Design draft provision service system

Abstract

The present invention relates to a design provision service system comprising: one or more user terminals for inputting one or more required keywords for producing an image constituting a desired design draft; and a design draft provision server for extracting one or more pieces of design draft information matching the one or more required keywords inputted from the user terminals, and transmitting same to the one or more user terminals. Accordingly, the effects are achieved whereby a design draft image and a related keyword, which match the required keywords, may be simultaneously provided, and a design requester and a design service provider may share and discuss the design draft, and thus the gap between the image desired by the design requester and the result may be minimized.

Description

Design draft provision service system

[0001] The present invention relates to a design draft provision service system, and more particularly, to a design draft provision service system that simultaneously provides a design draft image matching an essential keyword and related keywords, and shares the design draft with a design client and a design service provider, thereby minimizing the gap between the image desired by the design client and the resulting product.

[0002] For businesses and small business owners, design is essential for marketing, including logos, branding, posters, and program books.

[0003] Most design firms proceed with projects by setting a number of design drafts and revisions to avoid excessive schedules.

[0004] And the client is left with limited information and fears about whether they can be certain of the results.

[0005] If the image the client wants is different from the image the designer has in mind, it is inevitable that a difficult situation will arise where it is necessary to decide whether to get a refund or re-order.

[0006] For example, if you entrust a logo design to a design company, but you don't like it, you end up paying a lot of money again, or you receive the results from another interior designer through a friend's recommendation. There can be a lot of gaps between the image the client wants and the image the designer has in mind.

[0007] In order to solve such problems, the present invention provides a design draft service system that simultaneously provides a design draft image matching an essential keyword and related keywords, and shares the design draft with a design client and a design service provider, thereby minimizing the gap between the image desired by the design client and the result.

[0008] In order to achieve the above purpose, a design draft provision service system according to the features of the present invention is provided.

[0009] One or more user terminals entering one or more required keywords to create images that constitute a desired design draft;

[0010] A design draft providing server that extracts one or more design draft information matching the required keywords received from each of the one or more user terminals and transmits the extracted information to the one or more user terminals,

[0011] Each of the one or more user terminals receives a design selection signal of a design draft image for re-selecting a design draft image from one or more design draft information matching an adjective keyword received from the design draft providing server, and transmits the received design selection signal to the design draft providing server.

[0012] The above design draft providing server performs design work using the reselected design draft image corresponding to the reselected design selection signal.

[0013] The one or more design draft information may include one design draft image and associated keywords related to the one design draft image.

[0014] The above required keywords may include at least one of a noun keyword and an adjective keyword.

[0015] If the above required keywords include the above adjective keywords, the design draft providing server can search for design draft information matching the above adjective keywords and output the design draft image and detailed keywords together.

[0016] By the aforementioned configuration, the present invention has the effect of minimizing the gap between the image desired by the design client and the result by sharing and discussing the design draft with the design client and the design service provider.

[0017] Figure 1 is a drawing showing the configuration of a design draft provision service system according to an embodiment of the present invention.

[0018] Figure 2 is a block diagram briefly showing the internal configuration of a user terminal according to an embodiment of the present invention.

[0019] FIG. 3 is a block diagram briefly showing the internal configuration of a design draft providing server according to an embodiment of the present invention.

[0020] FIG. 4 and FIG. 5 are diagrams showing examples of essential keywords and design draft information (design draft image and associated keywords) according to an embodiment of the present invention.

[0021] FIG. 6 is a drawing showing an example of a home page screen of a user terminal provided by a design draft providing server according to an embodiment of the present invention.

[0022] The present invention is susceptible to various modifications and embodiments. Specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, but rather to encompass all modifications, equivalents, and alternatives falling within the spirit and technical scope of the present invention. Throughout the description of each drawing, similar reference numerals have been used to designate similar components.

[0023] Terms such as "first," "second," "A," and "B" may be used to describe various components, but these components should not be limited by these terms. These terms are used solely to distinguish one component from another. For example, without departing from the scope of the present invention, the first component could be referred to as the "second component," and similarly, the second component could also be referred to as the "first component." The term "and / or" includes any combination of multiple related items listed or any one of multiple related items listed.

[0024] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0025] The terminology used in this application is only used to describe specific embodiments and is not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, it should be understood that the terms "comprise" or "have" indicate the presence of a feature, number, step, operation, component, part, or combination thereof described in the specification, but do not exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0026] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and will not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0027] Hereinafter, with reference to the attached drawings, preferred embodiments of the present invention will be described in more detail. In order to facilitate an overall understanding in describing the present invention, identical reference numerals will be used for identical components in the drawings, and redundant descriptions of identical components will be omitted.

[0028] Figure 1 is a drawing showing the configuration of a design draft provision service system according to an embodiment of the present invention.

[0029] A design draft provision service system (100) according to an embodiment of the present invention may include a plurality of user terminals (110), a design draft provision server (120), and an external server (130).

[0030] A design draft provision service system (100) includes a user terminal (110) that runs a design draft recommendation service app to use a design draft provision program, a design draft provision server (120) that is connected to the user terminal (110) via a communication network (102), provides a service for a design draft provision program function through the design draft recommendation service app, manages statistical data on user app usage history related to design draft recommendation, and helps provide a final design draft image desired by a design client by sharing a design draft image with the user terminal (110) in advance.

[0031] The design draft provision server (120) serves as a platform that provides a design draft provision service to multiple user terminals (110) that are clients. The design draft provision server (120) can provide the design draft provision service by linking with an application installed on each user terminal (110).

[0032] The design draft provision service system (100) can be constructed to be linked to the design draft provision server (120) through a communication network (102) to exchange design draft related information for the design draft recommendation service.

[0033] The design draft recommendation service app is a management app that is installed on a user terminal (110) to receive a design draft provision program provided by a design draft provision server (120) or is provided as a web-based service.

[0034] The communication network (102) is not limited in its communication method, and may include, for example, a communication method utilizing a mobile communication network, wired Internet, wireless Internet, or broadcasting network, as well as short-range wireless communication between devices. For example, the communication network (102) may include any one or more of a network such as a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a broadband network (BBN), or the Internet. In addition, the communication network (102) may include any one or more of a network topology including, but not limited to, a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree, or a hierarchical network.

[0035] The user terminal (110) may connect to a design draft provision server (120) through a design draft recommendation service app and include an input unit for inputting one or more essential keywords for creating an image that constitutes a desired design draft.

[0036] The design draft provision server (120) can extract one or more design draft information matching the required keywords received from each user terminal (110) and transmit the information to each of one or more user terminals (110).

[0037] The external server (130) may be a server that stores design draft images in a database, or a server that transmits design draft images by connecting to the Internet.

[0038] Figure 2 is a block diagram briefly showing the internal configuration of a user terminal according to an embodiment of the present invention.

[0039] A user terminal (110) according to an embodiment of the present invention may include a communication module (111), a terminal control unit (112), a display unit (113), a memory unit (114), a storage unit (115), and an input unit (116).

[0040] The terminal control unit (112) can upload and activate the design draft recommendation service app downloaded from the design draft provision server (120) to the memory unit (114).

[0041] The terminal control unit (112) can display a graphical user interface on the display unit (113) through a design draft recommendation service app downloaded from the design draft provision server (120) through the communication module (111).

[0042] The terminal control unit (112) can operate a design draft recommendation service app to connect to a design draft providing server (120) and search for a design draft image through a homepage provided by the design draft providing server (120).

[0043] The input unit (116) represents an input means such as a keyboard, mouse, or touch pad, and receives one or more required keywords for creating an image that constitutes a desired design draft.

[0044] The terminal control unit (112) can transmit the required keyword inputted through the input unit (116) to the design draft provision server (120) through the communication module (111).

[0045] Here, the required keywords may include at least one of a noun keyword and an adjective keyword.

[0046] The terminal control unit (112) receives design draft information from the design draft providing server (120) and outputs it to the display unit (113). Here, the design draft information may include one design draft image and related keywords (30) related to one design draft image.

[0047] The terminal control unit (112) can store a plurality of design draft information received from the design draft providing server (120) in the storage unit (115), receive a design selection signal of a design draft image included in one or more pieces of design draft information through the input unit (116), and transmit the received design selection signal to the design draft providing server (120).

[0048] If the design draft provision server (120) includes an adjective keyword in the required keywords, it can search for design draft information matching the adjective keyword and output the design draft image and detailed keywords, which are objective and descriptive keywords, together.

[0049] FIG. 3 is a block diagram briefly showing the internal configuration of a design draft providing server according to an embodiment of the present invention, FIGS. 4 and 5 are diagrams showing examples of essential keywords and design draft information (design draft image and related keywords) according to an embodiment of the present invention, and FIG. 6 is a diagram showing an example of a home page screen of a user terminal provided by a design draft providing server according to an embodiment of the present invention.

[0050] A design draft providing server (120) according to an embodiment of the present invention may include a design transmission / reception unit (121), an object image extraction unit (122), an object information database unit (123), a design information database unit (124), a control unit (125), a morphological analyzer (126), a thesaurus unit (126a), an adjective dictionary unit (126b), a user interface unit (127), a learning unit (128a), and an artificial neural network engine unit (128) of an artificial neural processing network (128b).

[0051] The design transmitter / receiver (121) can receive images from a user terminal (110) or an external server (130).

[0052] The object image extraction unit (122) can divide an image into macro blocks each consisting of N×N pixels in size, and convert the divided macro blocks into a black and white image that displays a boundary line.

[0053] The object image extraction unit (122) can convert each image into an edge image using an edge detection algorithm for multiple macro blocks.

[0054] The edge detection algorithm refers to a method of determining an edge only in a section where the pixel values ​​of the boundary are found to differ significantly from a preset reference pixel value. In another embodiment, the edge detection algorithm can detect an edge by utilizing the rate of change in brightness values ​​between pixels using the Sobel operation method.

[0055] Edge detection algorithms can be used in any edge image that can extract the boundaries of people in the image.

[0056] The object image extraction unit (122) extracts a feature map of a thumbnail image frame using one of a deep neural network (DNN), a convolutional deep neural network (CNN), a recurrent neural network (RNN), and a deep belief neural network (DBN).

[0057] The object image extraction unit (122) can generate a feature map using a model that has already been trained by the learning unit based on deep learning. Deep learning is defined as a set of machine learning algorithms that attempt a high level of abstraction (summarizing key content or functions in a large amount of data or complex data) through a combination of various nonlinear transformation methods.

[0058] The object image extraction unit (122) extracts an area in which an object is estimated to exist in the image, and extracts a feature map representing features from the extracted area.

[0059] Here, the object is a concept that includes people, animals, and inanimate objects (buildings, bicycles, etc.), but in the present invention, it represents a human object.

[0060] The object image extraction unit (122) extracts at least one region in which the presence of an object is estimated from the image based on the extracted feature map. Methods for extracting the region may include, for example, faster RCNN, SSD (Single Shot Multibox Detector), and YOLO (You Only Look Once).

[0061] The object image extraction unit (122) includes a machine learning engine (e.g., TensorFlow, etc.) and can apply an object detection algorithm based on a deep learning model pre-trained with a learning image set.

[0062] When an image received from a user terminal (110) or an external server (130) is input, the object image extraction unit (122) can find multiple objects included in the image based on a deep learning model-based object detection algorithm and extract the characteristics of the objects.

[0063] The object image extraction unit (122) can extract features of an object detected in an image by applying a learning model modeled based on various attributes (e.g., color, texture, category, etc.) that can classify objects. In addition to the above-described method, features included in an image can be extracted using at least one algorithm among image recognition algorithms such as SIFT (scale invariant feature transform), SURF (speeded up robust features), HOG (histogram of oriented gradient), and MCT (modified census transform), which are known technologies.

[0064] The object image extraction unit (122) can extract a first feature of an object image and a first vector value corresponding thereto using a deep learning model that segments an object from an image.

[0065] The object information database section (123) registers in advance the features and vector values ​​of multiple object images and keywords related thereto and stores them respectively.

[0066] The object image extraction unit (122) can transmit a first vector value corresponding to a first feature of the extracted object image to the object information database unit (123).

[0067] The object information database unit (123) can search for an object image similar to the input object image by comparing a first vector value corresponding to a first feature of the input object image with a second vector value corresponding to a second feature of a plurality of pre-stored object images.

[0068] When searching for an object image similar to an input object image, the object information database unit (123) recognizes the input object image to obtain an object name and transmits keywords related to the object name to the object image extraction unit (122). Here, the keywords may be text information.

[0069] The object image extraction unit (122) can receive object images searched from the object information database unit (123) and keywords related to the object images and store them in the design information database unit (124).

[0070] The object image extraction unit (122) extracts each object image from one design draft image (20) input from the design transmission / reception unit (121), transmits each extracted object image to the object information database unit (123), recognizes each object image from the object information database unit (123), obtains the name of the object, receives a keyword related to the object image, and stores each object image included in one received image and the associated keyword related to each object image in the design information database unit (124).

[0071] The design information database unit (124) can create and store design draft information by matching a single design draft image (20) input from the design transmission / reception unit (121), multiple object images included in the design draft image (20), and associated keywords (30) related to each object image.

[0072] The design information database (124) can classify and store design draft images (20) by related keywords (30).

[0073] A plurality of user terminals (110) input one or more essential keywords (10) for creating an image constituting a desired design draft, and transmit the input essential keywords (10) to a design draft providing server (120). The essential keywords (10) can be configured through consultation between the service user and the service provider.

[0074] The design transmission / reception unit (121) of the design draft provision server (120) can receive essential keywords (10) from each user terminal (110).

[0075] The control unit (125) can transmit the required keyword (10) received from each user terminal (110) to the design information database unit (124) and receive one or more design draft information including related keywords matching the required keyword (10) from the design information database unit (124).

[0076] When there are multiple required keywords (10), the design information database (124) can search for pre-registered design draft information that includes all keywords identical to each required keyword (10) and transmit the searched information to the control unit (125).

[0077] When there are multiple required keywords (10), the design information database unit (124) can search for pre-registered design draft information that includes all similar keywords similar to each required keyword (10) and transmit the searched information to the control unit (125).

[0078] The design information database (124) may further include a morphological analyzer (126) that analyzes the part of speech of a word to determine whether the keyword matching the design draft image (20) is a noun or adjective.

[0079] In the case where there are multiple required keywords (10), the design information database (124) can analyze the part of speech of each required keyword (10) in a morphological analyzer (126) and classify the required keywords (10) into noun keywords that are nouns and adjective keywords that are adjectives.

[0080] The design information database section (124) may include synonyms that are different words with the same meaning registered in the Korean dictionary, such as egg, egg, laundry, and washing, and may include words by type (apple, watermelon, etc.) included in one upper category (fruit, etc.), and may include a synonym dictionary section (126a) that includes all words that have several to dozens of meanings defined for one word, such as “beautiful,” “beauty,” and “beautiful.”

[0081] The design information database section (124) may include an adjective dictionary section (126b) that classifies objective adjectives used to express facts and subjective adjectives used to express people's opinions. For example, objective adjectives refer to facts such as big, small, new, colorful, and indisputable, while subjective adjectives refer to things that require a person's subjective opinion, such as expensive, delicious, and luxurious.

[0082] The design information database section (124) can analyze the part of speech of each essential keyword (10) through a morphological analyzer (126) and classify them into noun keywords and adjective keywords using a thesaurus section (126a) and an adjective dictionary section (126b).

[0083] The design information database unit (124) searches for registered design draft information including keywords that match the noun keyword when there is one noun keyword according to the number of noun keywords, and extracts design draft information that matches the keyword when there is a similar keyword that matches the noun keyword or is similar to the noun keyword. For example, when the noun keyword is fruit, the design information database unit (124) can set not only keywords that are identical to fruit but also words classified into the fruit category, such as apples and watermelons, as similar keywords.

[0084] The design information database (124) searches for registered design draft information including multiple keywords that match each noun keyword when there are multiple noun keywords, and, when there are similar keywords that match multiple noun keywords or are similar to multiple noun keywords, it can extract design draft information that matches the multiple keywords.

[0085] The design information database unit (124) can transmit the extracted design draft information to the control unit (125).

[0086] The design information database section (124) determines whether the adjective keyword is an objective adjective or a subjective adjective using the adjective dictionary section (126b) when several required keywords (10) include noun keywords and adjective keywords, and if it is an objective adjective, the objective adjective is regarded as an objective and descriptive detailed keyword of the noun keyword, and can extract a plurality of design draft information (one design draft image (20) and a plurality of related keywords (30) related to one design draft image (20)) based on the noun keyword.

[0087] The design information database (124) may further include a high-quality style index indicating luxury in the design draft information.

[0088] A detailed explanation of the high-quality style index is provided below.

[0089] As shown in Fig. 4, for example, when the required keyword (10) is fruit, menu board, or fresh, the design information database (124) searches for and displays a design draft image (20) that includes a matching related keyword (30) of the required keyword (10).

[0090] As illustrated in FIG. 5, for example, if the required keyword (10) is a horror movie or a volcano, the design information database (124) can search for and display a design draft image (20) that includes a matching related keyword (30) of the required keyword (10).

[0091] The design information database section (124) can, when several required keywords (10) include noun keywords and adjective keywords, use the adjective dictionary section (126b) to determine whether the adjective keywords are objective adjectives and subjective adjectives, and, if they are subjective adjectives, generate question data querying the meaning of the subjective adjective, and extract multiple pieces of design draft information (one design draft image (20) and a plurality of associated keywords (30) related to one design draft image (20)) having associated keywords (30) that match the noun keywords.

[0092] The subjective adjective of the present invention may be, but is not limited to, “luxurious” and may also be other terms such as “delicious”.

[0093] The control unit (125) can generate design proposal information including the generated question data and multiple extracted design draft information and transmit it to the user terminal (110) via the design transmission / reception unit (121).

[0094] The design information database (124) can generate question data querying the meaning of luxurious when the subjective adjective is “luxurious.”

[0095] Here, the question data is a query that asks for words or reasons that evoke luxury in the selected design draft image (20) when a design draft image (20) of a luxurious image is selected from among multiple design draft images (20) on a user terminal (110).

[0096] For example, if the question data assumes that the adjective keyword is "luxurious," the user may be asked to input the reason why the user selected a design draft image (20) that matches the adjective keyword (luxurious) from among multiple design draft images (20) or a word associated with the selected design draft image.

[0097] The user terminal (110) receives a design selection signal for re-selecting a design draft image from design draft information matching the adjective keyword (luxurious) received from the design draft providing server (120), and receives question answer information in response to a question data input request signal from the design draft information.

[0098] Design selection signals represent images that are re-selected from multiple design draft images that match the adjective keyword (luxurious).

[0099] The question and answer information indicates the reason or word that evokes luxury in the selected image.

[0100] The user terminal (110) can generate design sharing information including a design selection signal and question answer information for reselecting a design design image from one or more design design information matching an adjective keyword received from a design design provision server and transmit the generated design sharing information to the design design provision server (120).

[0101] The design draft providing server (120) can perform design work using a reselected design draft image corresponding to a reselected design selection signal.

[0102] When the control unit (125) of the design draft provision server (120) receives design sharing information from a user terminal (110), it analyzes the design selection signal and question answer information included in the received design sharing information, sets the number of times a design selection signal corresponding to each design draft image is received as the number of design preferences, and stores the question answer information as an associated keyword (30) in the design information database unit (124).

[0103] The design information database (124) registers question answer information as an associated keyword (30) indicating a subjective adjective (luxury), and if there is a word in the input question answer information that matches a pre-registered associated keyword (30), the high quality status information of the corresponding design draft image becomes '1', and if there is no word in the input question answer information that matches a pre-registered associated keyword (30), the high quality status information of the corresponding design draft image becomes '0'.

[0104] The design information database section (124) has a design preference score preset for each design draft image in proportion to the number of design preferences, a high quality status score corresponding to high quality status information of '1' is preset, and a high quality status score corresponding to high quality status information of '0' is 0.

[0105] The design information database (124) stores the number of design preferences and high-quality status information for each design draft image, and a high-quality style index can be calculated by the following mathematical expression 1 based on the number of design preferences and high-quality status information.

[0106] The design information database (124) has a high-quality style index matched to each design draft image.

[0107]

[0108] The above design preference score generates a design selection signal indicating the selection of a luxurious design image that the design client thinks is selected from among several design design images received from the design design provision server (120) at the user terminal (110), and the design selection signal is accumulated by setting the number of design preference receptions received from the user terminal (110) at the design design provision server (120), and a preset design preference score is displayed in proportion to the accumulated number of design preference receptions.

[0109] The above high quality status score is generated when the design draft providing server (120) includes the subjective adjective luxurious in the required keywords (10), and when the luxurious design draft image is selected, question data inquiring about the meaning of luxurious is generated and transmitted to the user terminal (110), and the user terminal (110) receives question answer information in response to the question data, and when the design draft providing server (120) has a word in the question answer information that matches a pre-registered related keyword (30), the high quality status information of the corresponding design draft image becomes '1' and indicates a preset high quality status score corresponding to the high quality status information that is '1'.

[0110] W1 represents the first weight related to the design preference score, and W2 represents the second weight related to the high quality condition score, and the first and second weights are preset.

[0111] However, the first and second weights can be set differently depending on whether the emphasis is on selecting design draft images or on matching words and related keywords in the Q&A information. Generally, the first weight can be configured with a higher value than the second weight.

[0112] The design information database (124) calculates a design reselection index by substituting the high-quality style index calculated by mathematical formula 1 into mathematical formula 2.

[0113] The design information database section (124) may have a higher design reselection index for a design draft image, which may increase the likelihood that the design draft image will be used for design work.

[0114] The design information database (124) matches a high-quality style index and a design reselection index to each design draft image.

[0115]

[0116] The control unit (125) selects the top 10 design draft information having a high quality style index based on the high quality style index included in one or more design draft information from among the plurality of extracted design draft information, and generates design proposal information including the generated question data and the selected design draft information, and transmits the generated design proposal information to the user terminal (110) via the design transmission / reception unit (121).

[0117] If the design reselection index is greater than or equal to a preset standard, the control unit (125) can set a design draft image having a design reselection index greater than or equal to the preset standard as a design draft image for which design work will be performed.

[0118] The control unit (125) can set the design draft image with the highest design reselection index as the design draft image for which design work will be performed.

[0119] Each design draft image is matched with a high-quality style index and a design reselection index.

[0120] Accordingly, the control unit (125) can select a design draft image having both the highest values ​​for the high-quality style index and the design reselection index and set it as a design draft image for proceeding with the design work.

[0121] When the control unit (125) receives design sharing information from the user terminal (110), it can implement design draft information corresponding to the design selection signal of the design sharing information as a graphical user interface through the user interface unit (127) and provide it to the user homepage screen (40) of the app of the user terminal (110).

[0122] The user interface unit (127) may provide a home page screen (40) provided by a design draft providing server (120) to a user terminal (110) and provide a graphic interface that helps the user terminal (110) search for a design draft image.

[0123] The artificial neural network engine unit (128) according to an embodiment of the present invention may further include a learning unit (128a) that trains the artificial neural processing network (128b).

[0124] The artificial neural network engine unit (128) can input a required keyword (10) as input and output a design draft image (20) and related keywords (30) that match the required keyword (10) as output through a design draft information recommendation artificial intelligence algorithm.

[0125] The learning unit (128a) can learn through an artificial neural processing network (128b) using learning data stored in the design information database unit (124).

[0126] The learning unit (128a) can detect feature points from learning data, and can adjust the connection weights applied to the artificial neural processing network (128b) based on the learning results of the detected feature points.

[0127] The learning unit (128a) can calculate an error by comparing the output value generated from the output layer of the neural network with the desired expected value for the learning data, and adjust the connection weights applied to the recognition model of the neural network in a direction to reduce the error.

[0128] The learning unit (128a) is composed of an input layer, a hidden layer, and an output layer, and can train an artificial neural processing network (128b) to output one or more design draft information to the output layer through an artificial intelligence algorithm for recommending design draft information generated according to essential keywords (10) representing a design draft image.

[0129] The learning unit (128a) can detect feature vectors from essential keywords (10) and design draft information, and train an artificial neural processing network (128b) using the detected feature vectors.

[0130] The learning unit (128a) trains the artificial neural processing network (128b), and when the weights of the artificial neural processing network (128b) are adjusted and learning is completed, the artificial neural processing network (128b) can become the artificial neural network engine unit (128).

[0131] The learning unit (128a) performs learning of an artificial neural processing network (128b) using a training set consisting of essential keywords (10), and learns one or more design draft information recommended according to the deep learning feature values ​​and essential keywords (10) by type.

[0132] The learning unit (128a) learns one or more design draft information recommended according to the required keywords (10) from an artificial neural processing network (128b).

[0133] The learning unit (128a) receives required keywords (10) and generates a deep learning-based recognition model (artificial intelligence algorithm for recommending children's customized nutritional supplements) through optimization based on the prediction result of one or more design draft information recommended according to the required keywords (10).

[0134] The learning unit (128a) uses a training set consisting of essential keywords (10) as an input vector, and when it passes through the input layer, hidden layer, and output layer, it learns through supervised learning to generate one or more design draft information as an output vector according to the essential keywords (10).

[0135] The learning unit (128a) inputs learning data stored in the design information database unit (124) into the artificial neural processing network (128b) to output correct data, that is, adjusts the connection weights of the artificial neural processing network (128b) so that the loss function becomes the minimum value, and the resultant artificial neural processing network (128b) that has completed learning is created through this.

[0136] When learning data, test data, or new data not used for learning is input to an artificial neural processing network (128b) with updated connection weights, the learning unit (128a) can obtain the output result through the input layer - hidden layer - polling connected layer - output layer and output it as response data.

[0137] An artificial neural processing network (128b) that has completed learning can predict one or more design draft images corresponding to the input essential keywords (10) using a deep learning-based recognition model (design draft recommendation artificial intelligence algorithm).

[0138] The artificial neural processing network (128b) quantizes the essential keyword (10) and the design draft image to extract feature points, compares the extracted feature points with the feature vectors of the learning data learned in the learning unit (128a), and transmits the design draft image to the control unit (125) according to the essential keyword (10).

[0139] The artificial neural processing network (128b) outputs the required keyword (10), which is test data, as a result value of the response data (result of the design draft image) using a deep learning-based recognition model.

[0140] The operations according to the embodiments of this specification can be implemented as computer-readable programs or codes on a computer-readable recording medium. Computer-readable recording media include all types of recording devices that store data that can be read by a computer system. Furthermore, computer-readable recording media can be distributed across network-connected computer systems, allowing computer-readable programs or codes to be stored and executed in a distributed manner.

[0141] When the embodiment is implemented in software, the above-described technique can be implemented as a module (process, function, etc.) that performs the above-described function. The module can be stored in memory and executed by the processor. The memory can be internal or external to the processor and connected to the processor by various well-known means.

[0142] Additionally, the computer-readable recording medium may include hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc. The program instructions may include not only machine language codes produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc.

[0143] While some aspects of the present invention have been described in the context of a device, they may also represent a description of a corresponding method, wherein a block or device corresponds to a method step or a feature of a method step. Similarly, aspects described in the context of a method may also be described as a corresponding block or item or a feature of a corresponding device. Some or all of the method steps may be performed by (or using) a hardware device, such as, for example, a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, one or more of the most significant method steps may be performed by such a device.

[0144] In embodiments, a programmable logic device (e.g., a field programmable gate array) may be used to perform some or all of the functions of the methods described herein. In embodiments, the field programmable gate array may operate in conjunction with a microprocessor to perform one of the methods described herein. In general, the methods are preferably performed by some hardware device.

[0145] Although the present invention has been described above with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various modifications and changes may be made to the present invention without departing from the spirit and scope of the present invention as set forth in the claims below.

Claims

1. One or more user terminals that input one or more required keywords to create images that constitute the desired design draft; A design draft providing server that extracts one or more design draft information matching the required keywords received from each of the one or more user terminals and transmits the extracted information to the one or more user terminals, Each of the one or more user terminals receives a design selection signal of a design design image for re-selecting a design design image from one or more design design information matching an adjective keyword received from the design design provision server, and transmits the received design selection signal to the design design provision server. The above design draft providing server performs design work using the reselected design draft image corresponding to the reselected design selection signal. A design draft provision service system in which the one or more design draft information includes one design draft image and associated keywords related to the one design draft image.

2. In claim 1, The above required keywords are a design draft provision service system that includes at least one of a noun keyword and an adjective keyword.

3. In claim 2, The above design draft provision server is a design draft provision service system that searches for design draft information matching the above adjective keyword when the above required keyword includes the above adjective keyword and outputs the design draft image and detailed keywords together.

Images