KR20220000693A - Service providing apparatus and method for reviewing advertisement, and non-transitory computer readable medium having computer program recorded thereon - Google Patents

Abstract

The present invention relates to an apparatus for providing an advertisement examination service, a method thereof, and a non-transitory computer-readable medium having a computer program recorded thereon and, more specifically, to an apparatus for providing advertisement examination service, which uses a learning model learning examination results of judges who judge advertisement content created by advertisers to increase screening accuracy in the case of automatic screening of advertisement contents, a method thereof, and a non-transitory computer-readable medium having a computer program recorded thereon. According to the present invention, the method simultaneously examines whether each individual element forming advertisement content has passed examination and whether the entire advertisement content has passed the examination, thereby preventing examination elements from being omitted during an examination process. In addition, only when consensus is reached through cross-checking between examination results of examining individual elements forming advertisement content while being separately from advertisement content and the examination results of examining individual elements as a whole of advertisement content, the advertisement content is judged to have passed the examination, thereby providing an effect of greatly increasing examination accuracy. The method comprises a reception step, an extraction step, an examination step, and a judgement step.

Description

Service providing apparatus and method for reviewing advertisements, and a non-volatile recording medium in which a computer program is recorded, and a non-transitory computer readable medium having computer program recorded thereon

The present invention relates to a service providing apparatus and method for advertisement screening, and to a non-volatile recording medium in which a computer program is recorded, and more particularly, using a learning model in which the evaluation result of a judge who judges advertisement content generated by an advertiser is learned. The present invention relates to an apparatus and method for providing a service for advertisement screening, which can increase screening accuracy when automatically examining advertisement contents, and to a non-volatile recording medium in which a computer program is recorded.

Currently, advertisements for the promotion of products or services are becoming an indispensable means for increasing sales and profits, and with the recent development of communication networks, these advertisements are available in a variety of media through the Internet network than in the past media such as broadcasting media or electronic billboards. It is more and more preferred to be provided through a medium.

Accordingly, advertisers produce advertisement contents to provide advertisements through the Internet network and then provide them to the operating servers of various web sites or applications so that they are displayed on various advertisement publication targets such as the web sites or applications.

At this time, various problems may occur, such as when the advertisement is undermined by public morals or the operating principle of the advertisement target is violated.

However, in the past advertising media, including physical media such as broadcast media or electronic billboards, the number of advertisements that can be accommodated is limited, so there is no difficulty in manually examining whether advertisements are unsuitable for advertisement publication through a human reviewer. , when advertisements are published as advertisement contents through a website or application, various advertisement contents can be changed at any time in a predetermined time unit. Not only is it difficult to manually review the advertisement content, but there is also a significant problem in the time and cost required for reviewing advertisement content.

Accordingly, there is a demand for a new method for improving the screening efficiency of such advertisement contents.

Korean Patent Laid-Open Patent No. 10-2001-0000949 [Title of the invention: Advertising method using media]

The present invention learns the evaluation results of the judges on the advertisement contents provided by the advertiser and applies the advertisement contents to the learning model in which the evaluation criteria are learned to support automatic evaluation of advertisement contents, thereby increasing the convenience of screening and at the same time increasing the advertisement contents. After extracting the image object and text from the image object and text, the purpose of this is to increase the accuracy of the examination result by calculating the examination result for each of the image object and the text through the learning model.

In addition, in the present invention, if there is a difference between the evaluation result judged by the judge for advertisement content and the judgment result calculated through the learning model, the judgment accuracy is improved by re-learning the learning model based on the reviewer's corrections to improve the judgment accuracy. It aims to increase

A service providing method for advertisement review by a service providing apparatus according to an embodiment of the present invention includes the receiving step of receiving advertisement content, and the extraction step of extracting at least one of an image object and text that are not identified as text from the advertisement content And, when the image object is extracted from the advertisement content among the first learning model in which the first examination condition is set by learning the examination result for the image object and the second learning model in which the second examination condition is set by learning the examination result for the text A first evaluation result is obtained by applying to the first learning model, and when text is extracted from the advertisement content, a second evaluation result is obtained by applying to the second learning model, and both the image object and the text from the advertisement contents A screening step of acquiring a third screening result by applying the advertisement contents to a third learning model in which a third screening condition is set by learning the screening results for advertisement contents in which an image object and text are identified together when extracted, and the advertisement contents If there is no third examination result in the examination result obtained through the examination step corresponding to the above, it is determined whether or not the advertisement content is passed based on any one of the first and second examination results, and through the examination step When at least one of the first to third examination results is unsuccessful when a third examination result is obtained, the method may include a determining step of determining a final examination result of the advertisement content as a failure.

As an example related to the present invention, in the method for providing a service for advertisement screening, after the determining step, when at least one of the first and second screening results is a failure and the third screening result is a pass, the third learning model The method may further include the step of updating the third learning model by learning an examination result corresponding to a failure among the first and second examination results.

As an example related to the present invention, in the method for providing a service for advertisement screening, after the determining step, screening review request information including a final screening result for the advertisement content is generated and transmitted to a preset judge's terminal; When correction information for the final examination result or error information is received from the examiner's terminal in response to the examination review request information, the terminal corresponds to the correction information among the first to third learning models based on the correction information It may be characterized in that it further comprises a teaching step of learning and updating the correction information in a learning model to be used.

As an example related to the present invention, in the method of providing a service for advertisement screening, when at least one of the first and second screening results is a pass while the third screening result is unsuccessful after the determining step, the first and the first 2 Selecting from the third examination results the examination contents for the same examination items as the examination items of the passed specific examination result among examination results, and generating error information if the specific examination results and the examination contents do not match each other; Further comprising, in the teaching step, when generating the error information, generating examination review request information including the error information instead of the final examination result, transmitting the information to the examiner's terminal, and receiving correction information corresponding to the error information It may be characterized by learning and updating the learning model based on the learning model.

As an example related to the present invention, the first examination condition is a color standard for the text, a font standard, a font size standard, a standard for one or more examination items related to a violation of public morals, a trade name, a product name, and a service name in the text. Each of the criteria for identification of any one advertisement target name and at least one of the advertisement target name identified from the text is included as a screening item, and the first screening result includes screening for each screening item included in the first screening condition It may be characterized in that it includes the reason and whether the examination passed, and whether or not the pass was passed according to whether the examination was passed for each examination item.

As an example related to the present invention, the second examination condition includes one or more preset criteria for each property of an object including at least one of size, resolution, and color of the image object, and one or more criteria for each examination item related to violation of public morals , at least one of a criterion for identification of an advertisement target name of any one of a brand name, a product name, and a service name from the image object and at least one of the advertisement target name identified from the image object, respectively, as an examination item, and the second examination result includes: It may be characterized in that the second examination condition includes the reason for the examination for each examination item, whether the examination has passed, and whether or not the examination has been passed according to whether the examination has been passed for each examination item.

As an example related to the present invention, the third screening condition includes a size range for each of the text and image objects identified in the advertisement content and the size of the advertisement content according to the type and size of the target on which the advertisement content is posted. Criteria for determining the properties of advertisement content to be used, screening criteria for each preset text-related property determined according to the properties of the advertisement content, screening criteria for each preset image object related property determined according to the properties of the advertisement content, and the advertisement At least one of a criterion for the arrangement pattern of an image object and text identified in the content and a criterion for whether an advertisement target name derived from the image object is the same as an advertisement target name derived from the text is included as a screening item, respectively, , the third examination result includes the reason for each examination item included in the third examination condition, whether the examination passed, and whether or not the examination passed according to whether the examination passed by the examination item, and the examination item included in the third examination result Some of them may be characterized in that they correspond to the examination items of the first and second examination results.

As an example related to the present invention, the advertisement content may include publication target information including attribute values for each attribute of the publication target including the type and size of the publication target on which the advertisement content is posted. .

As an example related to the present invention, in the method of providing a service for advertisement screening, when the final screening result determined through the determining step is a pass, the advertisement content is published on the target of the advertisement content and the advertisement content is accessed. The method may further include an advertisement processing step of transmitting to a user terminal, and transmitting a final examination result including advertisement content and a reason for examination to an advertiser terminal corresponding to the advertisement content when the final examination result is unsuccessful. .

In the computer-readable nonvolatile recording medium according to an embodiment of the present invention, a computer program for causing a computer to function to perform the above-described service providing method for advertisement review may be recorded.

A service providing apparatus for advertisement review according to an embodiment of the present invention comprises: a communication unit for receiving advertisement content; an extraction unit for extracting at least one of an image object and text that are not identified as text from the advertisement content; When an image object is extracted from the advertisement content among the first learning model in which the first evaluation condition is set by learning the evaluation result for to obtain a first examination result, and when text is extracted from the advertisement content, it is applied to the second learning model to obtain a second examination result, and when both an image object and text are extracted from the advertisement content, an image object and a screening unit for acquiring a third screening result by applying the advertisement contents to a third learning model in which a third screening condition is set by learning the screening results for advertisement content identified together with a text, and the screening unit corresponding to the advertisement contents If there is no third examination result in the examination result obtained through the examination, it is determined whether the advertisement content is passed or not based on any one of the first and second examination results, and the third examination result is obtained through the examination step and a control unit configured to determine a final examination result of the advertisement content as a failure when at least one of the first to third examination results is unsuccessful.

The present invention extracts text and image objects from advertisement content and conducts examination based on the examination criteria according to the examination result learning of the examiner for each text and image object, and the examination result obtained by the examiner for the entire advertisement content It checks whether the examination results obtained by conducting the examination based on the examination criteria according to the learning match each other, and automatically determines whether or not the examination of advertisement content has passed based on the examination result to determine whether each individual element constituting the advertisement content has passed the examination. In addition to examining whether or not all of the advertisement contents have passed the examination, there are no screening factors that are omitted from the screening process, and the results obtained when individual elements constituting advertisement contents are separated from advertisement contents and individual elements as the entire advertisement contents Advertisement content is judged to have passed the examination only when consensus is reached through cross-checking between the examination results, which greatly improves the examination accuracy and the examination convenience and examination efficiency according to the automatic examination.

In addition, in the present invention, when the advertisement content consists only of text or image objects, the evaluation result can be quickly and accurately evaluated by easily evaluating the advertisement contents with the same and similar evaluation criteria as the examiner through the learning model in which the evaluation criteria of the judges as described above are learned. Not only can it be calculated automatically, but also accurate evaluation of the entire advertisement content including both the image object and the text by applying the screening criteria that are considered when the advertisement content is to be judged as the entire advertisement content for the advertisement content from which both the image object and the text are extracted Through cross-checking that compares the screening results for text or image objects extracted from advertisement contents with the screening results for the entire advertisement contents, when the screening results do not match each other, based on a more reasonable screening result Advertisement content is judged by automatically re-learning the learning model with the system or re-learning the learning model based on the correction information received through a request for review of the review result to the reviewer, thereby supporting the supplementation of the review criteria. It has the effect of increasing the accuracy and reliability of

In addition, in the present invention, when an error occurs because the examination result for each text and image object extracted from the advertisement content and the examination result for the advertisement content do not match with each other, error information is generated and provided to the examiner or the final examination result information can be provided to the reviewer, and when a correction of errors or final examination results is requested from the reviewer, the learning model can be re-trained based on the correction request to update the learning model to meet the reviewer's assessment criteria. , it is possible to significantly increase the screening accuracy through re-learning of this learning model.

1 is a configuration environment diagram of an advertisement screening system including a service providing apparatus for advertisement screening according to an embodiment of the present invention;
2 is a detailed configuration diagram of an apparatus for providing a service for advertisement review according to an embodiment of the present invention;
3 is a flowchart of a service providing method of a service providing apparatus for advertisement review according to an embodiment of the present invention;
4 to 6 are diagrams illustrating an operation of an apparatus for providing a service for advertisement review according to an embodiment of the present invention.

It should be noted that the technical terms used in the present invention are only used to describe specific embodiments, and are not intended to limit the present invention. In addition, the technical terms used in the present invention should be interpreted as meanings generally understood by those of ordinary skill in the art to which the present invention belongs, unless otherwise specifically defined in the present invention, and excessively comprehensive It should not be construed in the meaning of a human being or in an excessively reduced meaning. In addition, when the technical term used in the present invention is an incorrect technical term that does not accurately express the spirit of the present invention, it should be understood by being replaced with a technical term that can be correctly understood by those skilled in the art. In addition, the general terms used in the present invention should be interpreted according to the definition in the dictionary or according to the context before and after, and should not be interpreted in an excessively reduced meaning.

Also, the singular expression used in the present invention includes the plural expression unless the context clearly dictates otherwise. In the present invention, terms such as “consisting of” or “comprising” should not be construed as necessarily including all of the various components or various steps described in the invention, and some components or some steps may not be included. It should be construed that it may further include additional components or steps.

Also, terms including ordinal numbers such as first, second, etc. used in the present invention may be used to describe the elements, but the elements should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, a first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the accompanying drawings, but the same or similar components are assigned the same reference numerals regardless of reference numerals, and redundant description thereof will be omitted.

In addition, in the description of the present invention, if it is determined that a detailed description of a related known technology may obscure the gist of the present invention, the detailed description thereof will be omitted. In addition, it should be noted that the accompanying drawings are only for easy understanding of the spirit of the present invention, and should not be construed as limiting the spirit of the present invention by the accompanying drawings.

Hereinafter, detailed embodiments of the present invention will be described with reference to the drawings.

1 is a configuration environment diagram of an advertisement screening system including a service providing apparatus for advertisement screening according to an embodiment of the present invention.

As shown, the advertisement review system according to an embodiment of the present invention may be configured to include a service providing apparatus 100 that communicates with an examiner terminal, an advertiser terminal, and a user terminal through a communication network.

At this time, the advertisement review system may be implemented by more components than the components shown in FIG. 1, and the advertisement review system may be implemented by fewer components than that.

In addition, the communication network described in the present invention may include a wired/wireless communication network, and as an example of such a wireless communication network, a wireless LAN (WLAN), a DLNA (Digital Living Network Alliance), a Wibro (Wireless Broadband: Wibro), WiMAX (World Interoperability for Microwave Access: Wimax), GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), CDMA2000 (Code Division Multi Access 2000), EV-DO (Enhanced Voice-Data Optimized or Enhanced Voice- Data Only), Wideband CDMA (WCDMA), High Speed Downlink Packet Access (HSDPA), High Speed Uplink Packet Access (HSUPA), IEEE 802.16, Long Term Evolution (LTE), Long Term Evolution-A (LTE-A) Advanced), Wireless Mobile Broadband Service (WMBS), 5G mobile communication service, Bluetooth, LoRa (Long Range), RFID (Radio Frequency Identification), Infrared Data Association (IrDA), Ultra Wideband (UWB), ZigBee, Near Field Communication (NFC), Ultra Sound Communication (USC), Visible Light Communication (VLC), Wi-Fi, Wi-Fi Direct may be included. In addition, wired communication networks include wired LAN (Local Area Network), wired WAN (Wide Area Network), Power Line Communication (PLC), USB communication, Ethernet, serial communication, optical/coaxial A cable may be included.

In addition, the service providing apparatus 100 may be configured as a server, and may include various servers such as a web server, a database server, a web application server, and the like, and the service providing apparatus 100 includes a network load balancing mechanism, One or more of various software that allows the service device to operate on the Internet or other networks may be installed, and may be implemented as a computerized system through this. Also, the network may be an http network, and may be a private line, an intranet, or any other network.

In addition, the examiner terminal, the advertiser terminal, and the user terminal include a smart phone, a portable terminal, a mobile terminal, a personal digital assistant (PDA), and a portable multimedia player (PMP). terminal, telematics terminal, navigation terminal, personal computer, notebook computer, slate PC, tablet PC, ultrabook, wearable device, For example, watch-type terminal (Smartwatch), glass-type terminal (Smart Glass), HMD (Head Mounted Display), etc.), Wibro (Wibro) terminal, IPTV (Internet Protocol Television) terminal, smart TV, digital broadcasting terminal, Various terminals such as a television, a 3D television, a home theater system, an audio video navigation (AVN) terminal, an audio/video (A/V) system, a flexible terminal, and the like may be applied.

In addition, the examiner terminal, the advertiser terminal, and the user terminal each include a terminal communication unit for communicating with various devices and servers including the service providing apparatus 100, a terminal storage unit for storing various information, a terminal input unit for receiving user input, It may be configured to include a terminal display unit for displaying various information, and a terminal control unit for performing an overall control function of the terminal.

On the other hand, the service providing device 100 communicates with a shopping mall server that sells products of a plurality of different sellers online through a communication network, so that the advertisement content provided by the shopping mall server from the shopping mall server through a website or application is provided. After receiving a review request for the advertisement content and storing the advertisement content included in the examination request information received from the shopping mall server, the corresponding advertisement content may be reviewed as described below.

At this time, the service providing device 100 communicates with the seller terminals of different sellers who sell the goods through a communication network, and stores the product information received from the seller terminals in the product DB, and at the same time, through a preset web page. Product information is provided, and product information of various sellers can be provided through a corresponding web page to a user terminal accessing the web page. It may be configured as an online shopping mall server that performs payment processing for product information and enables delivery of products purchased by the user.

In this case, the seller terminal may be an advertiser terminal.

As described above, the service providing apparatus 100 may receive, from the advertiser terminal, the advertisement content posted on the advertisement publication target from the advertiser terminal.

In this case, the service providing apparatus 100 may receive the advertisement content from a separate external server that publishes the advertisement content received from the advertiser terminal to the advertisement publication target.

In addition, the advertisement publication target described in the present invention is a web page included in a website, a pop-up page, a banner included in the web page, an application page executed (provided) through the application, a specific advertisement area of the web page, Various objects into which advertisement content can be inserted, such as a specific advertisement area of an application page, may be included.

In addition, the service providing apparatus 100 includes an advertisement DB 101 for storing advertisement contents received from an advertiser terminal or an external server, and a screening result of a human judge on advertisement contents or text or image objects from a plurality of judge terminals. It may be configured to include the examination DB 102 in which examination result information is received and stored.

On the other hand, the service providing apparatus 100 uses a neural network-based learning model in which the evaluation results of a plurality of judges are learned instead of the existing judges corresponding to people examining the suitability for advertisement posting (or suitability for sending advertisements) of advertisement content. Through automatic evaluation of advertisement contents, it supports the rapid evaluation of numerous advertisement contents provided on the web or application basis, and by improving the evaluation accuracy by updating the learning model through continuous feedback. It is possible to increase the examination reliability for the automatic examination, which will be described in detail with reference to the drawings below.

2 is a detailed configuration diagram of the service providing apparatus 100 for ad screening according to an embodiment of the present invention, and FIG. 3 is a service providing method of the service providing apparatus 100 for ad screening according to an embodiment of the present invention. is a flowchart, and FIGS. 4 to 6 are diagrams illustrating an operation of the service providing apparatus 100 for advertisement review according to an embodiment of the present invention.

As illustrated, the service providing apparatus 100 may include a communication unit 110 , an extraction unit 120 , an examination unit 130 , an update unit 140 , and a control unit 150 .

In this case, at least one of the components constituting the service providing apparatus 100 may be included in the other one, for example, the communication unit 110 , the extraction unit 120 , the examination unit 130 , and the update unit 140 . may be configured to be included in the control unit 150 .

In addition, the service providing apparatus 100 may be implemented by more components than the components shown in FIG. 2 , or the service providing apparatus 100 may be implemented by fewer components than those shown in FIG. 2 .

The detailed configuration of each component will be described. First, the communication unit 110 receives the manual examination result information according to the manual examination of a plurality of examiners from an external device such as an examiner terminal, and receives the examination DB ( 102), and the examination unit 130 learns the first to third learning models preset in the examination unit 130 based on the manual examination results stored in the examination DB 102, and the first to third learning models are completed. A learning model may be generated, which will be described in more detail below (S1).

Also, the communication unit 110 may receive advertisement content from an advertiser terminal or an external server and store it in the advertisement DB 101 ( S2 ).

In this case, the service providing apparatus 100 may be configured to include the advertisement DB 101 as described above.

Also, the extraction unit 120 may select and extract advertisement contents stored in the advertisement DB 101 under the control of the control unit 150, and an image object that is not identified (or extracted) as text from the advertisement contents. ) (or object image) and at least one of text may be extracted (S3).

In this case, the extraction unit 120 may extract the image object and text from the advertisement content using various well-known algorithms or application programming interfaces (APIs) (eg, APIs provided by Google).

In addition, the advertisement content described in the present invention may be composed of an advertisement image or a frame image of an image including at least one of an image object and text, and advertisement contents are produced in addition to the image object and text. It may further include identification information for one advertiser (or advertiser terminal), publication target information for a publication target on which advertisement content is to be published, and the like.

For example, the advertisement content may include publication target information including attribute values for each attribute of the publication target including the type, size, and address of the publication target on which the advertisement content is posted.

In addition, the publication target information includes a web page address in which the advertisement content is inserted and published, a pop-up page address, an application ID, an area for an area where the advertisement content is inserted and displayed in a web page or an application page executed and displayed by the application. information may be included.

In addition, the text included in the advertisement content means an object identified as text, and the text extracted by the extraction unit 120 from the advertisement content is a letter, sign, number, character, etc. generated by identifying the text image. may mean text including , and the image object extracted from the advertisement content through the extraction unit 120 may mean an object that is not identified as text.

Therefore, the image object described in the present invention is defined as meaning an object extracted as text or not identified as text.

However, the entire image object is not extracted as text, but a text image may be included in the image object. For example, when an image object such as a shoe is extracted, a trademark marked on the shoe may also be included as a text image in the image object and extracted together.

In addition, the text extracted by the extraction unit 120 may include all texts identified by the extraction unit in the advertisement content, and the image object is included in the advertisement content while being added to the extraction unit 120 . MAY contain any object that is not identified (not extracted) as text by

In addition, the communication unit 110 may receive the result information of a plurality of examiners for a plurality of different image objects from various external devices, such as an examiner terminal, and manual examination by a plurality of examiners for a plurality of different texts. It is possible to receive result information, and it is possible to receive information on the results of manual examination by a plurality of judges for a plurality of different advertisement contents including advertisement images (or images from which text and image objects can be extracted) in which text and image objects are identified together. can receive In addition, the control unit 150 may store, in the examination DB 102 , information on the result of the examiner's manual examination received through the communication unit 110 in cooperation with the communication unit 110 .

In this case, the manual examination result information may be information on the examination result manually evaluated by the examiner on text, image object, or advertisement content.

In addition, the examination unit 130 determines the examination result of the examiner for a plurality of different image objects based on the manual examination result information stored in the examination DB 102 , the first learning model among the plurality of learning models preset in the examination unit 130 . By learning the model, the first learning model may be generated as a first learning model in which the first examination condition is automatically generated and set (S1).

In addition, the examination unit 130 learns the examination result of the examiner for the text on the basis of the manual examination result information stored in the examination DB 102 in the second learning model among the plurality of learning models preset in the examination unit 130 to learn the second learning model. The second learning model may be generated as a second learning model in which the second examination condition is automatically generated (S1).

In addition, the examination unit 130 is an advertisement purpose that includes an advertisement image in which an image object and text are combined or included or an image object and text are identified (extracted) based on the manual examination result information stored in the examination DB 102. By learning the review result of the reviewer for the content on the third learning model among the plurality of learning models preset in the review unit 130, the third learning model is automatically generated and set as a third learning model. can be (S1).

In this case, the first learning model, the second learning model, and the third learning model (first to third learning models) may be configured as independent models, respectively.

In addition, each of the first to third learning models may be configured as a deep learning algorithm.

In this case, the deep learning algorithm may be composed of one or more neural network models.

In addition, the neural network model (or neural network) described in the present invention may be composed of an input layer, one or more hidden layers, and an output layer, and the neural network model includes a Deep Neural Network (DNN). , Recurrent Neural Network (RNN), Convolutional Neural Network (CNN), Support Vector Machine (SVM), and the like, may be applied to various types of neural networks.

Accordingly, the examination unit 130 may work with the extraction unit 120 , and may receive at least one of an image object, text, and specific advertisement contents corresponding to specific advertisement contents from the extraction unit 120 . (S4).

In addition, the examination unit 130 checks the extraction state of the extraction unit 120 , and selects the text from the extraction unit 120 when the text is extracted in response to the specific advertisement content in the extraction unit 120 . received and the text provided from the extraction unit 120 may be applied to the first learning model to obtain a first examination result (S5, S8).

In this case, the first examination condition set in the first learning model is a color standard for the text, a font standard, a font size standard, a standard for one or more examination items related to a violation of public morals, a brand name, a product name, and a service name in the text. At least one of a criterion for identification of any one advertisement target name and an advertisement target name identified from the text may be included as examination items, respectively.

Accordingly, the examination unit 130 determines the reason for each examination item included in the first examination condition, whether the examination passed, and the first examination condition corresponding to the text extracted from the specific advertisement content through the first learning model. The first examination result including whether or not passed according to whether the examination has passed for each examination item may be calculated.

That is, the examination unit 130 determines whether the text does not deviate from the size range of the letters determined by the examiner to be suitable as an advertisement for the text, whether there is a word against public morals, whether the color is not complex enough to cause visual confusion, and whether the text The judge's judgment criteria such as whether at least one of a brand name, a product name, and a service name is included can be learned through the first learning model, and the text extracted from the specific advertisement content is judged using the judgment judgment criteria Information on the first examination result including the reason for passing the examination for each item and the basis for determining whether the examination for each item has passed and whether or not the examination has passed may be calculated.

In this case, the standard against the good morals described in the present invention may be based on whether or not words or image objects such as profanity, profanity, violence, and obscenity are included.

Also, the screening unit 130 may identify (extract) an advertisement target name including at least one of a brand name, a product name, and a service name from the text through deep learning-based learning.

Also, when an image object is extracted by the extraction unit 120 in response to specific advertisement content, the examination unit 130 may apply the image object to the second learning model to obtain a second examination result (S5, S8). ).

At this time, the second examination condition set in the second learning model includes at least one preset criterion for each property of an object including at least one of size, resolution, and color for the image object, one or more criteria for each examination item related to the violation of public morals, A criterion for identification of any one advertisement target name among a brand name, a product name, and a service name from the image object and at least one of an advertisement target name identified from the image object may be included as examination items, respectively.

Accordingly, the examination unit 130 corresponds to the image object extracted from the specific advertisement content through the second learning model, and determines the reason for each examination item included in the second examination condition, whether the examination passed, and the second examination condition. It is possible to calculate the second examination result including whether or not passed according to whether the examination has passed for each examination item.

That is, the examination unit 130 determines whether the image object meets the size and resolution determined by the examiner to be suitable as an advertisement, whether the product type identified from the image object is hindered (violating) morals, and the object included in the image object. Through the second learning model, the judge's judgment criteria, such as whether the color of the color or the background color is not complex enough to cause visual confusion, and whether a trade name, product name, or service name is identified from an image object, can be learned through the second learning model. Information on the second examination result, including the reason for the examination and whether the image object extracted from the specific advertisement content has passed the examination by examination item and the basis for determining whether the examination has passed by examination item using the examination determination criterion can be calculated.

Here, the screening unit 130 may identify a product type from an image object through deep learning-based learning, and an advertisement target name including at least one of a brand name, a product name, and a service name from the image object through the learning can do.

In this case, the image object may include various types of objects such as logos, products, and characters.

For example, the examination unit 130 may identify the logo or product from the image object through the second learning model and calculate a service name, product name, or brand name from the identified logo and product. This configuration uses an existing deep learning-based algorithm that classifies the type of a specific input object with respect to a specific input object and calculates a name for the type after repeatedly learning the image object and the name corresponding to the image object. 130) can be implemented.

As described above, the service providing apparatus 100 according to an embodiment of the present invention can easily be identical to and similar to an examiner through a learning model in which the evaluation criteria of the examiner as described above are learned when the advertisement content consists only of text or image objects. It is possible to automatically calculate the evaluation result by evaluating the advertisement content based on the evaluation criteria.

However, when the advertisement contents are advertisement contents that only text is identified or advertisement contents that include only image objects that are not identified as text, there is no need to judge the entire advertisement contents because only a single type of image element is included in the advertisement contents, When the text and image objects are mixed (both exist) in the advertisement content, it is necessary to judge the advertisement as a whole of the advertisement contents (or advertisement images constituting the advertisement contents).

In addition, when judging advertisements for all of the advertisement contents, the accuracy may be lower than that of the case where the evaluation is conducted only with text or only the image object is judged, or the evaluation criteria may be inconsistent with each other.

Accordingly, the service providing apparatus 100 according to the embodiment of the present invention determines the evaluation criteria for advertisement contents from which both an image object and text are extracted in consideration of the size of advertisement contents or attributes of advertisement contents such as the target of advertisement contents. When image objects and texts are mixed in and advertisement contents, the evaluation criteria that are considered when the advertisement content as a whole should be judged, such as the evaluation criteria considering the relationship between the image object and the text, such as whether there is a correlation or the arrangement pattern, are applied In this way, it is possible to support accurate evaluation of the entire advertisement content including both image objects and text, and cross-checking the results of each of the text and image objects extracted from advertisement contents with the evaluation results of the entire advertisement contents. Through cross check, when the examination results do not match, it can be supported to complement the mutual examination criteria based on a more reasonable examination result. This will be explained in detail.

First, when both the image object and the text are extracted from the specific advertisement content by the extraction unit 120 in conjunction with the extraction unit 120 (S4), the screening unit 130 is an advertisement in which the image object and the text are identified together. It is possible to obtain the third examination result by learning the examination result for the content for use and applying the specific advertisement content received (provided) from the extraction unit 120 to the third learning model in which the third examination condition is set (S8). ).

In this case, the third screening condition set in the third learning model is the size of the advertisement content according to the type and size of the target on which the advertisement content is posted, and the size range for each text and image object identified in the advertisement content. Criteria for determining a property of advertisement content including, a screening standard for each preset text-related property determined according to the property of the advertisement content, a screening standard for each preset image object related property determined according to the property of the advertisement content, the above At least one of the criteria for the arrangement pattern of the image object and text identified in the advertisement content and the criterion for whether the advertisement target name derived from the image object is the same as the advertisement target name derived from the text are included as examination items, respectively can do.

For example, the screening unit 130 determines the size of specific advertisement content and the size of the specific advertisement content based on the publication target information included in the specific advertisement content for a pop-up page, a banner, or a page area, which is a target of the specific advertisement content, through the third learning model. It is possible to determine an attribute of a specific advertisement content including a size range of text and image objects identified in the specific advertisement content.

In addition, the review unit 130 learns the advertisement content in the third learning model, identifies text and image objects from the advertisement content, and includes the same examination items as the examination items included in the first examination standard and the second examination standard. The third evaluation criterion can be set in the third learning model, but unlike the first and second learning models in which a single evaluation criterion is set for each evaluation item, the type of publication target according to the publication target information included in the advertisement content and a plurality of different examination criteria for the same examination items as the first and second examination criteria according to the size may be set in the third learning model to be included in the third examination standard.

As a detailed example of this, the review unit 130 sets the size range of the text used when determining whether the size of the text is appropriate according to the first screening criterion set in the first learning model to a plurality of each other according to the type and size of the publication target. It can be changed or subdivided into a different size range and set in the third learning model. Accordingly, the size range of the text according to the third screening criterion set in the third learning model is determined in consideration of the type and size of the publication target. 1 It may be different from the text size range according to the screening criteria.

That is, the review unit 130 determines whether or not the reviewer does not deviate from the size range of letters that the reviewer judges to be suitable as an advertisement for the text included (identified) in the advertisement content in consideration of the type and size of the publication target, and whether or not words that go against good morals The third learning model uses the same criteria as the first learning model for the judges' judgment criteria, such as whether there is, whether the colors are not complex enough to cause visual confusion, and whether the text contains at least one of a trade name, a product name, and a service name. In consideration of the type and size of the publication target while learning through For the third examination result, including the reason for the examination and whether or not the text has been passed (identified) from the specific advertisement content using the criteria, information can be calculated.

As another example, the review unit 130 identifies from the image object whether the reviewer meets the size and resolution determined to be suitable as an advertisement for the image object included (identified) in the advertisement content in consideration of the type and size of the publication target. Whether the type of product is not disturbed (violating) good morals, whether the color of the object or background color included in the image object is not complex enough to cause visual confusion, whether the brand name, product name or service name is identified from the image object While learning through the third learning model in the same way as the second learning model, the reviewer's judgment criteria, such as, etc., are generated and set in the third learning model by taking into account the type and size of the publication target. Whether or not the image object extracted (identified) from the specific advertisement content has passed the examination by each examination item and whether the examination has been passed by the examination item using the examination criterion corresponding to the type and size of the publication target according to the specific advertisement content It is possible to calculate the information on the third examination result including the examination reason for the basis of the determination of whether or not and whether or not it has passed.

In other words, through the third learning model, the screening unit 130 determines the selected screening criteria in consideration of the type and size of a publication target according to a specific advertisement content among a plurality of screening criteria set for each screening item identical to the first and second screening criteria. Accordingly, it is possible to calculate the third examination result including whether each examination item has passed and the reason for examination.

In this case, the review unit 130 may calculate the third examination result for whether each examination item has passed and the examination reason according to the selected examination criteria in consideration of only the size of the publication target.

Also, the type of publication target may be classified by a reviewer, and for example, a web page, a pop-up page, a banner, and an application page may be classified into different types of publication target.

In addition, the screening unit 130 judges the relationship between the image object and the text to be judged when the entire content, which is a screening item that does not exist in the first learning model and the second learning model, is to be judged. can be trained in the third learning model, for example, the advertisement target name derived from the reference for the arrangement pattern of the image object and text identified in the advertisement content and the advertisement target name derived from the image object is the same as the advertisement target name derived from the text A third examination condition each including a criterion for whether or not to do so as an examination item may be set in the third learning model.

Accordingly, the third examination result includes the reason for each examination item included in the third examination condition, whether the examination has passed, and whether or not the examination has passed according to whether the examination has been passed by each examination item. Some of the plurality of examination items may include examination items corresponding to (or identical to) the first and second examination results.

Accordingly, the examination unit 130 applies the specific advertisement content to the third learning model, corresponds to the specific advertisement content, and determines the reason for each examination item included in the third examination condition and whether the examination passed, and the third The third examination result including whether or not passed according to whether the examination has passed for each examination item included in the examination condition may be calculated.

Also, when the first examination result is calculated, the examination unit 130 generates first examination result information on the first examination result and provides it to the control unit 150, and when the second examination result is calculated, the second examination result information The second examination result information may be generated and provided to the control unit 150, and when the third examination result is calculated, third examination result information on the third examination result is generated and transmitted to the control unit 150 can provide

Meanwhile, the controller 150 may include a microprocessor, an engine, and the like. For example, the control unit 150 may include RAM, ROM, CPU, GPU, and a bus, and the RAM, ROM, CPU, GPU, etc. may be connected to each other through a bus.

The control unit 150 is configured to determine the advertisement content based on any one of the first and second evaluation results when there is no third evaluation result in the examination result obtained through the examination unit corresponding to the specific advertisement content from the examination unit 130 . You can decide whether or not to pass the content.

That is, when only a text or image object is identified in the specific advertisement content through the examination unit 130 and the third examination result through the third learning model is not calculated, the control unit 150 determines whether the first and final examination result information, which is a final examination result for whether the specific advertisement content is passed or not, may be calculated based on any one of the second examination results.

At this time, the control unit 150 receives only one of the first and second examination results from the examination unit 130 , and when any one of the received first and second examination results is a pass, the final result corresponding to the specific advertisement content It is determined that the examination result is passed, and when either one of the received first and second examination results is unsuccessful, the final examination result corresponding to the specific advertisement content may be determined as failed (S6, S7).

In addition, when the third screening result is obtained (received) through the screening unit 130, the control unit 150 determines that the final screening result of the specific advertisement content is rejected when at least one of the first to third screening results is unsuccessful. It is determined, and only when all of the first to third examination results are passed, the final examination result of the specific advertisement content may be determined as pass (S9 and S10).

In addition, the control unit 150 may control the communication unit 110 , the extraction unit 120 , and the examination unit 130 , and a specific advertisement corresponding to the specific advertisement content and corresponding to the final examination result when the final examination result is calculated. Content may be extracted from the advertisement DB 101 and included in the final examination result information corresponding to the final examination result.

As described above, the present invention extracts text and image objects from advertisement content, performs screening on each of the text and image objects based on the evaluation criteria according to the reviewer's learning of the examination results, and applies the obtained examination results to the entire advertisement contents. Each individual element composing advertisement content is determined by automatically determining whether the advertisement content has passed the examination based on the examination result and checking whether the examination results obtained by conducting the examination on the basis of the examination criteria according to the examination result learning The results and advertisements when individual elements composing advertisement contents are separated from advertisement contents are judged together to ensure that there are no screening factors omitted in the screening process Advertisement content is judged to have passed the screening only when consensus is reached through cross-checking between the screening results when individual elements are judged as a whole. can

Meanwhile, as shown in FIG. 5 , the service providing apparatus 100 further includes an update unit 140 , and the update unit 140 works with the control unit 150 to correspond to the specific advertisement content. When at least one of the first and second examination results is a failure and the third examination result is a pass, the third learning model learns the examination result corresponding to the failure among the first and second examination results in the third learning model The model can be updated.

That is, for the examination items of the first learning model and the second learning model, which are determined by learning only any one of text or image objects, the reliability is higher than that of the third learning model, which is determined as the entire content in which text and image objects are mixed. As can be seen, if the evaluation result is judged as failing in any one of the first and second learning models, but the third learning model is judged to be pass, and no consensus (concord) is made, the first and second learning models It is possible to update the third learning model by learning the information about the examination result on the examination result determined to be unsuccessful in the third learning model, and through this, the examination accuracy of the third learning model is separated from the advertisement content by text or image object. (extraction) can be supplemented to increase the accuracy of the examination.

In addition, the update unit 140 receives the first to third examination results from the examination unit 130 to the control unit 150 , and the third examination result is unsuccessful and at least one of the first and second examination results. When is passed, from among the first and second examination results, examination contents for the same examination items as one or more examination items included in the pass specific examination result may be selected and extracted from the third examination result.

Accordingly, when the specific examination result and the examination content do not match, the update unit 140 may generate error information including the specific examination result and examination content and provide it to the controller 150 .

Also, the communication unit 110 may communicate with the examiner's terminal through a communication network.

Accordingly, the control unit 150 generates examination review request information including at least one of a final examination result for advertisement content and corresponding error information, transmits it to a preset examiner's terminal, and receives the examination review from the examiner's terminal. When correction information on at least one of the final examination result and error information is received from the examiner terminal in response to the request information, a learning model corresponding to the correction information among the first to third learning models based on the correction information It can be updated by learning the correction information.

That is, the service providing apparatus 100 generates and provides error information to the examiner when an error occurs because the examination result for each text and image object extracted from the advertisement content and the examination result for the advertisement content do not match with each other. or provide information on the final evaluation result to the reviewer, and if an error or final review result is requested by the reviewer, the learning model is re-trained based on the correction request to adapt the learning model to meet the reviewer’s assessment criteria. It can be updated, and through re-learning of this learning model, the screening accuracy can be greatly improved.

As described above, in the present invention, when the advertisement content consists only of text or image objects, through the learning model in which the evaluation criteria of the judges as described above are learned, the advertisement contents are easily evaluated using the same and similar screening criteria as the judges, and the evaluation results are obtained. In addition to being able to automatically calculate quickly and accurately, the evaluation criteria that is considered when the advertisement content from which both image objects and text are extracted should be judged as the entire advertisement content is applied to the entire advertisement content including both image objects and texts. It can support accurate screening of advertisement content, and a more reasonable judgment when the screening results do not match each other through cross-checking that compares the screening results for text or image objects extracted from advertisement contents with the screening results for the entire advertisement contents. By supporting the supplementation of the evaluation criteria by automatically re-learning the learning model based on the results or by proceeding with the process of re-learning the learning model based on the correction information received through a request for review of the examination results to the examiner, It is possible to increase the accuracy and reliability of content review.

Meanwhile, as shown in FIG. 6 , when the final examination result of the specific advertisement content is passed based on the examination result provided by the examination unit 130 , the control unit 150 selects the specific advertisement content as the target of publication. Advertisement content may be published and transmitted to a user terminal accessing the publication target.

In this case, the control unit 150 may communicate with an external server including the target of advertisement content through the communication unit 110 , and transmit the advertisement content for which the final examination result is passed to the external server so that the advertisement content is transmitted. It can be published in the above publication target.

In addition, when the final evaluation result of the specific advertisement content is unsuccessful, the control unit 150 transmits the final evaluation result (final evaluation result information) including the specific advertisement content and the reason for the evaluation to the advertiser terminal corresponding to the specific advertisement content. can be transmitted

Meanwhile, in the above configuration, the user terminal includes a user terminal communication unit, a user terminal storage unit, a user terminal display unit, a user terminal voice output unit, and a user control unit. Not all components of the user terminal are essential components, and the user terminal may be implemented by more components than the components of the user terminal, or the user terminal may be implemented by fewer components.

In this case, the advertiser terminal and the reviewer terminal may include the same configuration as the user terminal.

The user terminal communication unit communicatively connects with any internal component or at least one external terminal through a wired/wireless communication network. In this case, any external terminal may include the service providing apparatus 100, an external server, and the like. Here, as wireless Internet technologies, wireless LAN (WLAN), DLNA (Digital Living Network Alliance), WiBro (Wireless Broadband: Wibro), WiMAX (World Interoperability for Microwave Access: Wimax), HSDPA (High Speed Downlink Packet Access) ), High Speed Uplink Packet Access (HSUPA), IEEE 802.16, Long Term Evolution (LTE), Long Term Evolution-Advanced (LTE-A), Wireless Mobile Broadband Service (WMBS), 5G mobile communication service, Bluetooth, LoRa (Long Range), RFID (Radio Frequency Identification), Infrared Data Association (IrDA), UWB (Ultra Wideband), ZigBee, Near Field Communication: NFC), ultrasonic communication (Ultra Sound Communication: USC), visible light communication (VLC), Wi-Fi (Wi-Fi), Wi-Fi Direct (Wi-Fi Direct), etc. may be included. In addition, wired communication technologies include wired LAN (Local Area Network), wired WAN (Wide Area Network), Power Line Communication (PLC), USB communication, Ethernet, serial communication, and optical/ A coaxial cable and the like may be included.

In addition, the user terminal communication unit may mutually transmit information with an arbitrary terminal through a universal serial bus (USB).

In addition, the user terminal communication unit technical standards or communication methods for mobile communication (eg, Global System for Mobile communication (GSM), Code Division Multi Access (CDMA), Code Division Multi Access 2000 (CDMA2000), EV-DO (Enhanced Voice-Data Optimized or Enhanced Voice-Data Only), Wideband CDMA (WCDMA), High Speed Downlink Packet Access (HSDPA), High Speed Uplink Packet Access (HSUPA), Long Term Evolution (LTE), Long Term Evolution (LTE-A) Term Evolution-Advanced (Term Evolution-Advanced, etc.)) and transmits and receives radio signals to and from the base station, the service providing apparatus 100, various servers, and the like on a mobile communication network.

In addition, the user terminal communication unit receives various types of information transmitted from the service providing apparatus 100 under the control of the user control unit.

The user terminal storage unit stores various user interfaces (UIs), graphic user interfaces (GUIs), and the like.

In addition, the user terminal storage unit stores data and programs necessary for the user terminal to operate.

That is, the user terminal storage unit may store a plurality of application programs (application programs or applications) driven in the user terminal, data for operation of the user terminal, and commands. At least some of these application programs may be downloaded from the external service providing device 100 through wireless communication. In addition, at least some of these application programs may exist on the user terminal from the time of shipment for basic functions (eg, incoming calls, outgoing functions, message reception, and outgoing functions) of the user terminal. Meanwhile, the application program may be stored in the user terminal storage unit, installed in the user terminal, and driven to perform an operation (or function) of the user terminal by the user control unit.

In addition, the user terminal storage unit is a flash memory type (Flash Memory Type), a hard disk type (Hard Disk Type), a multimedia card micro type (Multimedia Card Micro Type), a card type memory (eg, SD or XD memory, etc.) , magnetic memory, magnetic disk, optical disk, random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable memory (PROM) Read-Only Memory) may include at least one storage medium. Also, the user terminal may operate a web storage that performs a storage function of the user terminal storage on the Internet, or may operate in connection with the web storage.

In addition, the user terminal storage unit may store interface-related information, web page-related information, and content received through the user terminal communication unit under the control of the user control unit.

The user terminal display unit may display various contents such as various menu screens using a user interface and/or a graphic user interface stored in the user terminal storage unit under the control of the user control unit. Here, the content displayed on the display unit of the user terminal includes various text or image data (including various information data) and a menu screen including data such as icons, list menus, and combo boxes. In addition, the user terminal display unit may be a touch screen.

In addition, the user terminal display unit is a liquid crystal display (LCD), a thin film transistor liquid crystal display (TFT LCD), an organic light-emitting diode (Organic Light-Emitting Diode: OLED), a flexible display (Flexible Display), a three-dimensional display (3D display), an e-ink display (e-ink display), may include at least one of LED (Light Emitting Diode).

In addition, the user terminal display unit may be configured as a stereoscopic display unit for displaying a stereoscopic image.

A three-dimensional display method such as a stereoscopic method (glasses method), an auto stereoscopic method (glasses-free method), or a projection method (holographic method) may be applied to the stereoscopic display unit.

In addition, the user terminal display unit displays the interface-related information, web page-related information, and content received through the user terminal communication unit under the control of the user control unit.

The user terminal voice output unit outputs voice information included in a signal processed by a predetermined signal by the user control unit. Here, the user terminal voice output unit may include a receiver, a speaker, a buzzer, and the like.

In addition, the user terminal voice output unit outputs a guide voice generated by the user control unit.

In addition, the user terminal voice output unit may output voice information corresponding to the information received through the user terminal communication unit under the control of the user control unit.

The user control unit executes an overall control function of the user terminal.

In addition, the user control unit executes the overall control function of the user terminal by using the program and data stored in the user terminal storage unit. The user control unit may include RAM, ROM, CPU, GPU, and a bus, and the RAM, ROM, CPU, GPU, etc. may be connected to each other through a bus. The CPU can access the user terminal storage unit to perform booting using the O/S stored in the user terminal storage unit, and perform various operations using various programs, contents, data, etc. stored in the user terminal storage unit. can

In this case, the user control unit may be an application unit configured in the above-described user terminal, and of course, the application unit may be composed of application-related data stored in the user terminal storage unit and an application executed by the user control unit.

In addition, the user control unit performs a membership registration procedure for the user of the corresponding user terminal by interworking with the service providing apparatus 100 .

In addition, when performing the membership registration procedure, the user control unit must complete the authentication function through the identification means (eg, mobile phone, credit card, i-PIN, e-mail, etc. included) to perform the membership registration procedure for the service providing device 100 normally. can be completed

In addition, if the user of the user control unit is not a member registered in advance in the service providing device 100 , the user control unit is based on the user terminal (or the corresponding user device 100) based on the guide information related to membership registration provided from the service providing device 100 . member registration procedure for the user of the user terminal).

In addition, the user terminal may further include a user terminal interface unit that functions as an interface with all external devices connected to the user terminal. For example, the user terminal interface unit includes a wired/wireless headset port, an external charger port, a wired/wireless data port, a memory card port, a port for connecting a device equipped with an identification module, an audio I/ It may be composed of an input/output (O) port, a video input/output (I/O) port, and an earphone port. Here, the identification module is a chip storing various information for authenticating the user terminal's usage right, and includes a user authentication module (UIM), a subscriber identity module (SIM), and a universal user authentication module (Universal). Subscriber Identity Module: USIM) and the like. In addition, the device provided with the identification module may be manufactured in the form of a smart card. Accordingly, the identification module may be connected to the user terminal through the port. Such a user terminal interface unit receives data or power from an external device and transmits it to each component inside the user terminal, or transmits data inside the user terminal to an external device.

In addition, the user terminal interface unit is a path through which power from the cradle is supplied to the user terminal when the user terminal is connected to an external cradle, or a path through which various command signals input from the cradle by the user are transmitted to the user terminal can be Various command signals or corresponding power input from the cradle may be operated as signals for recognizing that the user terminal is correctly mounted on the cradle.

In addition, the user terminal may further include an input unit for receiving a command or control signal generated by a manipulation such as a button manipulation by the user or a signal according to an arbitrary function selection, or touch/scroll a displayed screen. may be

The input unit is a means for receiving at least one of a user's command, selection, data, and information, and may include a plurality of input keys and function keys for receiving numeric or character information and setting various functions.

In addition, the input unit includes a key pad, a dome switch, a touch pad (static pressure/capacitance), a touch screen, a jog wheel, a jog switch, a jog shuttle, and a mouse. , various devices such as a stylus pen, a touch pen, and the like may be used. In particular, when the user terminal display unit is formed in the form of a touch screen, some or all of the functions of the input may be performed through the user terminal display unit.

In addition, each component (or module) of the user terminal may be software stored in a memory (or user terminal storage) of the user terminal. The memory may be an internal memory of the user terminal, an external memory, or another type of storage device. Also, the memory may be a non-volatile memory. The software stored on the memory may include a set of instructions that, when executed, cause the user terminal to perform a specific operation.

The user terminal, the reviewer terminal, the advertiser terminal, and the service providing apparatus 100 described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components.

In addition, the components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), and a programmable logic unit (PLU). unit), a microprocessor, or any other device capable of executing and responding to instructions, may be implemented using one or more general purpose or special purpose computers.

The advertiser terminal, the reviewer terminal, the user terminal, and the service providing apparatus 100 may execute an operating system (OS) and one or more software applications executed on the operating system.

In addition, the user terminal, the advertiser terminal, the reviewer terminal, and the service providing apparatus 100 may access, store, manipulate, process and generate data in response to the execution of the software.

For convenience of understanding, each of the components is sometimes described as being used one, but those of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing. It can be seen that elements can be included.

For example, the advertiser terminal, the reviewer terminal, the user terminal, and the service providing apparatus 100 may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The software may include a computer program, code, instruction, or a combination of one or more thereof, and operates the advertiser terminal, the reviewer terminal, the user terminal, and the service providing apparatus 100 as desired. may be ordered, independently or collectively.

The software and/or data may be any type of machine, component, physical device, virtual equipment), a computer storage medium or device, or a transmitted signal wave (signal wave) may be permanently or temporarily embodied (embody).

The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The service providing method for advertisement screening according to the embodiment of the present invention described in the above embodiment can be written as a computer program, and codes and code segments constituting the computer program can be easily inferred by a computer programmer in the art. can In addition, the computer program is stored in a computer readable media, and is read and executed by a computer or the service providing apparatus 100 according to an embodiment of the present invention, thereby providing a service for advertisement review. method can be implemented.

The information storage medium includes a magnetic recording medium and an optical recording medium. A computer program for implementing the service providing method for advertisement screening according to an embodiment of the present invention may be stored and installed in the internal memory of the service providing apparatus 100 . Alternatively, an external memory such as a smart card storing and installing a computer program for implementing the service providing method for advertisement screening according to an embodiment of the present invention may be mounted in the service providing apparatus 100 through an interface.

The various devices and components described herein may be implemented by hardware circuitry (eg, CMOS-based logic circuitry), firmware, software, or a combination thereof. For example, it may be implemented using transistors, logic gates, and electronic circuits in the form of various electrical structures.

Those of ordinary skill in the art to which the present invention pertains may modify and modify the above-described contents without departing from the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention, but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. The protection scope of the present invention should be construed by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

100: service providing device 110: communication unit
120: extraction unit 130: examination unit
140: update unit 150: control unit

Claims (10)

As a service providing method for advertisement review of a service providing device,
Receiving step of receiving advertisement content;
an extraction step of extracting at least one of an image object and text not identified as text from the advertisement content;
When an image object is extracted from the advertisement content among the first learning model in which the first examination condition is set by learning the examination result for the image object, and the second learning model in which the second examination condition is set by learning the examination result for the text, the second A first evaluation result is obtained by applying to the learning model, and when text is extracted from the advertisement content, a second evaluation result is obtained by applying to the second learning model, and both image objects and text are extracted from the advertisement contents. a screening step of acquiring a third screening result by learning a screening result for advertisement content in which an image object and text are identified together, and applying the advertisement content to a third learning model in which a third screening condition is set;
If there is no third screening result in the screening result obtained through the screening step in response to the advertisement content, it is determined whether or not to pass the advertisement content based on any one of the first and second screening results, and a decision step of determining a final evaluation result of the advertisement content as disqualified when at least one of the first to third evaluation results is disqualified when a third evaluation result is obtained through the step;
A method of providing services for advertising screening, including.
The method according to claim 1,
after the decision step
When at least one of the first and second examination results is a failure and the third examination result is a pass, the third learning model learns the examination result corresponding to the failure among the first and second examination results to learn the third Service providing method for advertisement review, characterized in that it further comprises the step of updating the learning model.
The method according to claim 1,
after the decision step
Review request information including the final examination result for the advertisement content is generated and transmitted to a preset examiner's terminal, and the final examination result or error information from the examiner's terminal in response to the examination review request information from the terminal and a teaching step of learning and updating the correction information in a learning model corresponding to the correction information among the first to third learning models based on the correction information when receiving correction information for How to provide services for advertising review.
4. The method according to claim 3,
after the decision step
When the third examination result is failing and at least one of the first and second examination results is a pass, the contents of the examination for the same examination item as the examination item of the specific examination result which is a pass among the first and second examination results are recalled. Selecting from the third examination result, and generating error information when the specific examination result and the examination content do not match each other;
In the teaching step, when generating the error information, instead of the final examination result, examination review request information including the error information is generated and transmitted to the examiner's terminal, and the learning is performed based on the correction information received in response to the error information. A service provision method for advertisement review, characterized in that the model is trained and updated.
The method according to claim 1,
The first examination condition is the color standard for the text, the font standard, the font size standard, the standard for one or more examination items related to the violation of public morals and customs, and the advertisement target name of any one of a brand name, a product name, and a service name in the text. At least one of a criterion for identification and an advertisement target name identified from the text is included as an examination item, respectively, and the first examination result includes an examination reason for each examination item included in the first examination condition, whether the examination has passed, and the examination A method of providing a service for advertisement screening, characterized in that it includes whether or not the pass is passed according to whether the examination has been passed by item.
The method according to claim 1,
The second examination condition includes one or more preset criteria for each property of an object including at least one of size, resolution, and color for the image object, one or more criteria for each examination item related to the violation of public morals, a brand name, and a product name from the image object. and at least one of a criterion for identification of an advertisement target name among service names and an advertisement target name identified from the image object, respectively, as an examination item, and the second examination result includes the information included in the second examination condition. A method for providing a service for advertisement screening, characterized in that it includes the reason for the examination by each examination item, whether the examination has passed, and whether or not the examination has been passed according to whether the examination has been passed for each examination item.
The method according to claim 1,
The third examination condition is to determine the properties of the advertisement content including the size of the advertisement content and the size range for each text and image object identified in the advertisement content according to the type and size of the publication target on which the advertisement content is posted. criteria for the advertisement content, screening criteria for each preset text-related attribute determined according to the attribute of the advertisement content, screening criteria for each preset image object-related attribute determined according to the attribute of the advertisement content, and image object and text identified in the advertisement content At least one of a criterion for the arrangement pattern of , and a criterion for whether the advertisement target name derived from the image object is the same as the advertisement target name derived from the text is included as an examination item, and the third examination result includes the The reason for the examination by each examination item included in the third examination condition, whether the examination passed or not, and whether the examination passed according to the examination by the above examination items are included, and some of the examination items included in the third examination result are 2 A service provision method for advertisement review, characterized in that it corresponds to the examination items of the examination result.
8. The method of claim 7,
The method of claim 1, wherein the advertisement content includes publication target information including attribute values for each attribute of the publication target including the type and size of the publication target on which the advertisement content is posted.
The method according to claim 1,
If the final screening result determined through the determining step is pass, the advertisement content is posted on the target of the advertisement content and transmitted to a user terminal accessing the target. An advertisement processing step of transmitting the final examination result including the reason to the advertiser terminal corresponding to the advertisement content
Service providing method for advertisement review, characterized in that it further comprises.
a communication unit for receiving advertisement content;
an extraction unit for extracting at least one of an image object and text that are not identified as text from the advertisement content;
When an image object is extracted from the advertisement content among the first learning model in which the first examination condition is set by learning the examination result for the image object, and the second learning model in which the second examination condition is set by learning the examination result for the text, the second A first evaluation result is obtained by applying to the learning model, and when text is extracted from the advertisement content, a second evaluation result is obtained by applying to the second learning model, and both image objects and text are extracted from the advertisement contents. an examination unit for learning the examination result of the advertisement content in which the image object and the text are identified together, and applying the advertisement content to a third learning model in which a third examination condition is set to obtain a third examination result; and
If there is no third examination result in the examination result obtained through the examination unit corresponding to the advertisement content, it is determined whether or not to pass the advertisement content based on any one of the first and second examination results, and the examination step When a third screening result is obtained through
A service providing device for ad review comprising a.

Images