same type based on the main body images;
the first pictorial image is displayed on a master device of the plurality of display devices and the second pictorial image is displayed on a slave device of the plurality of display devices.
2. The method of generating a plurality of pictorial images based on artificial intelligence of claim 1 wherein the step of extracting pictorial elements from the voice pictorial instructions comprises:
identifying semantic information and voice emotion information of the voice painting instruction;
inquiring a preset drawing element matched with the semantic information and inquiring a preset drawing element matched with the voice emotion information;
and taking the preset drawing elements obtained by inquiry as the drawing elements.
3. The method of generating a plurality of pictorial images based on artificial intelligence of claim 1, wherein prior to the step of generating a first pictorial image using the pictorial elements based on a first artificial intelligence model, further comprising:
Inquiring a preset drawing element matched with the display size of the main equipment, and taking the preset drawing element as the drawing element.
4. The method of generating a plurality of pictorial images based on artificial intelligence of claim 1, wherein prior to the step of generating a first pictorial image using the pictorial elements based on a first artificial intelligence model, further comprising:
extracting audio features from the voice painting instruction, and determining a user type according to the audio features;
inquiring a preset drawing element matched with the user type as the drawing element.
5. The method of generating a plurality of pictorial images based on artificial intelligence of any of claims 1-4 wherein the pictorial element corresponding to the first pictorial image is a first pictorial element; the step of generating a second pictorial image using the first pictorial image based on a second artificial intelligence model includes:
acquiring a second painting element according to the first painting image and the association characteristics between the main image and the local image learned by a second artificial intelligence model; the method comprises the steps of,
inquiring a preset drawing element matched with the display size of the slave device to serve as a second drawing element;
A second pictorial image is generated based on the second pictorial element.
6. The method of generating multiple pictorial images based on artificial intelligence of claim 5 wherein the number of slave devices is multiple; the step of generating a second pictorial image based on the second pictorial element includes:
and generating a second painting image corresponding to each slave device according to the second painting element matched with the display size of each slave device and the second painting element acquired based on the first painting image.
7. The method of generating a plurality of pictorial images based on artificial intelligence of claim 1, wherein prior to the step of generating a second pictorial image using the first pictorial image based on a second artificial intelligence model, further comprising:
acquiring a plurality of painting images, and marking a main body image and a local image in each painting image by using a third artificial intelligent model;
generating a training sample according to the main body image and the local image corresponding to each drawing image;
the second artificial intelligence model is trained based on a plurality of the training samples.
8. The method of generating a plurality of pictorial images based on artificial intelligence of claim 1, further comprising:
Taking a display device which receives the voice painting instruction as the master device and taking other display devices except the master device as the slave devices;
or, the display device with the largest display size among the plurality of display devices is used as the master device, and other display devices except the master device are used as the slave devices.
9. An electronic device comprising a memory, a processor, and an artificial intelligence based generation program stored on the memory and executable on the processor, the artificial intelligence based generation program implementing the steps of the artificial intelligence based generation method of multiple pictorial images as claimed in any of claims 1 to 8 when executed by the processor.
10. A computer-readable storage medium, wherein a program for generating a plurality of drawing images based on artificial intelligence is stored on the computer-readable storage medium, and the program for generating a plurality of drawing images based on artificial intelligence realizes the steps of the method for generating a plurality of drawing images based on artificial intelligence according to any one of claims 1 to 8 when executed by a processor.
CN202310148885.9A 2023-02-22 2023-02-22 Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence Active CN116342739B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310148885.9A CN116342739B (en) 2023-02-22 2023-02-22 Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310148885.9A CN116342739B (en) 2023-02-22 2023-02-22 Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence

Publications (2)

Publication Number Publication Date
CN116342739A CN116342739A (en) 2023-06-27
CN116342739B true CN116342739B (en) 2023-09-26

Family

ID=86892140

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310148885.9A Active CN116342739B (en) 2023-02-22 2023-02-22 Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence

Country Status (1)

Country Link
CN (1) CN116342739B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111028828A (en) * 2019-12-20 2020-04-17 京东方科技集团股份有限公司 Voice interaction method based on screen drawing, screen drawing and storage medium
CN111125403A (en) * 2019-11-27 2020-05-08 浙江大学 Aided design drawing method and system based on artificial intelligence
CN113793398A (en) * 2020-07-24 2021-12-14 北京京东尚科信息技术有限公司 Drawing method and device based on voice interaction, storage medium and electronic equipment

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109242030A (en) * 2018-09-21 2019-01-18 京东方科技集团股份有限公司 Draw single generation method and device, electronic equipment, computer readable storage medium
US11386589B2 (en) * 2020-08-04 2022-07-12 Ping An Technology (Shenzhen) Co., Ltd. Method and device for image generation and colorization

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111125403A (en) * 2019-11-27 2020-05-08 浙江大学 Aided design drawing method and system based on artificial intelligence
CN111028828A (en) * 2019-12-20 2020-04-17 京东方科技集团股份有限公司 Voice interaction method based on screen drawing, screen drawing and storage medium
CN113793398A (en) * 2020-07-24 2021-12-14 北京京东尚科信息技术有限公司 Drawing method and device based on voice interaction, storage medium and electronic equipment

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于卷积神经网络的水墨画合成方法;彭玉元;赵朋朋;武有能;谢兵兵;;桂林理工大学学报(第01期);204-209 *

Also Published As

Publication number Publication date
CN116342739A (en) 2023-06-27

Similar Documents

Publication Publication Date Title
CN109410927B (en) Voice recognition method, device and system combining offline command word and cloud analysis
KR101992424B1 (en) Apparatus for making artificial intelligence character for augmented reality and service system using the same
CN115830171B (en) Image generation method based on artificial intelligence drawing, display equipment and storage medium
US20160283465A1 (en) Method for dialogue between a machine, such as a humanoid robot, and a human interlocutor; computer program product; and humanoid robot for implementing such a method
CN102110304A (en) Material-engine-based automatic cartoon generating method
TW200850057A (en) Method, system and user interface for automatically creating an atmosphere, particularly a lighting atmosphere, based on a keyword input
CN113094801B (en) Decoration simulation image generation method, device, equipment and medium
KR102334705B1 (en) Apparatus and method for drawing webtoon
CN116704085B (en) Avatar generation method, apparatus, electronic device, and storage medium
CN111324207A (en) Drawing display method and device and electronic equipment
CN104915458B (en) A kind of method, system and mobile terminal associated automatically when user searches for and applies
CN116824020A (en) Image generation method and device, apparatus, medium, and program
CN116342739B (en) Method, electronic equipment and medium for generating multiple painting images based on artificial intelligence
CN118052907A (en) Text map generation method and related device
KR20100020302A (en) Image analysis apparatus and method
CN113056066B (en) Light adjusting method, device, system and storage medium based on television program
CN116402049B (en) Method and device for generating decorated text set and image enhancer and electronic equipment
Pivato Representation of ethnicity as problem: Essence or construction
CN110990632A (en) Video processing method and device
CN116797725A (en) Vehicle-mounted scene generation method, device and system
CN115442495A (en) AI studio system
CN116583820A (en) Voice interaction method and device
CN107066567B (en) Topic detection-based user portrait modeling method and system in text conversation
CN110908525A (en) Input method, client side thereof and method for providing candidate pictures/videos
CN117271809B (en) Virtual agent communication environment generation method based on task scene and context awareness

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