KR102564646B1 - Method and system for automatically generating online product information - Google Patents

Abstract

The present invention relates to a method and system for automatically generating online product information that automatically generates and provides information on products sold or promoted online. The present invention provides an automatic online product information generation system that receives basic product information from a manager terminal and automatically generates product information to be posted online, wherein a management server to which the manager terminal accesses and information related to the product is stored. A product information collection unit that collects product information by crawling a web server connected to a network based on the basic information; a price calculation unit calculating a sales price of the product based on price information among the product information; an image selection unit selecting at least one image to be posted on a web page from among the product information collected by the product information collection unit; and a web page generation unit generating the web page for product information using the product information collected by the product information collection unit.

Description

Method and system for automatically generating online product information {METHOD AND SYSTEM FOR AUTOMATICALLY GENERATING ONLINE PRODUCT INFORMATION}

The present invention relates to a method and system for automatically generating online product information. More specifically, the present invention relates to a method and system for automatically generating online product information that automatically generates and provides information on products sold or promoted online.

BACKGROUND ART Electronic commerce refers to a transaction in which products or services are bought and sold on a network such as the Internet using a computer, that is, in a virtual space using an electronic medium (system). BACKGROUND Electronic commerce has rapidly developed with the development of electronic communication technology, and various products are transacted through electronic commerce.

In order to sell products using an online shopping mall, which is an example of electronic commerce, a product seller creates product information of a product to be sold and uploads it to the online shopping mall. The product information includes various information such as a product name, a product image, a manufacturer, and a seller. Typically, such product information is prepared according to a product registration form provided by an Internet shopping mall or is individually prepared by a worker as an individual web page.

Meanwhile, interest in so-called refurbished products has recently increased. Refurbished products refer to products that have been displayed or are resold due to returns or exchanges after initial packaging and shipment. These refurbished products are sold at a discount compared to normal products.

However, there are many cases in which it is difficult to accurately check product information due to not only various types of refurbished products, but also product release timing. Accordingly, product sellers dealing with refurbished products suffer from difficulties in effectively providing product information to online shopping malls.

In addition, depending on how product information using an online shopping mall is configured, the effect on consumer's interest or desire to purchase a product varies. Therefore, there is a problem in that sufficient advertisement or promotion is not achieved by simply collecting information.

Republic of Korea Patent Publication No. 10-2016-0042472

In order to solve the above problems, an object of the present invention is to provide a method and system for automatically generating online product information that automatically generates product information to be provided to an Internet shopping mall using basic product information such as a product name or code. to be

In addition, an object of the present invention is to provide a method and system for automatically generating online product information capable of increasing the effect of advertising or publicizing products.

The present invention provides an automatic online product information generation system that receives basic product information from a manager terminal and automatically generates product information to be posted online, wherein a management server to which the manager terminal accesses and information related to the product is stored. A product information collection unit that collects product information by crawling a web server connected to a network based on the basic information; a price calculation unit calculating a sales price of the product based on price information among the product information; an image selection unit selecting at least one image to be posted on a web page from among the product information collected by the product information collection unit; and a web page generation unit generating the web page for product information using the product information collected by the product information collection unit.

In one embodiment, the product DB includes a stocked product DB for storing a list of stocked products and a product information DB for storing the product information collected by the product information collection unit, and the management server, A product search unit for searching for a related product in the stocked product DB based on the basic information of the product input from the manager terminal is further included.

In addition, the product search unit may select a specific product from a plurality of products by using a priority order when a plurality of products are retrieved from the stock product DB.

Preferably, the price calculation unit adds a profit to any one of a standard selling price, an average product price, a maximum price or a minimum price of the product among the prices of the product collected by the product information collection unit, or multiplies a discount rate to sell the product. price can be calculated.

In one embodiment, the image selection unit checks the source of the image collected by the product information collection unit, and determines whether or not the image is selected according to the source, an image source analysis unit, a photographing angle of the image, An image characteristic analyzer that determines whether or not the image is selected according to at least one of the resolution of the image or whether a person or a background is included in the image, and analyzes the advertising efficiency according to the image to determine that the advertising efficiency is high. It may include at least one of advertisement efficiency analyzers that select a predicted image.

The image selection unit may further include an image correction unit generating a corrected image by modifying the image collected by the product information collection unit.

In one embodiment, the management server may further include a shopping mall posting unit that selects a sales server to post the web page or posts the web page to the sales server.

In one embodiment, the product DB further includes a dispute DB for storing dispute prediction information related to the sale of the product, and the shopping mall posting unit refers to the dispute DB when a dispute is expected in the sale of the product. , The web page is posted only to a closed sales server with limited access among the sales servers, and if a dispute is not expected in the sale of the product by referring to the dispute DB, the web page is posted to a closed sales server with limited access among the sales servers. It can be posted on both open sales servers where access is not restricted.

In addition, the present invention provides a method for automatically generating online product information in which a management server receives basic product information from an administrator terminal and automatically generates product information to be posted online, (a) a product search unit of the management server Searching for and selecting the product from the warehousing product DB; (b) collecting product information on a selected product by crawling a web server connected to a network by the product information collection unit of the management server; (c) selecting at least one product image to be used as the product information to be posted by the image selection unit of the management server; (d) determining information to be posted on the web page by the web page creation unit of the management server; and (e) generating the webpage by the webpage generation unit.

In one embodiment, in the step (a), when the number of products retrieved from the stocked product DB is plural, the product search unit may select a specific product from among the plurality of products by using a priority order.

In addition, in the step (d), the price calculation unit of the management server determines any one of a standard selling price, an average product price, a maximum price or a minimum price of the product among the prices of the product collected by the product information collection unit. You can calculate the selling price by adding the margin or multiplying by the discount rate.

In one embodiment, the step (c) may include the step of determining, by the image selector, the source of the image collected by the product information collection unit, and whether to select the image according to the source; Determining, by a selection unit, whether or not to select the image according to at least one characteristic of a photographing angle of the image, whether a person or a background is included in the image, or a resolution of the image, and the image selection unit, At least one of selecting an image predicted to have high advertising efficiency by analyzing advertising efficiency according to the method may be included.

In one embodiment, the web page generated in step (e) includes a product name unit displaying the name of the product, a product image unit displaying the product image selected in the image selection unit, and price information of the product. is displayed, the product code unit displays the barcode or item number of the product, the inventory quantity unit displays the quantity in stock of the product, the product description unit displays the description of the product, and the product information collection unit collects It may include a product information provision notice section in which the product information provision notice displayed is displayed.

In addition, the product information collecting unit may collect URL information of another web page for the product, and the web page may further include an external link according to the URL information.

The method of automatically generating online product information may further include (f) selecting a sales server to post the webpage or posting the webpage to the sales server, by the shopping mall posting unit of the management server. can

In one embodiment, in the step (f), if a dispute is expected in the sale of the product by referring to the dispute DB, the shopping mall posting unit posts the web page only to a closed sales server with limited access among the sales servers, and , If a dispute is not expected in the sale of the product by referring to the dispute DB, the webpage can be posted on both a closed sales server with restricted access and an open sales server with unrestricted access among the sales servers.

According to the present invention, product information to be provided to an Internet shopping mall is automatically generated using basic product information such as a product name or code, thereby eliminating the inconvenience of manually generating product information and providing a plurality of product information in a short time. There is an advantage in that it can be created and exposed to online shopping malls.

In addition, according to the present invention, there is an advantage in that the sales rate of the product can be increased by increasing the consumer's desire to purchase the product.

In addition, according to the present invention, even when product information is not sufficient, such as a refurbished product, there is an effect of effectively acquiring product information, automatically generating product information, and uploading the product information to an online shopping mall.

1 is a diagram schematically illustrating a system for automatically generating online product information according to a preferred embodiment of the present invention.
2 is a block diagram showing the configuration of a management server included in the system for automatically generating online product information according to a preferred embodiment of the present invention.
3 shows an example of a web page generated by a web page generating unit of the system for automatically generating online product information according to a preferred embodiment of the present invention.
4 is a flowchart illustrating a method of automatically generating online product information according to a preferred embodiment of the present invention.
5 is a block diagram showing the configuration of the image selection unit of the management server included in the system for automatically generating online product information according to a preferred embodiment of the present invention.
6 shows an example of an advertisement, (a) is an example of advertisement information in the form of an image, (b) is a diagram for explaining a background separation process of advertisement information, and (c) is a diagram from the background separated advertisement information. It is a diagram for explaining a process of extracting an advertisement target item.
7 is a graph showing an example of modeling for learning advertisement information.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, in adding reference numerals to components of each drawing, it should be noted that the same components have the same numerals as much as possible even if they are displayed on different drawings. In addition, in describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description will be omitted. In addition, although preferred embodiments of the present invention will be described below, the technical idea of the present invention is not limited or limited thereto and can be modified and implemented in various ways by those skilled in the art.

A product subject to automatic generation of online product information according to the present invention may be a new product that is being sold or a refurbished product that is subject to resale. In particular, in the case of a refurbished product, it is a product that is sold by re-posting product information online for resale by return, exchange, etc., and there may be various restrictions on the resale of the product. For example, if the seller of the original product and the seller of the refurbished product are different, the seller of the refurbished product may not be able to use the product information (eg, product image, price, sales conditions, etc.) posted by the original product seller. Alternatively, there may be a situation in which a manufacturer or distribution company prohibits or restricts the resale of a product when reselling the product. In addition, since refurbished products may be of various types and in small quantities, it may be practically difficult for sellers of refurbished products to newly create and post product information for each individual product.

In order to solve this problem, the present invention can promote convenience of product sellers by automatically generating product information on refurbished products as well as selling new products and posting them on an online shopping mall.

1 is a diagram schematically illustrating a system for automatically generating online product information according to a preferred embodiment of the present invention.

The system 1 for automatically generating online product information according to a preferred embodiment of the present invention includes a management server 100 to which a manager terminal 10 accesses. The management server 100 may be connected to a product DB 200 that stores product-related information.

The manager terminal 10 inputs basic information of a product to which product information is automatically generated to the management server 100, and visually outputs a web page for the product information automatically generated by the management server 100. perform the function of The manager terminal 10 may include input devices such as a keyboard, a touch panel, and a barcode reader. A manager may input basic product information into the manager terminal 10 using a keyboard, a touch panel, or a barcode reader.

The basic information of the product may be the product name, product number (model number or serial number), etc., and a code such as a barcode or QR code displayed on the product may be used as the basic information of the product.

In addition, the manager terminal 10 manages a list of warehousing products such as the name, product number (or model number or serial number), barcode, manufacturer, quantity, warehousing destination, warehousing date, and warehousing price of the warehousing product when warehousing the product. It can be transmitted to the server 100 and stored in the product DB 200. Such a list of goods received is exemplified in Table 1, and the information included in the list of goods received may be different from that illustrated in Table 1, of course.

turn product name part number barcode manufacturing company quantity Where to wear Receipt date stocking price One USB speaker KA123MU 88012349841 ○○○ 5 Mall A 2020-04-20 \52,000 ·
·
· ·
·
· ·
·
· ·
·
· ·
·
· ·
·
· ·
·
·

The manager terminal 10 transmits the list of warehousing products as illustrated in Table 1 to the management server 100, while inputting basic product information for each product into the management server 100 to create a web page for product information. may be requested to the management server 100. For example, when a plurality of different products are stocked at once, information on the plurality of products is included in the stocked product list and transmitted to the management server 100, and a web page for each stocked product (same type of product) is created. may be individually requested through the manager terminal 10. A manager can automatically create a web page for an individual product in the management server 100 by inputting the name, product number, or barcode of each product in the manager terminal 10 .

In one embodiment, the stocked product is a refurbished product, the refurbished product is received from a warehouse and stored, and a list of stocked products may be provided to the manager of the manager terminal 10 at the time of storage. The manager may transmit basic information of the stocked goods through the manager terminal 10 to the management server 100 and request automatic creation of a product information web page for each individual product.

The management server 100 generates product information (ie, a web page for product information) to be presented to consumers based on basic product information input from the manager terminal 10 . Specifically, the management server 100 collects product information based on basic product information using the network 500 constituting the Internet, selects appropriate product information, and automatically creates a web page for product information. do. The management server 100 may be connected to the product DB 200 . The product DB 200 may store information on a specific product that has already been collected or, in the case of a new product, store basic product information according to a barcode or QR code of the product. In addition, product information collected by the management server 100 may be newly stored in the product DB 200 .

The management server 100 may be connected to a plurality of web servers 300 through the network 500 . The web server 300 may be a target for which the management server 100 crawls information on a specific product. Such a web server 300 can be understood as any object that can be accessed through the network 500, such as a server for a manufacturer of products, various Internet shopping malls, and personal blogs.

The management server 100 classifies a plurality of product information collected by crawling in the web server 300 according to a predetermined criterion, extracts optimal product information, and creates a web page for product information of the corresponding product.

A web page for product information generated by the management server 100 may be transmitted to and posted to sales servers 400 and 410 for selling products. In one embodiment, the sales server may be a closed sales server 400 accessible only to specific members or an open sales server 410 openly accessible by a general search engine.

2 is a block diagram showing the configuration of a management server included in the system for automatically generating online product information according to a preferred embodiment of the present invention.

Referring to FIG. 2, the management server 100 includes a product search unit 110, a product information collection unit 120, a price calculation unit 130, an image selection unit 140, a web page generation unit 150, and A shopping mall posting unit 160 may be included.

On the other hand, the product DB 200 connected to the management server 100 includes a stocked product DB 210 in which a list of stocked products (eg, a product list illustrated in Table 1) is stored, and a product information collection unit 120 ) May include a product information DB 220 for storing the collected information. In addition, the product DB 200 may further include a dispute DB 230 for storing product or company information for which various disputes may be expected in relation to product sales.

The product search unit 110 performs a function of searching related products in the warehousing product DB 210 based on basic product information input from the manager terminal 10 . The product search unit 110 searches for related products in the warehousing product DB 210 using the product name or product number or a code displayed on the product (for example, a barcode or QR code). If there is only one product retrieved from the warehousing product DB 210 (for example, it refers to a product displayed in one row in Table 1, and does not mean that the quantity is one), select the corresponding product as a related product. Then, for the selected product, work for creating a web page for product information is progressed. When there are a plurality of products retrieved from the warehousing product DB 210, a specific product may be selected from among the plurality of products by using a priority order. The order of priority may be determined by the date of arrival of the product, the destination of the product, the price of the product, and the like. For example, a web page for product information may be created starting from the earliest date of arrival of the product. As another embodiment, when a plurality of products are searched for, the management server 100 may present the plurality of searched products to the manager terminal 10 and request selection of the manager terminal 10 .

The product information collection unit 120 collects product information by crawling the web server 300 connected to the network 500 based on basic product information. The product information collection unit 120 collects content existing on the web, and can perform a function of parsing the HTML page provided by the web server 300 and collecting information related to the product to be collected. there is. Product information collected here includes product image, price, sales quantity, and other web pages that provide product information (e.g., Naver (www.naver.com), Danawa (www.danawa.com), etc.) It may be a URL of a web page that provides comparison information, open market shopping malls such as G-Market and 11th Street, etc.), product advertisements, product information announcements in accordance with the 「Act on Provision of Information on Products, etc. in Electronic Commerce, etc.」, There are no restrictions on the information collected. Information collected by the product information collection unit 120 may be stored in the product information DB 220 .

The price calculation unit 130 calculates the selling price of the product based on the price information of the product. The price calculation unit 130 calculates any one of the standard selling price, the average product price, and the highest or lowest price of the product among the prices of the products collected by the product information collection unit 120 as the selling price of the product. can In one embodiment, the price calculation unit 130 may calculate the sales price of the product by adding a predetermined profit margin to the stock price of the product. In another embodiment, the price calculation unit 130 multiplies a predetermined discount rate by any one of the standard selling price, the average product price, and the highest or lowest price of the product to calculate the discounted price as the selling price of the product. can

The image selection unit 140 selects at least one image to be posted on a web page from among the images of products collected by the product information collection unit 120 . The image selector 140 may select an image by considering the source of the image, the characteristics of the image, or the advertisement efficiency according to the image. Also, the image selection unit 140 may be able to select a modified image by modifying the collected images.

The web page generator 150 creates a web page for product information using the collected product information. In one embodiment, the webpage generating unit 150 may create a webpage in HTML format.

3 shows an example of a web page generated by a web page generating unit of the system for automatically generating online product information according to a preferred embodiment of the present invention.

Referring to FIG. 3 , the web page 600 includes a product name unit 610 displaying product names, a product image unit 620 displaying a product image selected in the image selection unit 140, and a price calculation unit 130. Price display unit 630 for displaying the price information calculated on the product code unit 640 for displaying the product's barcode or item number, inventory quantity unit 650 for displaying the stock quantity of the product, and product description for displaying the product. It may include a description unit 660 and a product information provision notice unit 670 displaying the product information provision notification collected by the product information collection unit 120 .

The price display unit 630 may include a normal selling price of the product, a discount selling price that is an actual selling price and a discount rate, and shipping cost information.

In the product description unit 660 , descriptions of products collected by the product information collection unit 120 may be presented as text or images. An image presented in the product description unit 660 may also be selected in the image selection unit 140 . That is, the image selection unit 140 may select a plurality of images including a representative image of a product to be displayed on the product image unit 620 and a description image of the product to be displayed on the product description unit 660 . In addition, product advertisements collected by the product information collection unit 120 may be further included in the product description unit 660 .

In one embodiment, an external link 662 may be further included in the product description unit 660 . The product information collecting unit 120 may collect URLs of other web pages including product information, which may be displayed on the product description unit 660 as an external link 662 .

Referring back to FIG. 2 , the shopping mall posting unit 160 transmits the webpage automatically generated by the webpage generating unit 150 to the selling servers 400 and 410 to be posted. Here, the shopping mall posting unit 160 may select the sales servers 400 and 410 on which the webpage generated by the webpage generating unit 150 will be posted.

In one embodiment, the shopping mall posting unit 160 may select the sales servers 400 and 410 by referring to the dispute DB 230 . The dispute DB 230 may store information about a company that has requested prohibition of selling a specific product or information about a company with a possibility of a dispute. If the manufacturer or distributor of the product for which the web page is generated is included in the dispute DB 230, the shopping mall posting unit 160 causes the web page for the corresponding product to be posted only on the closed sales server 400. Otherwise, the shopping mall posting unit 160 causes the web page for the product to be posted not only on the closed sales server 400 but also on the open sales server 410 .

4 is a flowchart illustrating a method of automatically generating online product information according to a preferred embodiment of the present invention.

When there is a request for automatic generation of product information of a specific product from the manager terminal 10 to the management server 100, for example, when basic product information is input from the manager terminal 10, the product of the management server 100 The search unit 110 searches for the corresponding product in the warehousing product DB 210 (S10).

When the number of searches is singular (one) (S20), the product search unit 110 selects the corresponding product and completes the product selection (S40). When the number of searches is plural, the product search unit 110 determines the priority of product selection (S30), and selects a specific product according to the priority to complete the product selection (S40).

The product information collecting unit 120 collects product information on the selected product (S50). Collected product information may be stored in the product information DB 220 .

The image selector 140 selects at least one product image to be used (S60).

The webpage generation unit 150 determines information to be posted on the webpage (S70). In step S70, the price information calculated by the price calculation unit 130 may be determined as one of information to be posted on the web page. In addition, at least one of the product images selected by the image selector 140 may be determined as information to be posted on the web page. Also, at least one of the product information selected by the product information collecting unit 120 may be selected as information to be posted on the web page.

The webpage generation unit 150 creates a webpage using the determined webpage posting information (S80).

The shopping mall posting unit 160 selects the sales server 400 or 410 where the generated web page is to be posted and/or posts the generated web page to the sales server 400 or 410 (S90). In one embodiment, the shopping mall posting unit 160 may select the sales servers 400 and 410 to post the generated web page using the dispute DB 230 .

Next, detailed configuration and functions of the image selection unit 140 of the management server 100 according to the present invention will be further described.

5 is a block diagram showing the configuration of the image selection unit of the management server included in the system for automatically generating online product information according to a preferred embodiment of the present invention.

The image selection unit 140 may include an image source analysis unit 142 . The image source analysis unit 142 may check the source of the image collected by the product information collection unit 120 and determine whether the image is selected according to the image source. For example, if the source of the image is a blog or cafe operated by an individual, or a distribution company, the image may be excluded from selection. This is because there may be restrictions on the use of images, such as copyrights that prohibit arbitrary use. If it is analyzed that the image is provided by the manufacturer of the product, the source analysis unit 142 may determine that the image can be freely used for selling the product.

The image selection unit 140 may include an image characteristic analysis unit 144 . The image characteristic analysis unit 144 analyzes the characteristics of the image, such as the shooting angle of the image (eg, perspective view, front view, side view, etc.), whether the image includes a person or background other than the product itself, and the resolution of the image. can do. The image characteristic analysis unit 144 may determine whether an image is selected according to the characteristics of the image. For example, if a person is included in the image, the image characteristic analysis unit 144 may determine not to use the corresponding image for reasons such as portrait rights.

The image selection unit 140 may include an image correction unit 146 . The image correction unit 146 generates a modified image by modifying the image collected by the product information collection unit 120 . For example, when a person or an unnecessary background is included in the image collected by the product information collection unit 120, the image correction unit 146 may correct the image by extracting only the product image excluding the person or background. .

The image selector 140 may include an advertisement efficiency analyzer 148 . The advertisement efficiency analyzer 148 analyzes advertisement efficiency according to images, and the image selector 140 selects an image predicted by the advertisement efficiency analyzer 148 to have high advertisement efficiency.

In one embodiment, the advertisement efficiency analyzer 148 includes an item identification unit 148a for identifying items such as a specific product or prop in an image used in an advertisement, and an advertisement information learning unit 148b for learning advertisement information. and an advertisement efficiency prediction unit 148c that predicts advertisement efficiency.

Images collected by the product information collection unit 120 may be transmitted to the advertisement efficiency analysis unit 148 .

6 shows an example of an advertisement, (a) is an example of advertisement information in the form of an image, (b) is a diagram for explaining a background separation process of advertisement information, and (c) is a diagram from the background separated advertisement information. It is a diagram for explaining a process of extracting an advertisement target item.

Referring to (a) of FIG. 6 , an example of advertisement information (AD) in the form of an image can be confirmed.

Referring to (b) of FIG. 6 , the item identification unit 148a may perform an image separation process of separating the background BG of the advertisement information AD and the object OB. Here, as the image separation process, any one of a variety of known image separation algorithms may be selected and applied. In one embodiment, "Cho Hyun-jun, Kim David, Song Jae-bok, "Automatic object recognition and object segmentation learning dataset generation using Mask R-CNN," Journal of Korea Robotics Society (2019) 14(1):031-039 A background removal algorithm disclosed in " can be applied.

Referring to (c) of FIG. 6 , the item identification unit 148a may extract various items OB1, OB2, OB3, and OB4 from the object OB of the advertisement information AD. The item extraction unit 230 may perform an item extraction process for extracting items, such as the upper garment OB1, the bag OB2, the lower garment OB3, and the shoes OB4, if present in the object. Here, any one of various item extraction algorithms already known in the item extraction process may be selected and applied. In addition, the item identification unit 148a may identify an advertisement target item from among the extracted various items OB1, OB2, OB3, and OB4.

In one embodiment, the item identification unit 148a may identify an advertisement target item from among various item images using text information related to advertisement information. When the text information is 'wide pants', the item identification unit 148a may identify the bottom OB3 corresponding to the text information as an item to be advertised.

The advertisement information learning unit 148b has a function of learning advertisement efficiency from existing advertisement information collected from the advertisement platform of the web server 300, advertisement results related to the existing advertisement information (eg, advertisement click-through rate), and advertisement targets. do Such related information may be collected by the product information collection unit 120, or it may be possible to have a separate advertisement information collection unit (not shown) in the advertisement efficiency analysis unit 148.

The advertisement information learning unit 148b may generate an advertisement efficiency prediction model for calculating advertisement efficiency by learning the collected advertisement information. The advertisement information learning unit 148b may extract various analysis elements from the collected existing advertisement information. The advertisement information learning unit 148b accumulates the correlation between the characteristics of the extracted analysis elements and advertisement efficiency, organizes statistics, and generates an advertisement efficiency prediction model by performing modeling to match the linear prediction graph. Here, iterative deep learning and weights may be applied to the modeling.

7 is a graph showing an example of modeling for learning advertisement information.

Referring to FIG. 7 , PG represents a linear prediction graph, Y-axis represents advertisement efficiency, and X-axis represents characteristic values for each analysis element of advertisement information. Here, the characteristic value for each analysis element may be a characteristic value for one analysis element or a vector value (d1, d2, d3, d4, d5) obtained by vectorizing characteristic values for a plurality of analysis elements.

Based on the product information input from the user terminal 10, the advertisement information learning unit 148b performs deep learning on various analysis elements of advertisement information of related products to derive attribute values to be used for predicting advertisement efficiency. can

The advertisement information learning unit 148b may perform deep learning on first analysis elements such as background properties of advertisement information, properties of objects, advertisement target items, and character properties of objects.

The advertisement information learning unit 148b determines the color arrangement of the image in the advertisement information, the ratio of the advertisement target item to the entire image, and the vector elements (sharpness, color tone, contrast ratio, out-focusing ratio, sharpness, etc.) of the image in the advertisement information. Deep learning can be performed on the analysis element.

In addition, the advertisement information learning unit 148b may perform deep learning on a third analysis factor related to the atmosphere (refreshing, gloomy, cheerful, etc.) of the image in the advertisement information. Here, the mood of an image is an emotional part that a person recognizes when viewing a corresponding image, and can be statistically analyzed based on contrast and color tone. In the case of the first to third analysis elements, they may be higher or lower concepts of image elements used by the image processing unit 220 .

The advertisement information learning unit 148b may derive attribute analysis values for the first to third analysis elements.

When the deep learning for the first to third analysis elements is completed, the advertisement efficiency prediction unit 148c inputs the attribute analysis values for the first to third analysis elements into the pre-generated prediction efficiency model to determine the advertisement information. Advertising effectiveness can be predicted.

The advertisement efficiency prediction unit 148c may calculate an advertisement efficiency prediction value by applying weights to additional analysis factors.

Meanwhile, the advertisement efficiency prediction value (CRTRateSUM) according to an embodiment of the present invention may be calculated as in Equation 1 below.

In Equation 1 above, the CTR Rate may be calculated by Equation 2 below.

In Equations 1 and 2, m is the number of image attributes in the advertisement information, ImgPR.i is the advertising efficiency prediction value for the ith image attribute, TrendPR.i is the trend weight of the ith image attribute, and ImgDrat.i is the ith image attribute. Its own weight, ProPR.n is the weight for the item attribute itself, AdPR.n is the weight for the advertising media attribute, and BestPR.n is the matching weight for the popular item.

The trend weight of an image attribute is weighted according to the recent issue, search rank, sales rank, sociality rank, and seasonality of the image attribute. In the case of image attributes that meet the latest trend, a relatively high weight is assigned. .

The weight of the image property itself is weighted according to the importance of each image property. For example, a relatively high weight may be assigned to an object-related property rather than a background-related property.

The advertising media attribute weight is a weight assigned according to the advertising platform. A high weight may be assigned to an advertising platform recording a relatively high ad click-through rate.

The matching weight for a popular item is a weight assigned based on recent sales performance of the item to be advertised.

Various weights are related to their influence on advertising efficiency, and can be analyzed through deep learning or machine learning.

The above description is merely an example of the technical idea of the present invention, and those skilled in the art can make various modifications, changes, and substitutions without departing from the essential characteristics of the present invention. will be. Therefore, the embodiments disclosed in the present invention and the accompanying drawings are not intended to limit the technical idea of the present invention, but to explain, and the scope of the technical idea of the present invention is not limited by these embodiments and the accompanying drawings. . The protection scope of the present invention should be construed according to the claims below, and all technical ideas within the equivalent range should be construed as being included in the scope of the present invention.

10: Administrator terminal
100: management server
200: Product DB
300: web server
400, 410: sales server
500: network

Claims (16)

In the online product information automatic generation system for automatically generating product information to be posted online by receiving basic product information from a manager terminal,
It includes a management server to which the manager terminal accesses, and a product DB for storing information related to the product,
The management server,
a product information collection unit collecting product information by crawling a web server connected to a network based on the basic information;
a price calculation unit that calculates a sales price of the product based on price information among the product information;
an image selection unit selecting at least one image to be posted on a web page from among the product information collected by the product information collection unit; and
a web page generator for generating the web page for product information using the product information collected by the product information collection unit;
including,
The image selection unit,
An image source analysis unit for determining the possibility of copyright by checking the source of the image collected by the product information collection unit, and determining whether the image is selected according to the source; and
Whether or not to select the image is determined according to at least one characteristic of a photographing angle of the image, whether a person or a background is included in the image, or a resolution of the image, but when a person is included in the image, the use of the image Characterized in that it includes an image characteristic analysis unit that excludes or corrects the image,
Online product information automatic generation system. According to claim 1,
The product DB includes a warehousing product DB in which a list of warehousing products is stored and a product information DB in which the product information collected by the product information collection unit is stored;
The management server further includes a product search unit for searching for a related product in the stocked product DB based on basic information of the product input from the manager terminal. According to claim 2,
The system for automatically generating online product information, characterized in that the product search unit selects a specific product from a plurality of products by using a priority order when a plurality of products are searched in the stock product DB. According to claim 1,
The price calculation unit calculates the selling price by adding a profit or multiplying a discount rate to any one of a standard selling price, an average product price, a maximum price and a minimum price of the product among the prices of the product collected by the product information collection unit. Characterized in that, online product information automatic generation system. According to claim 1,
The image selection unit,
The system for automatically generating online product information, characterized in that it further comprises an advertisement efficiency analyzer for analyzing advertisement efficiency according to the image and selecting an image predicted to have high advertisement efficiency. According to claim 5,
The system of claim 1 , wherein the image selection unit further comprises an image correction unit generating a corrected image by modifying the image collected by the product information collection unit. According to any one of claims 1 to 6,
The management server may further include a shopping mall posting unit that selects a sales server to post the web page or posts the web page to the sales server. According to claim 7,
The product DB further includes a dispute DB for storing dispute prediction information related to the sale of the product;
The shopping mall posting department,
If a dispute is expected in the sale of the product by referring to the dispute DB, the web page is posted only to a closed sales server with limited access among the sales servers,
Characterized in that, if a dispute is not expected in the sale of the product by referring to the dispute DB, the web page is posted on both a closed sales server with restricted access and an open sales server with unrestricted access among the sales servers.
Online product information automatic generation system. An automatic online product information generation method in which a management server receives basic product information from an administrator terminal and automatically generates product information to be posted online,
(a) searching for and selecting the product from a warehousing product DB by the product search unit of the management server;
(b) collecting product information on a selected product by crawling a web server connected to a network by the product information collection unit of the management server;
(c) selecting at least one product image to be used as the product information to be posted by the image selection unit of the management server;
(d) determining information to be posted on the web page by the web page creation unit of the management server; and
(e) generating the web page by the web page generating unit;
including,
In step (c),
determining, by the image selection unit, the source of the image collected by the product information collection unit to determine the possibility of copyright restriction, and determining whether to select the image according to the source; and
The image selection unit determines whether or not to select the image according to at least one characteristic of a photographing angle of the image, whether a person or a background is included in the image, or a resolution of the image, but the person is included in the image characterized in that it includes the step of excluding the use of the image or modifying the image,
How to automatically generate online product information. According to claim 9,
In step (a), if there are a plurality of products retrieved from the stocked product DB, the product search unit selects a specific product from among the plurality of products using a priority order. According to claim 9,
In the step (d), the price calculation unit of the management server sets a profit on any one of the standard selling price, average product price, maximum price or minimum price of the product among the prices of the product collected by the product information collection unit. A method of automatically generating online product information, characterized in that the sales price is calculated by adding or multiplying the discount rate. According to claim 9,
In step (c),
The method of automatically generating online product information, characterized in that, further comprising the step of selecting, by the image selector, an image predicted to have high advertising efficiency by analyzing advertising efficiency according to the image. According to claim 9,
The web page generated in step (e),
A product name unit displaying the name of the product, a product image unit displaying the product image selected in the image selection unit, a price display unit displaying price information of the product, and a product code unit displaying the barcode or product number of the product. characterized in that it comprises a stock quantity unit displaying the inventory quantity of the product, a product description unit displaying a description of the product, and a product information provision notification unit displaying product information provision notification collected by the product information collection unit. How to automatically generate online product information. According to claim 13,
The method of automatically generating online product information, characterized in that the product information collection unit collects URL information of another web page for the product, and the web page further includes an external link according to the URL information. According to any one of claims 9 to 14,
(f) selecting, by the shopping mall posting unit of the management server, a sales server to post the web page or posting the web page to the sales server; According to claim 15,
In step (f),
The shopping mall posting department,
If a dispute is expected in the sale of the product by referring to the dispute DB, the web page is posted only to a closed sales server with limited access among the sales servers,
Characterized in that, if a dispute is not expected in the sale of the product by referring to the dispute DB, the web page is posted on both a closed sales server with restricted access and an open sales server with unrestricted access among the sales servers.
How to automatically generate online product information.

Images