KR102654932B1 - Method for providing advertisement effect analysis service and server using the same - Google Patents

Abstract

The server according to an embodiment of the present invention is a server that provides an advertising effect analysis service, and uses video data provided from at least one camera installed around the billboard to convert at least a portion of the entire area of the image into an analysis area. and a fluidity analysis unit that sets at least some of the floating objects as analysis targets and analyzes the fluidity and density of the analysis targets within the analysis area, and uses video data to provide information on the number and gender of users around the billboard. , and includes an advertising effect analysis unit that extracts age information, gaze information, and appearance information, and analyzes the advertising effect of the advertisement displayed on the billboard based on this.

Description

Method of providing advertising effect analysis service and server using the same {METHOD FOR PROVIDING ADVERTISEMENT EFFECT ANALYSIS SERVICE AND SERVER USING THE SAME}

The present invention relates to a method of providing an advertising effect analysis service and a server using the same. More specifically, a method of providing an advertising effect analysis service by analyzing video data provided from at least one camera installed around a billboard and the same. It's about the server used.

As an advertising technique, there is an advertising technique that delivers advertisements to users using fixed billboards installed outdoors, in public places, markets, apartments, bus stop shelters, etc., or mobile billboards mounted on various means of transportation such as buses and taxis. Advertising techniques such as these are continuously growing due to the steady demand for daily contact media.

However, due to the nature of such advertising techniques, it is difficult to accurately measure advertising effectiveness, so advertising fees paid to advertisers are customarily set based on advertising location, frequency of advertising exposure, etc., or accuracy is low due to manual measurement of floating population. It is set through a falling method. Moreover, it is difficult to accurately reflect the advertising effect according to the floating population's attention to advertising in the advertising cost setting.

The problem to be solved by the present invention is to provide a method of providing an advertising effect analysis service that can set a reasonable advertising cost by accurately analyzing the advertising effect, and a server using the same.

Another problem to be solved by the present invention is to provide a method for providing an advertising effect analysis service that can calculate a more appropriate advertising effect in response to the target audience of the advertisement depending on the type of advertisement output from the billboard, and to provide a server using the same. .

Another problem to be solved by the present invention is a method of providing an advertising effect analysis service that can measure the liquidity and density of moving objects (for example, people, cars, two-wheeled vehicles, etc.) in a specific area, and a server using the same. It is done.

The problems of the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

In order to solve the problems described above, the server according to embodiments of the present invention provides an advertising effect analysis service, using image data provided from at least one camera that captures images around the billboard. , a fluidity analysis unit that sets at least some of the entire area of the image as an analysis area and sets at least some of the moving objects as an analysis target to analyze the fluidity and density of the analysis target in the analysis area; And using the video data, the number of users around the billboard, gender information, age information, gaze information, and appearance information are extracted, and based on this, the advertising effect of the advertisement output from the billboard is analyzed. It may include an analysis unit.

In order to solve the problems described above, the method according to embodiments of the present invention provides an advertising effect analysis service, and includes receiving image data provided from at least one camera that captures images around the billboard. Step, using the image data, set at least some of the entire area of the image as an analysis area, set at least some of the moving objects as the analysis target, and determine the liquidity and Analyzing density, and using the image data, extracting the number of users around the billboard, gender information, age information, gaze information, and appearance information, and based on this, advertising for the advertisement output from the billboard It may include a step of analyzing the effect.

In order to solve the problems described above, the computer program according to the embodiments of the present invention is a computer program stored in a computer-readable medium, where the computer program includes instructions, and when executed by a computer system, the instructions can cause the computer system to perform the method.

In the case of a method for providing an advertising effect analysis service according to embodiments of the present invention and a server using the same, the advertising effect can be analyzed by accurately reflecting the interest of actual users in the advertisement output from the billboard, and based on this, A reasonable advertising cost can be set.

In addition, in the case of a method for providing an advertising effect analysis service according to embodiments of the present invention and a server using the same, an image of an analysis object (moving object) corresponding to the target group of the advertisement is displayed depending on the type of advertisement output from the billboard. By performing analysis and analyzing the advertising effect, it is possible to calculate an advertising effect more suitable for the relevant advertisement.

In addition, in the case of a method for providing an advertising effect analysis service and a server using the same according to embodiments of the present invention, not only the advertising cost setting function described above, but also moving objects in a specific area (for example, people, cars, two-wheeled vehicles, etc.) It can provide a function to measure the liquidity and density of.

The effects according to the present invention are not limited to the details exemplified above, and further various effects are included within the present invention.

1 is a diagram schematically showing an advertising effect analysis system according to embodiments of the present invention.
Figure 2 is a diagram schematically showing an advertising effect analysis server according to embodiments of the present invention.
Figure 3 is a block diagram showing an advertising effect analysis server according to embodiments of the present invention.
FIG. 4 is a block diagram showing an example of a liquidity analysis unit included in the advertising effect analysis server of FIG. 3.
FIG. 5 is a diagram for explaining an example of the operation of the liquidity analysis unit of FIG. 4.
FIG. 6 is a diagram for explaining an example of the operation of the target setting unit included in the liquidity analysis unit of FIG. 4.
FIG. 7 is a diagram for explaining an example of the operation of the area setting unit included in the liquidity analysis unit of FIG. 4.
FIG. 8 is a block diagram showing an example of an advertising effect analysis unit included in the advertising effect analysis server of FIG. 3.
FIG. 9 is a block diagram showing an example of a second analysis unit included in the advertising effect analysis unit of FIG. 8.
FIG. 10 is a diagram for explaining an example of the operation of the gaze detection unit included in the second analysis unit of FIG. 9.
FIG. 11 is a diagram for explaining an example of the operation of the characteristic detection unit included in the second analysis unit of FIG. 9.
FIG. 12 is a diagram for explaining another example of the operation of the characteristic detection unit included in the second analysis unit of FIG. 9.
FIG. 13 is a flow chart to explain an example of the operation of the model learning unit included in the advertising effect analysis unit of FIG. 8.
FIG. 14 is a block diagram showing another example of an advertising effect analysis unit included in the advertising effect analysis server of FIG. 3.
Figures 15A and 15B are diagrams for explaining an example of communication between an advertising effect analysis server and a camera according to embodiments of the present invention.
Figures 16a and 16b are diagrams for explaining another example of communication between an advertising effect analysis server and a camera according to embodiments of the present invention.

While describing each drawing, similar reference numerals are used for similar components. In the attached drawings, the dimensions of the structures are enlarged from the actual size for clarity of the present invention. Terms such as first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, a first component may be named a second component, and similarly, the second component may also be named a first component without departing from the scope of the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise.

In this application, terms such as “comprise” or “have” are intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but are not intended to indicate the presence of one or more other features. It should be understood that this does not exclude in advance the possibility of the existence or addition of elements, numbers, steps, operations, components, parts, or combinations thereof.

Additionally, when a part is said to be “connected” to another part, this includes not only the case where it is directly connected, but also the case where it is connected with another element in between.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the attached drawings. The same reference numerals are used for the same components in the drawings, and duplicate descriptions for the same components are omitted.

1 is a diagram schematically showing an advertising effect analysis system according to embodiments of the present invention.

Referring to FIG. 1, the advertising effect analysis system 1 according to embodiments of the present invention may include a server 100 (or an advertising effect analysis server, an advertising effect analysis service providing server). Depending on the embodiment, the advertising effect analysis system 1 may further include at least one camera 200, a billboard 300, and an advertiser terminal 400.

The server 100 may operate an advertising brokerage platform between the billboard 300 and the advertiser terminal 400. For example, the server 100 may operate an advertising brokerage platform by performing network communication between the billboard 300 and the advertiser terminal 400. For example, the server 100 may communicate with the advertiser terminal 400 to receive an advertisement request from the advertiser, allocate an advertisement to the billboard 300, and allow a predetermined advertisement to be output from the billboard 300. there is.

The billboard 300 may be a printed billboard equipped with printed materials or a display device that outputs advertisements in video form. The billboard 300 performs the function of continuously showing one print to users, or showing various prints or images to users in a preset order, according to a preset period, or in a random manner. You can do it. That is, in the following description, the billboard 300 refers to a billboard that is currently commonly used offline, and may refer to a billboard that simply displays printed matter or images provided by advertisers to users. Meanwhile, when the billboard 300 is a display device, the billboard 300 may provide various information in addition to advertisements.

Networks include wired networks such as Local Area Network (LAN), Wide Area Network (WAN), or Value Added Network (VAN), mobile radio communication networks, satellite communication networks, and Bluetooth (Bluetooth). ), Wibro (Wireless Broadband Internet), HSDPA (High Speed Downlink Packet Access), LTE (Long Term Evolution), 5G (5th Generation Mobile Telecommunication), etc., but are not limited to this. .

Meanwhile, the advertiser terminal 400 is a smartphone, tablet PC (personal computer), mobile phone, video phone, e-book reader, desktop PC, laptop PC, netbook computer, and workstation. It may be, but is not limited to, an electronic device such as a workstation, server, personal digital assistant (PDA), media box, game console, electronic dictionary, or wearable device, and may be connected to a server via a network. Any device that can transmit and receive data with 100 can be adopted as the advertiser terminal 400 regardless of its type.

In one embodiment, the server 100 may receive data about the surrounding image of the billboard 300 through communication with the camera 200 and analyze the advertising effect of the advertisement output from the billboard 300 based on this. there is.

For example, the camera 200 may be installed around the billboard 300. For example, when the billboard 300 is composed of an outdoor (Out of Home (OOH) billboard, etc.), the camera 200 is installed around the billboard 300 and captures images around the billboard 300. circuit Television; CCTV), but is not limited thereto. As another example, the camera 200 may be installed on the billboard 300. For example, when the billboard 300 is composed of a bus stop shelter billboard, etc., the camera 200 may be installed on the structure that constitutes the billboard 300. However, the camera 200 is not limited to this, and may be configured in various ways as long as it is capable of capturing images around the billboard 300.

The server 100 may perform communication with the camera 200 and receive data (image data) about the image captured around the billboard 300 from the camera 200. The server 100 extracts the number of users around the billboard 300, gender information, age information, gaze information, appearance information, etc. by analyzing the video data, and based on this, provides information about the advertisement output from the billboard 300. You can analyze advertising effectiveness according to users' attention to advertising.

Additionally, in one embodiment, the server 100 analyzes the image data received through communication with the camera 200 to identify moving objects (e.g. For example, the liquidity of people, cars, two-wheeled vehicles, etc.) can be analyzed.

In this way, the advertising effect analysis system 1 (or server 100) according to embodiments of the present invention receives image data from the camera 200 that captures the surroundings of the billboard 300, and analyzes the image data based on the image data. Thus, it is possible to analyze not only the advertising effect according to the user's attention to the advertisement output from the billboard 300, but also the fluidity of objects around the billboard 300.

For a detailed description of this, FIGS. 2 to 16B may be referred to.

Figure 2 is a diagram schematically showing an advertising effect analysis server according to embodiments of the present invention.

Referring to Figures 1 and 2, the server 100 (hereinafter referred to as “advertising effect analysis server”) according to embodiments of the present invention includes a communication unit 110, a processor 120, and a database 130. can do. Meanwhile, in FIG. 2, the advertising effect analysis server 100 is shown as including a database 130, but this is merely illustrative and the embodiment of the present invention is not limited thereto. For example, depending on the embodiment, the advertising effect analysis server 100 may be configured without including the database 130.

The advertising effect analysis server 100 can transmit and receive information to the camera 200 and the like through the communication unit 110. To this end, the communication unit 110 may be implemented with various communication technologies. For example, the communication unit 110 includes Wi-Fi, Wideband CDMA (WCDMA), High Speed Downlink Packet Access (HSDPA), High Speed Uplink Packet Access (HSUPA), High Speed Packet Access (HSPA), and Mobile WiMAX. WiMAX), WiBro, LTE (Long Term Evolution), 5G, Bluetooth, IrDA (infrared data association), NFC (Near Field Communication), Zigbee, and wireless LAN technologies will be applied. You can. In addition, when the advertising effect analysis server 100 is connected to the Internet and provides the corresponding service, the communication unit 110 may follow TCP/IP, a standard protocol for information transmission on the Internet, and GPS (Global Positioning System) technology. You can also use .

The database 130 may store information for advertising effect analysis services. For example, the database 130 may store information about liquidity data analyzed by the advertising effect analysis server 100. As another example, the database 130 may store information about data for AI model training conducted through image analysis of image data provided from the camera 200.

The processor 120 can access the database 130 through the communication unit 110. Meanwhile, when the advertising effect analysis server 100 does not include the database 130 and uses an external database as described above, the advertising effect analysis server 100 may access the external database through the communication unit 110. You will be able to.

In addition, the processor 120 included in the advertising effect analysis server 100 may use the camera 200, etc. (e.g., not only the camera 200 but also the billboard 300 described with reference to FIG. 1, etc.) through the communication unit 110. Communication can be performed with the advertiser terminal (400, etc.).

Hereinafter, with reference to FIGS. 3 to 16B, operations and processes performed on the processor 120 of the advertising effect analysis server 100 according to embodiments of the present invention will be described in more detail.

Figure 3 is a block diagram showing an advertising effect analysis server according to embodiments of the present invention.

FIG. 4 is a block diagram showing an example of a liquidity analysis unit included in the advertising effect analysis server of FIG. 3.

FIG. 5 is a diagram for explaining an example of the operation of the liquidity analysis unit of FIG. 4.

*FIG. 6 is a diagram for explaining an example of the operation of the target setting unit included in the liquidity analysis unit of FIG. 4.

FIG. 7 is a diagram for explaining an example of the operation of the area setting unit included in the liquidity analysis unit of FIG. 4.

FIG. 8 is a block diagram showing an example of an advertising effect analysis unit included in the advertising effect analysis server of FIG. 3.

FIG. 9 is a block diagram showing an example of a second analysis unit included in the advertising effect analysis unit of FIG. 8.

FIG. 10 is a diagram for explaining an example of the operation of the gaze detection unit included in the second analysis unit of FIG. 9.

FIG. 11 is a diagram for explaining an example of the operation of the characteristic detection unit included in the second analysis unit of FIG. 9.

FIG. 12 is a diagram for explaining another example of the operation of the characteristic detection unit included in the second analysis unit of FIG. 9.

FIG. 13 is a flow chart to explain an example of the operation of the model learning unit included in the advertising effect analysis unit of FIG. 8.

Referring to Figures 1 to 3, the advertising effect analysis server 100 according to embodiments of the present invention may receive video data (VD) from the camera 200. Meanwhile, as described with reference to FIG. 1, the video data VD may include image (eg, still image or video) information captured around the billboard 300.

In one embodiment, the advertising effect analysis server 100 extracts the number, gender information, age information, gaze information, appearance information, etc. of users around the billboard 300 based on the video data (VD), and based on this, The advertisement effect is analyzed according to the user's attention to the advertisement output from the billboard 300, and the effect of the advertisement is analyzed according to the user's level of attention to the advertisement, and the moving object (e.g., person, car, two-wheeled vehicle, etc.) Liquidity can be analyzed.

To this end, the advertising effect analysis server 100 may include a liquidity analysis unit 160 and an advertising effect analysis unit 170. Depending on the embodiment, the advertising effect analysis server 100 may further include a liquidity data storage unit 180 and an advertising cost calculation unit 190.

The liquidity analysis unit 160 may receive image data VD from the camera 200 . Here, the video data VD may include information about an image (for example, a still image or video) captured around the billboard 300 described with reference to FIG. 1 .

Based on the image data VD, the fluidity analysis unit 160 analyzes the fluidity of a specific moving object (e.g., a person, a car, a two-wheeled vehicle, etc.) for at least a portion of the entire area of the captured image. , first analysis data (AD1) can be generated.

In one embodiment, the fluidity analysis unit 160 uses Vision AI technology based on an image recognition algorithm to set a specific region among the entire region of the captured image and detect fluid objects such as people, cars, and two-wheeled vehicles. By selecting at least one moving object (or analysis object) among the objects, the fluidity and density of the analysis object within the specific area can be measured. For example, a video analytics solution (VAS) is installed in the liquidity analysis unit 160 to analyze captured image information included in video data (VD) based on artificial intelligence, and to analyze the analysis target within the specific area. The fluidity and density can be measured.

As an example, further referring to FIG. 5, the liquidity analysis unit 160 analyzes the captured image (e.g., the image displayed in the area indicated by “A” in FIG. 5) included in the image data VD, and , by measuring the fluidity and density of the analysis target within the specific area, the area of the specific area selected in the captured image, the number of moving objects (e.g., people, etc.) located within the specific area, and the fluidity and density. Information can be extracted. For example, the liquidity analysis unit 160 extracts the area of the specific area as "825 m ² ", as shown in the area indicated by "B" in FIG. 5, and calculates the number of people in the specific area as " 824 people" can be extracted, and the density can be extracted as "0.99 people/m ² ".

To describe the specific configuration of the liquidity analysis unit 160 in more detail, referring further to FIG. 4, the liquidity analysis unit 160 includes a target setting unit (TSP), an area setting unit (ASP), and a first analysis unit. It may include part (AZP1).

The target setting unit (TSP) receives image data (VD), sets at least one moving object among the moving objects in the captured image as an analysis target, and generates target setting data (TSD) containing information about this. can do. For example, target setting data (TSD) may include unique identifier information assigned to the analysis target.

For example, further referring to FIG. 6 , a plurality of moving objects FO may be included in a captured image included in the image data VD. For example, the captured image shown in FIG. 6 may include moving objects (FO) of a car and a person (however, this is based on the captured image of FIG. 6 and may vary depending on the area, time, etc. where the image was captured). Accordingly, various floating objects (FO) such as two-wheeled vehicles, animals, etc. may be included).

First, the target setting unit (TSP) can detect the types of a plurality of moving objects (FO) in the captured image. For example, the target setting unit (TSP) may extract a plurality of moving objects (FO) from a captured image and then detect which moving object each of the plurality of moving objects (FO) is. For example, as shown in FIG. 6, the target setting unit (TSP) determines that one of the plurality of moving objects (FO) in the captured image is a car (shown as “car” in FIG. 6). It can be detected that the remaining two floating objects (FO) are composed of people (shown as “Person” in FIG. 6).

Thereafter, the target setting unit (TSP) may select at least one moving object (FO) to be set as the analysis target (TG) among the plurality of moving objects (FO) of which the type is specified. For example, the target setting unit (TSP) may select a person as the analysis target (TG) among a plurality of moving objects (FO), but this is simply an example and the embodiment of the present invention is not limited thereto. For example, the target setting unit (TSP) may select a car as an analysis target (TG) among a plurality of moving objects (FO), or select both a car and a person as an analysis target (TG).

In one embodiment, the moving object selected as the analysis target (TG) by the target setting unit (TSP) may be a preset one. For example, if it is set in advance to perform fluidity and density analysis only for a specific analysis target, the target setting unit (TSP) extracts a plurality of floating objects (FO) and their types from the captured image, and Only the corresponding analysis target (TG) can be extracted from the floating objects (FO).

However, this is merely illustrative, and the embodiments of the present invention are not limited thereto. For example, the target setting unit (TSP) may extract only the corresponding analysis target (TG) from a plurality of moving objects (FO) based on a specific analysis target entered by the user's input.

The area setting unit (ASP) may receive image data (VD), set at least a portion of the entire area in the captured image as an analysis area, and generate area setting data (ASD) containing information about this. .

For example, further referring to FIG. 7 , a portion of the entire area of the captured image included in the image data VD may be set as the analysis area AA.

In one embodiment, the area selected as the analysis area (AA) by the area setting unit (ASP) may be preset. For example, if it is set in advance to perform fluidity and density analysis targeting only a specific area, the area setting unit (ASP) can extract only the specific area from the entire area of the captured image as the analysis area (AA).

However, this is merely illustrative, and the embodiments of the present invention are not limited thereto. For example, the area setting unit (ASP) may extract a specific input area from the entire area of the captured image as an analysis area (AA), based on the specific area input by the user.

Referring again to FIG. 4, the first analysis unit (AZP1) analyzes the image based on the target setting data (TSD) and area setting data (ASD) provided from the target setting unit (TSP) and the area setting unit (ASP). By detecting the number of analysis objects (TG) present in the analysis area (AA) of the captured image included in the data (VD), the fluidity and density of the analysis objects (TG) within the analysis area (AA) are analyzed. , Based on this, first analysis data AD1 can be generated.

Depending on the embodiment, the first analysis unit AZP1 (or the liquidity analysis unit 160) determines the liquidity and density of the analysis target (TG) in the analysis area (AA) for the video based on the video data (VD). By measuring in real time, first analysis data (AD1) can be generated. However, the embodiment of the present invention is not limited to this, and the first analysis unit AZP1 (or the liquidity analysis unit 160) generates still images based on the image data VD, that is, image data for a specific point in time. Based on (VD), the first analysis data (AD1) may be generated by measuring the fluidity and density of the analysis target (TG) in the analysis area (AA).

Referring again to FIG. 3, the first analysis data AD1 generated by the liquidity analysis unit 160 may be provided to the liquidity data storage unit 180, and the liquidity data storage unit 180 may store it. .

In one embodiment, the advertising effect analysis server 100 according to embodiments of the present invention stores the first analysis data (AD1) stored in the liquidity data storage unit 180, that is, the analysis target (TG) in the analysis area (AA). ), automatically informs the user (person in charge) of the location of the analysis area (AA) and the liquidity and density of the analysis target (TG) at specified stages such as interest, caution, warning, and risk, etc. It can perform its function. For example, when the advertising effect analysis server 100 reaches a standard set based on the first analysis data (AD1), it may perform a function of automatically notifying the user (related organization) of the relevant content. Accordingly, the multiple density of a specific area set as the analysis area (AA) can be predicted in real time, and safety accidents can be detected and prevented.

Meanwhile, the liquidity data storage unit 180 may be configured as part of the database 130 described with reference to FIG. 2, but the embodiment of the present invention is not limited thereto.

Meanwhile, at least some of the functions of the liquidity analysis unit 160 described with reference to FIGS. 4 to 7 can also be performed within the advertising effect analysis unit 170, which will be described in more detail with reference to FIG. 14. .

The advertising effect analysis unit 170 may receive video data VD from the camera 200.

The advertising effect analysis unit 170 extracts the number of users around the billboard 300, gender information, age information, gaze information, appearance information, etc. based on the video data (VD), and based on this, the billboard 300 Second analysis data (AD2) can be generated by analyzing users' advertisement attention (or advertisement effect) for advertisements output from .

In one embodiment, the advertising effect analysis unit 170 learns an AI model using model learning data (AMD) built on images captured around the billboard 300 (e.g., still images or videos). You can analyze captured images through this. For example, the advertising effect analysis unit 170 may use a convolutional neural network (CNN), a recurrent neural network (RNN), or an artificial neural network combining CNN and long short-term memory (LSTM) RNN, and a random forest ( You can also use a Random Forest (Random Forest) classifier. Depending on the embodiment, the advertising effect analysis unit 170 uses a cascade regression forest and a fast radial symmetry transform (FRST) to minimize the impact of changes in lighting to accurately collect data from the camera 200. Videos can be analyzed.

To describe the specific configuration of the advertising effect analysis unit 170 in more detail, referring further to FIG. 8, the advertising effect analysis unit 170 may include a second analysis unit AZP2. Depending on the embodiment, the advertising effect analysis unit 170 may further include a model learning unit (MLP) and a memory (MEM).

The second analysis unit (AZP2) analyzes the captured image around the billboard 300 using the model learning data (AMD) provided from the memory (MEM) based on the video data (VD), and analyzes the billboard (300). Information on the number, gender, age, appearance, etc. of users passing by, information on users looking at advertisements displayed on the billboard 300, etc. can be extracted. Here, the appearance information may include various external features such as tops, bottoms, shoes, and accessories in addition to the facial features of users passing around the billboard 300 or watching advertisements output from the billboard 300. , Information about users looking at advertisements output from the billboard 300 may include gaze information about users.

For example, further referring to FIG. 9 , the second analysis unit AZP2 may include a gaze detection unit (GDP), a characteristic detection unit (CDP), and an analysis data generation unit (ADG).

The gaze detection unit (GDP) uses model learning data (AMD) to analyze video data (VD) to determine the gaze of users looking at the billboard 300 in the captured image, that is, the video captured around the billboard 300. Information (GD) can be analyzed and generated.

For example, further referring to FIG. 10, the gaze detection unit (GDP) collects gaze information (GD1, GD2, GD3, GD4) can be extracted. For example, as shown in Figure 10, the gaze detection unit (GDP) extracts first gaze information (GD1) that the gaze of the first analysis object (TG1) is pointing from the center to the upper left, and extracts the second analysis object ( Extracting second gaze information (GD2) in which the gaze of TG2) is directed from the center to the bottom, extracting third gaze information (GD3) in which the gaze of the third object of analysis (TG3) is directed from the center to the bottom right, Fourth gaze information (GD4) where the gaze of the fourth analysis target (TG4) stays at the center can be extracted.

Meanwhile, as described later, based on the gaze information (GD) of users looking at the billboard 300, the analysis data generation unit (ADG) included in the second analysis unit (AZP2) of the advertising effect analysis unit 170 ) refers to whether users located around the billboard 300 are actually watching the advertisement output from the billboard 300, the time they actually look at the advertisement output from the billboard 300, and the advertisement displayed from the billboard 300. It is possible to extract which content of the video the user is interested in, and based on this, the advertising cost calculation unit 190 can reflect this in calculating advertising costs.

On the other hand, when the billboard 300 is composed of a bus stop, shelter billboard, etc., the camera 200 that generates video data (VD) can be installed on the structure constituting the billboard 300, making it possible to secure a short-distance video. Thus, more accurate detection of gaze information (GD) may be possible.

Next, the feature detection unit (CDP) analyzes the video data (VD) using the model learning data (AMD) to identify users around the billboard 300 in the captured image, that is, the video captured around the billboard 300. Alternatively, characteristic information (CD) of users looking at advertisements output from the billboard 300 may be analyzed and generated.

In one embodiment, the characteristic information (CD) analyzed and generated by the characteristic detection unit (CDP) includes the gender, age, appearance, etc. of users around the billboard 300 or users looking at advertisements output from the billboard 300. It may contain information about

For example, further referring to FIG. 11, the characteristic detection unit (CDP) is a plurality of analysis objects (TG5, TG6), for example, characteristic information (CD1, CD2) for each of the advertising viewing users, and the analysis object The gender and age information of people (TG5, TG6) can be detected. For example, as shown in FIG. 11, the characteristic detection unit (CDP) detects the first characteristic information (CD1) as 'male' and '32 years old (Age: 32)' for the fifth analysis target (TG5). ), or the second characteristic information (CD2) indicating that the sixth analysis target (TG6) is 'Female' and 'Age: 35' can be detected.

In addition, in order to specifically explain the information on the appearance of users included in the above-described characteristic information (CD), further referring to FIG. 12, the characteristic detection unit (CDP) generates model learning data (AMD) for the image data (VD). ) can be used to distinguish analysis targets (for example, users) within the video data (VD). For example, the feature detection unit (CDP) is included in the model training data (AMD) and uses the Instance Segmentation Model, which performs object detection, to individually identify analysis targets (e.g., users) in the video. can be distinguished. In addition, the feature detection unit (CDP) extracts clothing information such as tops, bottoms, shoes, and accessories as appearance information using model learning data (AMD) for each separate analysis object (e.g., users). In addition, the color model included in the model training data (AMD) can be used to extract information about the color of the clothing information to analyze and generate characteristic information (for example, third characteristic information (CD3)). You can.

Meanwhile, as will be described later, a second analysis by the advertising effect analysis unit 170 is performed based on characteristic information (CD) of users around the billboard 300 or users looking at advertisements output from the billboard 300. The analysis data generation unit (ADG) included in the unit (AZP2) can analyze the advertising effect according to users' gender, age, and appearance information, and based on this, the advertising cost calculation unit 190 can reflect this in calculating advertising costs. You can.

Referring again to FIG. 9 , the analysis data generator (ADG) may analyze the advertising effect based on gaze information (GD) and characteristic information (CD) and generate second analysis data (AD2) accordingly.

For example, as described above, the analysis data generator (ADG) determines whether users located around the billboard 300 are actually watching advertisements output from the billboard 300 based on the gaze information (GD). , using information related to advertising effects, such as the time the user is actually looking at the advertisement output from the billboard 300, and whether the user is interested in what content of the advertising video output from the billboard 300, etc. The effect of advertisements output from (300) can be analyzed. For example, when users located around the billboard 300 actually watch advertisements output from the billboard 300 or actually look at the displayed advertisements for a relatively long time, the analysis data generator (ADG) It can be analyzed that the effect of advertising is relatively large. As another example, depending on which content of the advertising video output from the billboard 300 the user is interested in, the analysis data generator (ADG) may analyze that the advertising effect for the specific content of the advertisement is large. there is.

Additionally, as another example, as described above, the analysis data generator (ADG) may analyze the advertisement effect according to the gender, age, and appearance information of users watching the advertisement based on the characteristic information (CD). For example, if the advertisement output from the billboard 300 is a fashion advertisement, the effectiveness of the advertisement can be analyzed by analyzing whether it matches the fashion advertisement according to the gender, age, and appearance information of the users watching the advertisement. there is. Specifically, when a fashion advertisement output from the billboard 300 advertises a top of a specific color (for example, a jacket, etc.), the analysis data generator (ADG) has appearance information similar to the top of the specific color. It can be analyzed that the greater the number of users, the greater the advertising effect. As another example, if the advertisement output from the billboard 300 is an advertisement for health food, the analysis data generator (ADG) determines that the gender and age of the users watching the advertisement match the gender and age recommended for the health food. By analyzing whether or not the advertisement is being used, the advertising effectiveness of the advertisement can be analyzed.

In this way, the analysis data generator (ADG) analyzes the effect of the advertisement by analyzing the impact of the advertisement output from the billboard 300 on actual users based on gaze information (GD) and characteristic information (CD), Based on this, second analysis data (AD2) can be generated. For example, the second analysis data AD2 may include, but is not limited to, a numerical value for the advertising effectiveness index.

Referring to FIG. 8, as described above, model training data (AMD) may be provided from memory (MEM). Here, the model learning data (AMD) may be generated by the model learning unit (MLP) performing a labeling operation on the image data (VD) provided from the camera 200.

For example, further referring to FIG. 13, the model learning unit (MLP) may acquire raw data (S1010). Here, the raw data may correspond to image data (VD) provided from the camera 200.

Next, the model learning unit (MLP) may construct learning data by performing a labeling operation on the acquired raw data (S1020). Here, a labeling operation on the acquired raw data is performed to construct learning data (S1020), which involves at least one of the above-described gender, age, and appearance for the image data (VD) provided from the above-mentioned camera 200. Accordingly, the process may be divided into identical or similar classes and the work of constructing model training data (AMD) may be carried out accordingly. For example, the model learning part (MLP) can build model learning data (AMD) by continuously analyzing video data (VD) (raw data) using artificial intelligence and machine learning. . Meanwhile, the model learning data (AMD) constructed by the model learning unit (MLP) may be stored on the memory (MEM).

Finally, the second analysis unit (AZP2) may proceed with AI model learning (S1030) using the constructed learning data. Here, the AI model learning using the constructed learning data (S1030) is performed by the above-described second analysis unit (AZP2) using the model learning data (AMD) stored in the memory (MEM) to display the billboard 300. ) It may be analyzing images captured from surrounding images (for example, still images or moving images).

Referring again to FIG. 3, the advertising cost calculation unit 190 may receive the second analysis data AD2 from the advertising effect analysis unit 170. Here, the second analysis data AD2 includes information about the advertisement effect (for example, an advertisement effect index), and the advertisement cost calculation unit 190 can reflect this advertisement effect in the advertisement cost calculation. Additionally, the advertising cost calculated by the advertising cost calculation unit 190 may be provided to the advertiser.

Accordingly, in the advertising effect analysis server 100 according to embodiments of the present invention, the advertising effect analysis server 100 does not calculate the advertising effect through conventionally calculated or inaccurate methods such as manual measurement. ) includes an advertising effect analysis unit 170 and an advertising cost calculation cost 190, so that the advertising effect can be analyzed by accurately reflecting the interest of actual users in the advertising output from the billboard 300, and based on this, A reasonable advertising cost can be set.

In addition, as described above, the advertising effect analysis server 100 according to embodiments of the present invention not only sets reasonable advertising costs, but also determines the liquidity and density of moving objects (e.g., people, cars, two-wheeled vehicles, etc.) within a specific area. It can provide a function to measure .

FIG. 14 is a block diagram showing another example of an advertising effect analysis unit included in the advertising effect analysis server of FIG. 3.

Meanwhile, FIG. 14 shows a modified embodiment of the embodiment described with reference to FIG. 8 in relation to the advertising effect analysis unit 170_1. In order to avoid redundant explanation, the description will focus on differences from the above-described embodiment, Parts that are not specifically described follow the above-described embodiments, and the same numbers indicate the same components and similar numbers indicate similar components.

Referring to FIG. 14, the advertising effect analysis unit 170_1 includes a target setting unit (TSP_1), an area setting unit (ASP_1), a second analysis unit (AZP2_1), a memory (MEM), and a model learning unit (MLP). can do.

Here, as described above, at least some functions of the liquidity analysis unit 160 in FIG. 3 may be configured to be included in the advertising effect analysis unit 170_1. That is, the advertising effect analysis unit 170_1 may be configured to include a target setting unit (TSP_1) and an area setting unit (ASP_1).

Accordingly, the second analysis unit AZP2_1 can receive target setting data (TSD) and area setting data (ASD) from the target setting unit (TSP_1) and the area setting unit (ASP_1). In addition, when performing image analysis on the image data (AD) using the model learning data (AMD), the second analysis unit (AZP2_1) analyzes a specific area based on the target setting data (TSD) and the area setting data (ASD). Second analysis data AD2 may be generated by performing image analysis only on a specific moving object (eg, analysis object TG) in the analysis area AA. Accordingly, depending on the type of advertisement output from the billboard 300, not all users around the billboard 300 or users watching the advertisement output from the billboard 300, but the second analysis unit AZP2_1 The area around the billboard 300 is set as the analysis area (AA), and according to the type of advertisement output from the billboard 300, video analysis is performed only for the analysis object (TG) corresponding to the target group of the advertisement to determine the advertising effect. By analyzing, it is possible to calculate advertising effects more appropriate for the relevant advertisement (for example, in handbag fashion advertisements, women in their 20s or 30s are set as the target of analysis (TG), or in health food advertisements, women in their 40s or older are selected. set to analysis target (TG)).

Figures 15A and 15B are diagrams for explaining an example of communication between an advertising effect analysis server and a camera according to embodiments of the present invention.

Figures 16a and 16b are diagrams for explaining another example of communication between an advertising effect analysis server and a camera according to embodiments of the present invention.

Meanwhile, Figures 15a and 15b show a case where the billboard (300 in Figure 1) is an outdoor billboard, and Figures 16a and 16b show a case where the billboard (300 in Figure 1) is a bus stop shelter billboard.

First, referring to FIGS. 15A and 15B, the camera 200 (e.g., CCTV, etc.) installed around the billboard (300 in FIG. 1) consisting of an outdoor billboard is an advertising effect analysis server through the communication unit 110. You can communicate with (100) wired or wirelessly. For example, as shown in FIG. 15A, the camera 200 (e.g., CCTV, etc.) installed around the billboard (300 in FIG. 1) is connected to the router 111 and the router 112 (e.g., LTE router, etc.) communicates with the advertising effect analysis server 100, or, as shown in FIG. 15B, the camera 200 (e.g., CCTV, etc.) installed around the billboard (300 in FIG. 1) is connected to the router. It is possible to communicate with the advertising effect analysis server 100 through (111) and a wired cable (113).

Next, referring to FIGS. 16A and 16B, the camera 200 installed within the installation structure of the billboard (300 in FIG. 1) consisting of a bus stop shelter billboard is connected to the advertising effect analysis server 100 through the communication unit 110. You can communicate by wire or wirelessly. For example, as shown in FIG. 16A, the camera 200 installed within the installation structure of the billboard (300 in FIG. 1) is transmitted through the router 111 and the router 112 (e.g., LTE router, etc.). The camera 200, which communicates with the advertising effect analysis server 100 or is installed within the installation structure of the billboard (300 in FIG. 1), as shown in FIG. 16b, transmits advertisements through the router 111 and the wired cable 113. It can communicate with the effect analysis server 100.

However, the communication method shown in FIGS. 15A to 16B is simply illustrative, and the embodiment of the present invention is not limited thereto, and the advertising effect analysis server ( 100) and the camera 200 can communicate.

Embodiments according to the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc., singly or in combination. Program instructions recorded on the computer-readable recording medium may be specially designed and configured for the present invention, or may be known and usable by those skilled in the computer software field. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magneto-optical media such as floptical disks. media), and hardware devices specifically configured to store and perform program instructions, such as ROM, RAM, flash memory, etc. Examples of program instructions include not only machine language code such as that created by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform processing according to the invention and vice versa.

As described above, in the case of a method for providing an advertising effect analysis service according to embodiments of the present invention and a server using the same, the advertising effect can be analyzed by accurately reflecting the interest of actual users in the advertisement output from the billboard. And based on this, reasonable advertising costs can be set.

In addition, in the case of a method for providing an advertising effect analysis service according to embodiments of the present invention and a server using the same, an image of an analysis object (moving object) corresponding to the target group of the advertisement is displayed depending on the type of advertisement output from the billboard. By performing analysis and analyzing the advertising effect, it is possible to calculate an advertising effect more suitable for the relevant advertisement.

In addition, in the case of a method for providing an advertising effect analysis service and a server using the same according to embodiments of the present invention, not only the advertising cost setting function described above, but also moving objects in a specific area (for example, people, cars, two-wheeled vehicles, etc.) It can provide a function to measure the liquidity and density of.

The advertising effect analysis server according to various embodiments of the present invention can be described as follows.

In order to solve the problems described above, the server according to embodiments of the present invention provides an advertising effect analysis service, using image data provided from at least one camera that captures images around the billboard. , a fluidity analysis unit that sets at least some of the entire area of the image as an analysis area and sets at least some of the moving objects as an analysis target to analyze the fluidity and density of the analysis target in the analysis area; And using the video data, the number of users around the billboard, gender information, age information, gaze information, and appearance information are extracted, and based on this, the advertising effect of the advertisement output from the billboard is analyzed. It may include an analysis unit.

In one embodiment, the liquidity analysis unit includes a target setting unit that sets the analysis target to generate target setting data based on the image data, and sets the analysis area based on the image data to generate area setting data. An area setting unit that generates, and based on the target setting data and the area setting data, detects the number of the analysis objects present in the analysis area in the image, and determines the fluidity and density of the analysis objects in the analysis area. It may include a first analysis unit that generates the first analysis data by analyzing .

In one embodiment, the advertising effect analysis unit includes a model learning unit that performs a labeling operation on the video data to build model training data, a memory that stores the model training data, and a memory that stores the model training data and the video data and the model training data. Based on this, it may include a second analysis unit that generates the second analysis data for the advertising effectiveness index by performing video analysis on the video data.

In one embodiment, the second analysis unit includes a gaze detection unit that extracts gaze information of users looking at an advertisement output from the billboard based on the image data and the model learning data, the image data, and the model learning data. Based on this, a characteristic detection unit that extracts characteristic information about the number, gender, age, and appearance of users around the billboard, and an analysis data generator that generates the second analysis data based on the gaze information and the characteristic information. It can be included.

In one embodiment, the characteristic information about the appearance may include information about the users' clothing.

In one embodiment, the advertising effect analysis unit includes a model learning unit that performs a labeling operation on the video data to build model training data, a memory that stores the model training data, and, based on the video data, the analysis target. a target setting unit that sets and generates target setting data; an area setting unit that sets the analysis area and generates area setting data based on the image data; and the image data, the model learning data, and the target setting data. , and based on the area setting data, by performing image analysis on the analysis target present in the analysis area among the entire area of the image, the first information on the advertising effectiveness index corresponding to the analysis area and the analysis target 2 It may include a second analysis unit that generates analysis data.

In one embodiment, the server may further include an advertising cost calculation unit that calculates an advertising cost corresponding to an advertising effectiveness index based on the second analysis data.

The advertising effect analysis method according to various embodiments of the present invention can be described as follows.

In order to solve the problems described above, the method according to embodiments of the present invention provides an advertising effect analysis service, and includes receiving image data provided from at least one camera that captures images around the billboard. Step, using the image data, set at least some of the entire area of the image as an analysis area, set at least some of the moving objects as the analysis target, and determine the liquidity and Analyzing density, and using the image data, extracting the number of users around the billboard, gender information, age information, gaze information, and appearance information, and based on this, advertising for the advertisement output from the billboard It may include a step of analyzing the effect.

Computer programs according to various embodiments of the present invention can be described as follows.

In order to solve the problems described above, the computer program according to the embodiments of the present invention is a computer program stored in a computer-readable medium, where the computer program includes instructions, and when executed by a computer system, the instructions can cause the computer system to perform the method.

Although the present invention has been described above with reference to embodiments, those skilled in the art will understand that various modifications and changes can be made to the present invention without departing from the spirit and scope of the present invention as set forth in the claims. You will be able to.

1: Advertising effectiveness analysis system
100: Advertising effect analysis server
110: Department of Communications
120: processor
130: database
160: Liquidity analysis unit
170, 170_1: Advertising effectiveness analysis department
180: Liquidity data storage unit
190: Advertising cost calculation unit
200: Camera
300: billboard
ADG: Analysis data generation unit
ASP: Area Setting Part
AZP1: First analysis unit
AZP2: Second analysis unit
CDP: Characteristic detection unit
GDP: Gaze detection unit
MEM: Memory
MLP: model learning unit
TSP: Target setting part

Claims (1)

In the server that provides advertising effect analysis services,
Using image data provided from at least one camera that captures images around the billboard, at least some of the entire area of the image is set as an analysis area, and at least some of the moving objects are set as analysis targets, a liquidity analysis unit that generates first analysis data by analyzing the liquidity and density of the analysis target within the analysis area; and
Using the video data, the number of users around the billboard, gender information, age information, gaze information, and appearance information are extracted, and based on this, the advertising effect for the advertisement output from the billboard is analyzed to conduct a second analysis. Includes an advertising effect analysis unit that generates data,
The liquidity analysis unit,
a target setting unit that sets the analysis target based on the image data and generates target setting data;
an area setting unit that sets the analysis area based on the image data and generates area setting data; and
Based on the target setting data and the area setting data, by detecting the number of the analysis target present in the analysis area in the image and analyzing the fluidity and density of the analysis target in the analysis area, the first It includes a first analysis unit that generates analysis data,
The advertising effect analysis department,
a model learning unit that constructs model learning data by performing a labeling operation on the image data;
a memory storing the model training data; and
A second analysis unit configured to generate the second analysis data for an advertising effectiveness index by performing video analysis on the video data based on the video data and the model training data,
The second analysis unit,
A gaze detection unit that extracts gaze information of users looking at advertisements output from the billboard based on the video data and the model learning data;
a characteristic detection unit that extracts characteristic information about the number, gender, age, and appearance of users around the billboard based on the video data and the model learning data; and
An analysis data generator that generates the second analysis data based on the gaze information and the characteristic information,
The characteristic information about the appearance includes information about the clothes of the users,
The fluidity analysis unit uses Vision AI technology based on an image recognition algorithm to set a specific area among the entire area of the captured image, select people, cars, and two-wheeled vehicles as moving objects, and analyze within the specific area. Measures the liquidity and density of the object,
The advertising effect analysis unit uses Cascade Regression Forest and Fast Radial Symmetry Transform (FRST) to minimize the impact of changes in lighting and analyzes images collected from the camera.
The second analysis unit analyzes the advertising effect by analyzing images only for the analysis object corresponding to the target group of the advertisement according to the type of advertisement output from the billboard,
Based on the information on the liquidity and density of the analysis object in the analysis area, the user is automatically informed of the location of the analysis area and the liquidity and density of the analysis object at each stage of interest, caution, warning, and risk,
The advertising effect analysis department,
a model learning unit that constructs model learning data by performing a labeling operation on the image data;
a memory storing the model training data;
a target setting unit that sets the analysis target based on the image data and generates target setting data;
an area setting unit that sets the analysis area based on the image data and generates area setting data; and
Based on the image data, the model learning data, the target setting data, and the area setting data, image analysis is performed on the analysis object present in the analysis area among the entire area of the image, A server comprising a second analysis unit that generates the second analysis data on the advertising effectiveness index corresponding to the analysis target.