Advertisement effect analyzer and method thereof

Abstract

To provide an advertisement effect measuring device and method for analyzing and measuring an advertisement effect by using information such as the number of times of viewing an advertisement by a person, the time of viewing the advertisement and the degree of attention.SOLUTION: An advertisement effect analysis system for measuring an advertisement effect of an advertisement medium (19) includes a vision sensor (13) for photographing a person (17) and a vehicle (18) located within a field of view of the advertisement medium, a computing device (50) for receiving and analyzing a video / image photographed by the vision sensor, and a communication device for transmitting analyzed data or an advertisement effect measurement result to a terminal (15) or a server (30), wherein the computing device includes one or more programs for analyzing data related to the person and the vehicle based on the received video / image.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a device and method for analyzing the effectiveness of advertisements displayed on outdoor advertising media, and analyzes information such as people, vehicles, and weather around the advertising media obtained through sensors, and measures the effectiveness of the advertisement based on the information. [Background technology]

[0002] Outdoor advertising, which is carried out indoors or outdoors and targets multiple people (pedestrians, people in vehicles, etc.), is recognized as an effective advertising method because it allows for forced exposure of the advertisement. However, it is difficult to objectively analyze the advertising effectiveness of outdoor advertising. Conventionally, the effectiveness has been analyzed through conventional estimation based on the floating population around the advertising medium.

[0003] Online advertising allows for targeting of viewers (e.g., computer users) based on the content of the ad, and measuring advertising effectiveness is relatively easy compared to outdoor advertising. However, with outdoor advertising, it is not easy to measure advertising effectiveness by analyzing how interested and focused people around the advertising medium (e.g., pedestrians, people in vehicles, etc.) are in the ad. Therefore, a method is needed to accurately measure advertising effectiveness through outdoor advertising media based on objective data. Summary of the Invention [Problem to be solved by the invention]

[0004] In order to solve the above problems, the present invention provides an apparatus and method for measuring advertising effectiveness that can analyze and measure advertising effectiveness using information such as the number of times people view an advertisement, the time they view the advertisement, and the level of attention.

[0005] This device and method obtain information on pedestrians and the flowing population through vision sensors such as cameras, and analyze and quantify whether people within the field of view of advertising media recognize and pay attention to the advertisement, thereby providing users with meaningful analytical data on the effectiveness of advertising.

[0006] Furthermore, by applying deep learning technology to analyze the effectiveness of outdoor advertising media, a method is provided to improve the efficiency and accuracy of advertising effectiveness measurement.

[0007] The problems to be solved by the present invention are not limited to the above-mentioned problems, and unmentioned problems will be clearly understood by a person having ordinary skill in the art of the present invention from this specification and the attached drawings. [Means for solving the problem]

[0008] The present invention provides an apparatus for measuring the advertising effectiveness of advertising media, which includes a vision sensor that photographs people and vehicles positioned within the field of view of the advertising media, a computing device that receives and analyzes the video / images photographed by the vision sensor, and a communication device that transmits the analyzed data or advertising effectiveness measurement results to a terminal or server, and the computing device includes one or more programs that analyze person and vehicle related data based on the received video / images.

[0009] The device for measuring advertising effectiveness includes at least one of the following programs: a human state analysis program that analyzes people's exposure, viewing, and attention states to advertisements; a human number counting program that counts the number of people exposed to advertisements, the viewing population, and the attention population; a human distribution analysis program that analyzes people's gender and age distribution; a human behavior analysis program that analyzes people's movement routes, staying times, and inflow and outflow populations; a spatial analysis program that divides areas within a shooting space and analyzes population density, staying population, etc.; a vehicle analysis program that analyzes the number, type, and moving speed of vehicles; and a potential audience counting program that counts people in vehicles within the field of view of the advertising medium as potential audiences.

[0010] The present invention provides a method for measuring the advertising effectiveness of advertising media, comprising the steps of transmitting videos / images captured by a vision sensor to a computing device, analyzing the videos / images captured by the computing device, and transmitting the analyzed data or advertising effectiveness measurement results to a terminal or a server, wherein the vision sensor captures images of people and vehicles positioned within the field of view of the advertising media, and the analyzing step analyzes person and vehicle-related data from the captured videos / images using a program. In the analyzing step, the computing device or server can analyze the captured videos / images, and the server transmits the analyzed data or advertising effectiveness measurement results to a terminal or another server.

[0011] This method for measuring advertising effectiveness involves analyzing various collected information through a program, combining it with the generated data, generating metadata, and transmitting it to a terminal or server. The metadata includes exposed population, viewing population, attention population, gender distribution, age distribution, viewing rate, attention rate, maximum exposure day, maximum exposure time, average number of times viewed, average viewing time, average dwell time, etc.

[0012] In addition, we provide a method for measuring advertising effectiveness by sharing data collected by multiple advertising media sensors between advertising media to complement and / or improve accuracy of the data, predicting the flow of the population based on this, preparing necessary advertisements in advance, and displaying the advertisements at the scheduled time.

[0013] The solutions to the problems of the present invention are not limited to the solutions described above, and solutions not mentioned will be clearly understood by those skilled in the art to which the present invention pertains from this specification and the accompanying drawings. [Effects of the Invention]

[0014] Through the present invention, the effectiveness of advertisements displayed in advertising media can be objectively analyzed, and advertising industry professionals such as advertisers, advertising agencies, advertising media companies, and advertising media agencies, as well as other users (hereinafter referred to as "users"), can check advertising analysis data and advertising effectiveness measurement results in real time. Users can develop various advertising strategies based on the advertising analysis data.

[0015] This invention can also be used to analyze visitors to offline spaces such as exhibitions, event venues, and pop-up stores. By analyzing the gender, age, dwell time, movement patterns, and gaze of visitors staying at pop-up stores, exhibition halls, large shopping centers, and various stores using cameras and edge computing, offline marketing results can be measured. Furthermore, by utilizing artificial intelligence (AI) technology, the accuracy and efficiency of advertising effectiveness measurement can be improved, providing the basis for an advertising solution and advertising service platform that allows advertisers, advertising media companies, and other users to coexist.

[0016] The effects of the present invention are not limited to the effects described above, and effects not mentioned will be clearly understood by those skilled in the art to which the present invention pertains from this specification and the accompanying drawings. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a diagram illustrating the configuration of a system for analyzing and measuring advertising effectiveness according to an embodiment of the present invention. [Figure 2] 1 is a diagram illustrating a method for analyzing and measuring advertising effectiveness according to an embodiment of the present invention. [Figure 3] 1 is a flowchart illustrating a method for analyzing and measuring advertising effectiveness on a computing device according to an embodiment of the present invention. [Figure 4] 1 is a flowchart illustrating a method for analyzing and measuring advertising effectiveness at a server according to an embodiment of the present invention. [Figure 5]1 is a diagram illustrating analysis data of advertising effectiveness and measurement results of advertising effectiveness in a terminal according to an embodiment of the present invention; [Figure 6] 10 is a flowchart illustrating another embodiment of an advertising effectiveness analysis method according to the present invention. [Figure 7] 1 is a diagram illustrating an embodiment of outputting an outline box of an object and coordinates of key points of a person's body according to the present invention; [Figure 8] 10 is a flowchart illustrating another embodiment of an advertising effectiveness analysis method according to the present invention. [Figure 9] 1 is a diagram illustrating a configuration of an advertisement effectiveness analysis device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0018] The objects, features, and embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, in describing the embodiments, technical contents well known in the technical field to which the present invention pertains will be omitted in order to clearly convey the gist of the present invention without detracting from the gist of the present invention.

[0019] The examples described in this specification are intended to explain the concept of the present invention to those skilled in the art to which the present invention pertains, and the present invention is not limited to the examples described in this specification. The scope of the present invention should be construed as including modifications and alterations that do not deviate from the concept of the present invention.

[0020] The terms used in this specification are generally used in the widest possible sense, taking into consideration the functions of the present invention. However, these terms may change depending on the intentions of those skilled in the art, precedents, the emergence of new technologies, etc. However, when a specific term is defined and used with a specific meaning, the meaning of the term will be separately described. Therefore, the terms used in this specification should be interpreted based on the substantive meaning of the term and the overall content of this specification, rather than simply by the name of the term.

[0021] In this specification, if it is determined that a detailed description of a known structure or function related to the present invention would obscure the gist of the present invention, the detailed description thereof will be omitted.

[0022] Furthermore, the numbers and terms (e.g., first, second, etc.) used in the course of describing the present invention throughout this specification are merely identification symbols for distinguishing one component from another, and do not imply an order of operations or priority, etc., unless otherwise defined in the specification.

[0023] The suffixes "module" or "section" for components used in the following examples are given and / or used for the convenience of writing the specification and do not have any meaning or role that is distinguishable from each other.

[0024] In the following examples, singular expressions include plural expressions unless the context clearly indicates otherwise.

[0025] As used herein, the symbol " / " should be interpreted as including one or more possible combinations of the associated items. For example, "and / or" includes one or more possible combinations of the associated items. That is, "transmit A / B" should be interpreted as transmitting only A, and / or transmitting only B, and / or transmitting both A and B.

[0026] In the following examples, the terms include or have / has mean that the features or components described in the specification are present, but do not exclude the possibility that one or more other features, components, and steps may also be added.

[0027] The drawings are for the purpose of easily explaining the present invention, and the configurations and shapes shown in the drawings may be exaggerated as necessary to facilitate understanding of the present invention. Therefore, the present invention is not limited to those shown in the drawings.

[0028] In some implementations, the order of certain processes may be performed out of the order described. For example, two processes described in succession may be performed substantially simultaneously, or may be performed in the reverse order from that described.

[0029] When a component is referred to as being "coupled" and / or "connected" to another component, it should be understood that the component may be directly coupled and / or connected to the other component, but there may also be other components between them. On the other hand, when a component is referred to as being "directly coupled" and / or "directly connected" to another component, it should be understood that there are no other components between them. Other expressions describing the relationship between components, such as "between" and "immediately between" or "adjacent to" and "directly adjacent to," should be interpreted in the same way. For example, when components are referred to as being electrically connected in this specification, it includes not only cases where the components are directly electrically connected, but also cases where there are other components between them and they are indirectly or electrically connected.

[0030] The methods described in the embodiments may be embodied in the form of program instructions that can be executed by various computer means and stored on a computer-readable medium. The computer-readable medium may include, alone or in combination, program instructions, software, algorithms, data files, data structures, and the like. The program instructions stored on the medium may be specially designed and constructed for the embodiments, or may be well known and available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language code, such as produced by a compiler, but also high-level language code that can be executed by a computer using an interpreter, etc.

[0031] These computer program instructions can be loaded onto a processor of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that the instructions generate means for performing the functions described in the flowchart blocks. Alternatively, the instructions can be stored in computer-usable or computer-readable memory that can direct a computer or other programmable data processing device to implement the functions in a particular way, such that the stored instructions can produce an article of manufacture containing instruction means for performing the functions described in the flowchart blocks. Computer program instructions can also be loaded onto a computer or other programmable data processing device to perform a series of operational steps that cause a computer-executed processor to provide the steps for performing the functions described in the flowchart blocks. A hardware device can be configured to operate with one or more software modules to perform the operations of an embodiment, or vice versa.

[0032] Also, each block may represent a module, segment, or portion of code including one or more executable instructions for performing a specific logical function. In some alternative implementations, the functions described in the blocks may occur out of order. For example, two blocks shown as successive may be performed substantially simultaneously, or the blocks may be performed in reverse order depending on the corresponding functions.

[0033] As used herein, the term "unit" refers to software or a hardware component such as a Field Programmable Gate Array (FPGA) or an Application Specific Integrated Circuit (ASIC). A "unit" performs a specific function, but is not limited to software or hardware. A "unit" may be configured to reside on an accessible storage medium or to implement one or more processors. Thus, according to some embodiments, a "unit" includes components such as software components, object-oriented software components, class components, and task components, as well as processors, functions, attributes, procedures, subroutines, program code segments, drivers, firmware, microcode, circuits, data databases, data structures, tables, arrays, and variables. The functionality provided within a component and a "unit" may be combined into fewer components and units or further separated into additional components and units. Moreover, a component and a "unit" may be embodied to implement one or more CPUs within a device or security multimedia card. According to various embodiments of the present disclosure, a "unit" may include one or more processors. The present invention will now be described in detail with reference to the accompanying drawings.

[0034] FIG. 1 is a diagram illustrating the configuration of a system for analyzing and measuring advertising effectiveness according to an embodiment of the present invention.

[0035] As shown in FIG. 1, the advertising effectiveness analysis system includes advertising media 19 that display advertisements, sensors 13 that collect situational information around the advertising media, computing devices 50 that analyze information (videos / images, weather, etc.) collected from the sensors 13, and a server 30 or a terminal 15 that displays data analyzed by the computing devices 50 and advertising effectiveness analysis results.

[0036] The advertising medium 19 is a device that displays advertisements in the form of printed matter or videos / images to various audiences (e.g., pedestrians, people in vehicles, etc.). The advertising medium 19 functions to continuously display a single printed matter to people, and / or to display various videos / images, information, news, etc. to people in real time according to a pre-set order / time, randomly, or according to the results of advertising effectiveness analysis. The advertising medium 19 includes various types of display-type advertising media such as electronic billboards, digital signage, etc., such as liquid crystal display devices (LCDs) and organic light-emitting display devices (OLEDs). The advertising medium 19 may be equipped with other electronic devices or equipment.

[0037] The advertising medium 19 can be installed indoors or outdoors, on a traveling device 20, on the roof or exterior wall of a building, or alone on the roadside (for example, as a street sign). The advertising medium can also be installed on the outside or inside of a traveling device that travels on a road. In this case, the traveling device 20 includes various means of transportation such as a mobile advertising vehicle, a passenger car, a bus, a truck, a motorcycle, a bicycle, etc. The size of the advertising medium can be changed depending on the location where it is installed and the purpose of use.

[0038] The advertising medium 19 can display advertisements transmitted from the server 30 using a pedestrian pattern that is associated with the driving pattern of the traveling device 20 and the characteristics of pedestrians around the traveling device 20. That is, the advertising medium 19 is provided outside the traveling device 20 and can display advertisements provided by advertisers based on the driving pattern and pedestrian pattern, making it applicable to adaptive advertising.

[0039] The sensor 13 is a device that collects information about the situation around the advertising medium 19. The sensor 13 includes various sensors, such as a camera, CCTV, LiDAR (Light Detection and Ranging) and other vision sensors that capture images of the area around the advertising medium 19, a thermometer, a hygrometer, and other sensors that collect information about the area around the advertising medium 19, such as the weather where the advertising medium 19 is located. The vision sensor may be a commercially available webcam, camera, CCTV, or other device. When the vision sensor is built into the advertising medium, an on-device compact camera may be used. Weather information includes temperature, humidity, PM2.5, precipitation, snowfall, sunrise information, sunset information, wind direction, wind speed, etc. The sensor may further include a GPS, a microphone, and a speaker.

[0040] The computing device 50 is a device that analyzes information collected from the sensor 13, combines the collected information with other data, converts the collected information into data in a format required by the user, and transmits advertising effectiveness analysis data and advertising effectiveness measurement results to the terminal 15 or the server 30. For example, it analyzes people and vehicle-related data from videos / images taken by a camera. The computing device can be installed inside or outside the advertising medium and / or installed in a separate location from the advertising medium and can communicate via wired / wireless.

[0041] The server 30 analyzes the information collected by the sensor 13 and / or combines it with other data, converts it into data in a format required by the user, and transmits it to the terminal 15 or another server. Only one server can be used, or multiple servers can be operated as needed, and the server can transmit advertising effectiveness analysis data and advertising effectiveness measurement results to other servers. The server can be a commonly used network server, and various types of servers such as a general PC, a high-performance PC, or a cloud can be used.

[0042] The computing device 50 and server 30 described above contain various data and one or more programs necessary for analyzing advertising effectiveness. The computing device and server analyze people's poses and the direction of pedestrians' movements (towards or away from the advertising medium), etc. They also analyze the number of people facing the advertising medium 19, the number of people looking at the advertising medium (those who recognize / pay attention to the advertisements displayed on the advertising medium), the amount of time people are paying attention to the advertising medium, etc.

[0043] The terminal 15 is a device that displays data analyzed by the computing device 50 or the server 30 and advertising effectiveness measurement results to the user. Here, users include advertising industry professionals such as advertisers, advertising agencies, advertising media companies, advertising media agencies, and other users. The terminal includes various terminals such as PCs, tablets, laptops, and smartphones, and provides web and / or app-type programs that allow users to easily check advertising analysis data and advertising effectiveness measurement results.

[0044] Information and data collected by sensors on multiple advertising media located within close proximity can be shared between advertising media. This allows for data complementation and / or increased analysis accuracy. For example, by sharing at least one of actual audience data, potential audience data, vehicle data, pedestrian movement direction, gender and age data, advertisement display information, and advertising media environment information between multiple advertising media, information on people near the advertising media, distance, and environmental information can be obtained, and the flow of the mobile population can be predicted. For example, advertisements tailored to a person's gender, age, and situation can be prepared in advance and displayed at the appropriate time. This further increases the effectiveness of advertising.

[0045] FIG. 2 is a diagram illustrating a method for analyzing and measuring advertising effectiveness according to an embodiment of the present invention.

[0046] 1, it has been explained that the advertising medium 19 can be installed indoors or outdoors, on a traveling device (e.g., a mobile advertising vehicle), on the roof or exterior wall of a building, or can be installed alone on a roadside, etc. Figure 2 illustrates an embodiment focusing on the advertising medium and related devices, regardless of the installation location of the advertising medium.

[0047] The sensor 13 includes a vision sensor such as a camera that takes pictures of the area around the advertising medium, and a sensor that collects information about the area around the advertising medium, such as the weather in the area where the advertising medium is installed, and may further include a GPS, a microphone, a speaker, etc.

[0048] The vision sensor (e.g., a camera) captures people 17 and vehicles 18 located around and within the field of view of the advertising medium. The vision sensor's function can be adjusted according to the measurement range, which indicates the size of the space to be captured and the number of people captured on the screen, and the performance of the vision sensor and computing device can be optimized and used according to the size of the space within the field of view of the advertising medium. By determining the resolution, angle of view, focal length, IR (Infrared) conversion function, etc. according to the measurement range, it is possible to detect people at long distances and capture images even at night.

[0049] The present invention can be applied by dividing the measurement range into two or more groups. For example, the performance of the vision sensor and the computing device can be differentiated / optimized and applied according to the size of the viewing space of the advertising medium as follows: a. Small spaces (e.g. elevators, menu ordering media): For media that are 0.5m to 10m away from the advertising media and that can expose approximately 15 people at the same time, analysis can be performed using a calculation board consisting of only a CPU without a small PC or GPU, and photography can be done with a webcam or small camera that does not have a wide angle of view. b. Medium-sized spaces (digital signage installed in offices, cafeterias, stores, etc., and indoor and outdoor media): Media that are located between 1m and 40m away from the advertising media and can expose approximately 50 people at the same time. Analysis can be performed using a small PC, GPU, or NPU computing board, and footage can be captured using webcams, small cameras, and CCTV with more than 1 million images. c. Large spaces (electronic billboards on building exteriors, roadside advertising media / outdoor billboards, etc.): These media are located at a distance of 5m to 130m from the advertising media and can expose approximately 150 people at the same time. Analysis can be performed using high-performance PCs or computing boards such as GPUs and NPUs, and images can be captured using high-resolution cameras and CCTV with more than 2 million images.

[0050] The sensor 13 may be installed on the outside, top, side, or bottom of the advertising medium, and / or may be installed inside the advertising medium (built-in), or may be installed in another location separate from the advertising medium.

[0051] The advertising medium 19 refers to a device that displays advertisements in the form of print or video / image to people. The advertising medium displays one print continuously, or displays video / image advertisements and information in real time according to a pre-set order / time, randomly, or according to the results of advertising effectiveness analysis. When the advertising medium displays not one advertisement but other advertisements or information in a switching manner, it is an advertising medium in the form of a display such as an LCD / LED electronic billboard or digital signage.

[0052] The computing device 50 is a device that periodically or in real time analyzes information collected from the sensor 13 and includes one or more programs that combine / convert the collected information with other data or analyze data necessary for measuring advertising effectiveness. For example, it includes a program that analyzes people and vehicle-related data from video / images captured by a camera. Such programs can be embedded in the computing device, server, cloud, etc.

[0053] In detail, the programs include at least one of the following programs: a human state analysis program that analyzes a person's exposure state, viewing state, and / or attention state; a human number counting program that counts the number of people exposed to an advertisement, viewing an advertisement, and / or paying attention to an advertisement; a human distribution analysis program that analyzes the gender and / or age distribution of people; a human behavior analysis program that analyzes a person's movement path, staying time, and / or inflow and outflow population, etc.; a spatial analysis program that analyzes the density of people or the number of people staying in the shooting space, etc.; a vehicle analysis program that analyzes the number, type, and / or movement speed of vehicles; and a potential audience counting program that counts people in vehicles within the field of view of the advertising medium as potential audience members, but is not limited to the above programs.

[0054] The computing device analyzes the exposure status of people around the advertising medium, whether they are exposed to the field of view where they can see the advertisement, their viewing status (whether they are actually viewing the advertisement), and whether they are paying attention to the advertisement for a preset period of time or longer (e.g., more than one second). Based on this, the number of people exposed to the advertisement, the number of people viewing the advertisement, and the number of people paying attention to the advertisement can be calculated.

[0055] The data on gender and / or age distribution for each population is analyzed based on the population exposed to the ad, the population viewing the ad, and the population paying attention to the ad. The data on gender and / or age distribution for each population is also generated by analyzing the time people spend in view of the ad media and their movement paths. The data on the time spent in view of the ad media, the inflow and outflow population, and movement paths are also analyzed. The area within the filming space is divided into sections to generate data on the population density and population staying in each area.

[0056] Vehicles moving on roads within the field of view of the advertising medium are classified by type (cars, buses, trucks, motorcycles, etc.), and their numbers are tallied. Vehicle speeds are analyzed to generate data. People within the field of view of the advertising medium who recognize and pay attention to the advertisement are classified as actual audience members, while people in vehicles within the field of view of the advertising medium are classified as potential audience members. Weights (e.g., 1.5 people for cars, 10 people for buses, 1 person for trucks, and 1 person for motorcycles) are assigned to the aggregated figures by vehicle type to generate advertising exposure population data for potential audience members. The weights mentioned above are not fixed values, but can be varied depending on various factors such as the location of the advertising medium, weather, and the user's advertising strategy.

[0057] Deep learning technology, which has learned from millions of images, can be applied to detect and track pedestrians, and analyze their gender, age, and whether they paid attention to advertisements. Deep learning technology and other artificial intelligence (AI) models measure the flow of people captured by cameras, improving the accuracy and efficiency of determining the gender and age of pedestrians, and analyzing whether they recognized / attended advertisements, and quantifying this to measure advertising effectiveness.

[0058] The data analyzed by the computing device or the results of measuring the advertising effectiveness are transmitted to the terminal 15 or the server 30 via the network. At this time, the data is converted into a format required by the user and transmitted to the terminal or the server 30 periodically or in real time.

[0059] The computing device may further include a data association program that collects advertisement display information displayed on the advertising medium, collects advertising environment information such as weather in the location where the advertising medium is located through a sensor, analyzes the collected information through a program, and combines it with data to generate metadata.

[0060] The metadata includes exposed population, viewing population, attention population, gender distribution, age distribution, viewing rate, attention rate, maximum exposure day, maximum exposure time, average number of appearances, average viewing time, average staying time, and the like.

[0061] The computing device collects information (advertising display information) about advertisements displayed in each advertising medium by date and time, combines the data or generates metadata combined chronologically, and provides advertising analysis data and advertising effectiveness measurement results for each advertisement.

[0062] Time series merging refers to the process of combining two or more sets of time series data into one dataset by aligning and merging the data at the same time interval between the two or more sets. This type of time series merging has the advantages of integrated analysis, data consistency, and improved performance of predictive models.

[0063] Metadata is data that describes data; that is, additional information that helps better classify and organize raw data, makes it easier to search for specific data, and helps better understand, analyze, and manage data. For example, in a book, the data would be the book's contents, and the metadata would be the book's title, author, publication year, ISBN number, subject keywords, etc. In a digital photograph, the data would be the photographic image, and the metadata would be the date and time it was taken, camera settings (e.g., shutter speed, aperture value (F-stop)), and location (GPS coordinates), etc.

[0064] Weather information (e.g., temperature, humidity, PM2.5, precipitation, snowfall, sunrise / sunset information, wind direction, wind speed, etc.) for the location of the advertising media by date and time is collected, and metadata is generated that is combined with advertising analysis data. In the case of mobile advertising media such as running devices, the current location is updated through location sensors such as GPS, and information on speed and surrounding environment is collected and combined with advertising analysis data.

[0065] The computing device 50 also quantifies the performance of the advertising medium 19. The computing device 50 chronologically combines the advertisement display time of the advertising medium 19 and transmits the quantified value of the advertising performance for each advertisement to the terminal 15 or the server 30 periodically or in real time. The advertising performance includes a quantified value that combines the actual audience-based advertising medium performance (e.g., performance by time zone / date / period / weather, exposed / viewed / attention population, gender / age distribution, etc.) and the potential audience-based advertising medium performance (e.g., by time zone / date / period / weather, exposed population, gender / age distribution, etc.).

[0066] The computing device 50 may be installed inside (built-in) or outside the advertising medium 19, or may be installed in a different location separate from the advertising medium 19. The computing device 50 may include edge devices such as PCs, tablets, smartphones, and laptops, and may also perform video analysis deep learning models. The operating system (OS) of the computing device may be embedded with various software, such as Windows, Linux, Android, or iOS-based. The deep learning model built into the edge device and used to analyze advertising effectiveness is lightweight and optimized for on-device characteristics, so it can be run using only a central processing unit (CPU) without a calculation accelerator chip such as a graphics processing unit (GPU) or neural processing unit (NPU).

[0067] The terminal 15 is a device that displays data analyzed by the server 30 or the computing device 50 and advertising effectiveness measurement results to the user. Users (e.g., advertising industry professionals such as advertisers, advertising agencies, advertising media companies, advertising media agencies, etc.) can check the advertising analysis data and advertising effectiveness measurement results provided by the computing device 50 in real time through the terminal 15. Based on this, the user can develop various advertising strategies. Terminals include various devices such as PCs, tablets, laptops, and smartphones.

[0068] The server 30 collects various data collected by the sensors of the advertising media in one place, analyzes the data, combines and converts the data into a form required by the user, and transmits it to the terminal 15 or another server. The computing device 50 connected to the advertising media 19 can analyze the information collected by the sensors 13. When the number of advertising media 19 is large, or when providing users with an advertising effectiveness measurement platform service, it may be efficient to use a server for analysis. The server can transmit advertisements desired by advertisers to the advertising media 19.

[0069] The computing device 50 can perform calculations with low power consumption. Video / image transmission from the sensor 13 to the computing device 50 can be via wired or wireless communication, and for wired communication, connection can be made via a LAN cable, USB cable, etc. The sensor 13 can be supplied with power independently from the advertising medium 19 or the computing device 50, or can be supplied with power from the computing device 50 via PoE, USB, etc. along with video / image transmission.

[0070] FIG. 3 is a flow chart illustrating a method for analyzing and measuring advertising effectiveness on a computing device according to an embodiment.

[0071] 3, the vision sensor captures images of people and vehicles within its field of view that can see advertisements displayed on advertising media 19 (S301), and transmits the captured video / images periodically or in real time to a computing device 50 (S303). The video / images can be transmitted from the vision sensor to the computing device in various ways, such as wired or wireless communication.

[0072] The computing device analyzes the received captured video / images (S305). At this time, the computing device uses one or more programs to analyze the people and vehicle-related data required for measuring advertising effectiveness in the received video / images. In the above analysis process, pedestrian and vehicle detection, object tracking, pose analysis, gender analysis, age analysis, etc. in the video / images are performed periodically or in real time. The programs can be embedded in the computing device, server, cloud, etc.

[0073] The program performs at least one of the following: analyzing people's exposure, viewing, and / or attention states; counting the number of people exposed to the advertisement, the number of people viewing the advertisement, and / or the number of people paying attention to the advertisement; analyzing people's gender and / or age distribution; analyzing people's behavior such as movement paths, staying time, and / or inflow and outflow population; analyzing the population density and / or staying population in the shooting space; analyzing the number, type, and / or movement speed of vehicles; or counting people in vehicles within the field of view of the advertising medium as potential audience members.

[0074] The computing device can convert the analyzed data into de-identified data and store it in a storage device, server, or cloud (S306). The original videos / images of people and vehicles are not stored, and only de-identified data is stored, protecting personal information. De-identified data is de-identified data that cannot identify people or vehicles in the videos / images because it assigns an arbitrary ID to people and vehicles during the video analysis process and records information such as gender, age, time spent, and movement.

[0075] The analyzed data or advertising effectiveness measurement results are transmitted to the terminal or server, and the transmitted data includes people's status, number of people, people's distribution, people's behavior, space analysis, vehicle analysis, actual / potential audience data, etc.

[0076] The collected data, analyzed data, or advertising effectiveness measurement results are transmitted to the terminal or server periodically (e.g., in units of seconds / minutes / hours, daily, etc.) or in real time (S307). At this time, the transmitted data includes actual audience data (e.g., person / vehicle ID, vehicle type, exposure / viewing / attention time, gender, age, dwell time, movement direction, movement speed, current time, etc.), potential audience data, advertisement presentation information (advertisement ID, advertisement type, transmission start time, transmission end time, video length, current time, etc.), advertisement media environment information (media ID, location, latitude, longitude, movement speed, weather, current time, etc.), etc. Data can be transmitted from the computing device to the terminal or server in a variety of ways, including wired communication and wireless communication.

[0077] The computing device periodically / in real time transmits and updates metadata, advertising media performance, advertising results, related statistics, program analyzed data, advertising effectiveness measurement results, etc. to the terminal or server. The user can check the transmitted data and information in real time through the terminal.

[0078] FIG. 4 is a flowchart illustrating a method for analyzing and measuring advertising effectiveness at a server according to an embodiment of the present invention.

[0079] As shown in Figure 3, the advertising effectiveness can be measured by analyzing the information collected on the computing device, but the advertising effectiveness can also be measured by analyzing the information collected on the server. In this case, the server analyzes the information collected in parallel with the computing device or independently.

[0080] As shown in Figure 4, the vision sensor captures images of people and vehicles within the field of view of the advertising media (S401) and transmits the captured video / images to a server periodically or in real time (S403). The server analyzes the received video / images (S405). The server uses one or more programs to analyze the person- and vehicle-related data required for measuring advertising effectiveness from the received video / images. The programs can be embedded on the server or in the cloud. The analyzed data can be converted into non-identifiable data and stored in a storage device, server, or cloud (S406). The analyzed data or advertising effectiveness measurement results can then be transmitted to a terminal, computing device, or another server (S407).

[0081] At this time, the server analyzes various data collected by the sensors, combines / converts the data into a format required by the user, and transmits it to the terminal, computing device, or another server. Although the data can be analyzed on the computing device, if there are a large number of advertising media and a large amount of data, it may be more efficient to use a server for analysis when providing users with an advertising effectiveness measurement platform.

[0082] FIG. 5 is a diagram illustrating advertising effectiveness analysis data and advertising effectiveness measurement results in a terminal according to an embodiment of the present invention.

[0083] The data analyzed by the computing device 50 or the server 30 and the results of measuring advertising effectiveness are transmitted to and updated on the terminal 15. The user can check the transmitted data and measurement results in real time through the terminal 15. The data and information that can be checked through the terminal 15 include the location of the advertising medium 19, advertising effectiveness analysis equipment and location, exposed population, viewing population, attention population, floating population, gender distribution, age distribution, viewer rating, attention rate, maximum exposure day, maximum exposure time, average number of views, average viewing time, average dwell time, metadata, advertising medium performance, advertising results, other related statistical values, graphs, etc. Alternatively, the user can select only the data they want to know and set it to be displayed on the terminal.

[0084] At this time, the terminal visualizes the analyzed data and information, improving the user's readability and ease of use. The user can specify the date or period they wish to check, and can download the advertising analysis data and advertising effectiveness measurement results for the selected period in various report formats (e.g., Word, PDF, Excel, etc.). Through this, the present invention can achieve both data analysis and data visualization in measuring the advertising effectiveness of advertising media.

[0085] FIG. 6 is a flowchart illustrating another embodiment of an advertisement analysis method.

[0086] The advertising effectiveness measurement program analyzes people and vehicles in videos / images captured by vision sensors in real time to generate advertising effectiveness measurement data. The program can be installed on a computing device, server, or cloud.

[0087] As shown in Figure 6, the analysis process can be roughly divided into three parts: data collection, object detection and analysis, and process optimization. Specific processes within each part can be performed sequentially or simultaneously in parallel. The specific processes within each part are as follows:

[0088] The first part, data collection, uses vision sensors to collect video / image information on people and vehicles in real time.

[0089] The second part, object detection and analysis, uses technologies such as object detection, body pose estimation, head pose estimation, person feature extraction and attribute recognition, and vehicle attribute recognition to perform audience characteristics (e.g., gender, age), behavior (e.g., viewing, attention), and vehicle aggregation analysis.

[0090] Object detection technology detects objects such as people and vehicles, and outputs the outline of each object in box coordinates, as shown in Figure 7(a). Body pose estimation technology simultaneously outputs the coordinates of the main points of the human body (17 points, including the eyes, nose, ears, shoulders, elbows, pelvis, knees, and feet) and the coordinates of the body's outline box, as shown in Figure 7(b). Detection ensemble technology can be applied to combine the results of object detection and body pose estimation (e.g., to find the intersection of the body's outline box coordinates) to more accurately detect the body's outline. Using only object detection or body pose estimation technology can result in low accuracy in the body's outline box and intermittent inability to detect people. Therefore, detection ensemble technology can be applied to improve detection accuracy. In addition, Detection Ensemble can combine not only Object Detection and Body Pose Estimation results but also Head Pose Estimation results to detect the outline of a person more accurately.

[0091] Videos / images of people or vehicles detected by Object Detection and those corrected by Detection Ensemble technology are analyzed for gender and age through Person Feature Extraction + Attribute Recognition technology, and gaze direction is extracted through Head Pose Estimation technology. Vehicle attribute recognition classifies the vehicle type (passenger car, bus, truck, motorcycle, etc.) and analyzes movement speed, direction, etc.

[0092] The third process optimization part utilizes detection ensemble, object tracking, and ReID matching technologies to improve analysis speed and accuracy.

[0093] Filmed video is a collection of consecutive images, and people and vehicles moving in the video change position slightly in consecutive images. The object detection and analysis part mentioned above corresponds to object detection and analysis for a single image, but if multiple consecutively transmitted images are only detected up to the second part, there is a problem that the same person is recognized as a different person in each image. To solve this problem, object tracking technology is used.

[0094] Object tracking technology is a technology that identifies an object detected in a previous image as the same object in the next image in a series of images. In other words, it tracks the same person or vehicle in a series of images and assigns the same ID to the same object. By assigning the same ID to the same object according to the movement of people and vehicles, it is possible to confirm the characteristics (e.g., gender and age) and behavior (e.g., viewing, attention, speed, direction of movement, etc.) data of the same object.

[0095] In this process, if the accuracy of object tracking is not high, the object can be tracked continuously without interruption, and to improve the accuracy, ReID Matching technology is used. ReID Matching technology is an algorithm that determines whether the objects are the same based on the overlapping area between the outline boxes of objects detected in consecutive images.

[0096] Additionally, prediction accumulation technology can be further utilized. Prediction accumulation technology is a data conversion technology that converts measurement data accumulated by object ID through object tracking into the desired data format using arithmetic logic, conditional logic, and majority logic. For example, if the gaze direction value of data accumulated for ID 176 (e.g., a person) over a 0.3-second period is within the viewing angle five times and not within the viewing angle once, ID 176 is recorded as being in a viewing state for 0.3 seconds. This majority-voting logic is also a method of classifying gender and age data analyzed in real time in the direction that is most dominant.

[0097] Through this process, the characteristics and behavior of people and vehicles can be analyzed in real time to measure the effectiveness of advertising.

[0098] FIG. 7 is a diagram illustrating an embodiment of outputting the outline box of an object and the coordinates of the main points of a person's body.

[0099] The object detection technology uses a deep learning model to detect people, vehicles, etc., and outputs the outline of each object in box coordinates as shown in Figure 7(a). The pose estimation technology simultaneously outputs the coordinates of the main points of the human body (17 points, such as eyes, nose, ears, shoulders, elbows, pelvis, knees, and feet) and the box coordinates of the human outline as shown in Figure 7(b). The number of points is not limited to 17, and it is possible to output coordinates of more or less than 17 points depending on the settings.

[0100] FIG. 8 is a flowchart illustrating another embodiment of the advertising effectiveness analysis method.

[0101] According to the embodiment of Figure 8(a), a vision sensor detects pedestrians and vehicles located around an advertising medium (S801). Figure 8(b) is a diagram showing a screen of an embodiment in which a vision sensor detects pedestrians and vehicles.

[0102] It is important for object detection to detect all individual objects, including people and vehicles, in all videos and images. Furthermore, accuracy decreases if the detected object box points to random coordinates or is mistaken for another type of object. This invention provides two technologies (Object Detection and Body Pose Estimation) in parallel. Furthermore, detection ensemble technology improves the accuracy of the object box coordinates detected by each technology. Detection ensemble detects the outline of a person more precisely by using the intersection and intersection of the outline boxes of the person output by each model. Furthermore, it upscales the input image resolution and optimizes the detection threshold for each technology to detect even distant objects. Of course, accuracy increases when shooting in real time (e.g., 30 FPS or higher) using a high-resolution camera with a high-performance GPU.

[0103] The system predicts pedestrian attributes based on the detected information (S803) and analyzes advertisement recognition / attention (S805). Using an image analysis AI model (e.g., Vision Transformer (ViT)), various pedestrian attributes (e.g., exposure time, gender, age, movement direction, etc.) can be predicted based on the image. Exposure time refers to the amount of time a pedestrian is exposed to an advertisement. Gender and age are determined by capturing and analyzing features that can distinguish the gender / age from the pedestrian's clothing through an AI technology-based deep learning model.

[0104] When analyzing pedestrians' recognition / attention to an advertisement (S805), head pose estimation technology can be used to determine whether they are recognizing / attention to the advertisement. A 3D vector of the head's direction can be predicted from a 2D upper body image. The pedestrian's head pose is estimated and restored to a 3D coordinate system, and the probability of viewing the advertisement is predicted through this. The time spent recognizing / attention to the advertisement is determined through a pose algorithm to determine whether it is simple recognition or attention to the advertisement. Additionally, the size and position of the outdoor advertising medium and the relative position of the camera are calibrated to define a world coordinate system. Whether the head direction is correct relative to the world coordinate system can be accurately analyzed.

[0105] When detecting and tracking pedestrians (S807), if tracking is lost, analysis performance can be supplemented through pedestrian re-identification (S809). Pedestrian detection and tracking may fail intermittently, but a pedestrian re-identification algorithm can be used to supplement this. By applying the re-identification algorithm, pedestrians can be tracked over a long period of time, and the analyzed information can be accumulated to improve the final prediction performance.

[0106] Object tracking is important in videos and multiple images. If object tracking is not provided for multiple images received in succession, the same person may be recognized as a different person. Object tracking technology is an algorithm that identifies an object detected in a previous image in a series of images as the same object in the next image. This allows the same object (e.g., person, vehicle, etc.) to be assigned the same ID, and data on the characteristics (e.g., gender, age) and behavior (e.g., viewing, attention, speed, direction of movement, etc.) of the same object can be secured. In this process, ReID Matching technology can be applied to improve the accuracy of object tracking.

[0107] The system analyzes the advertising effectiveness by pedestrian attributes and operates a dashboard that displays advertising analysis data and advertising effectiveness measurement results on the user's device (S811). It can also compile the number of vehicles passing around the advertising media based on vehicle information collected through sensors (S813).

[0108] This invention provides machine learning and deep learning technologies to analyze and measure advertising effectiveness. Information collected by sensors such as cameras can be converted into machine-learnable features through AI image recognition algorithms, and a server or computing device can analyze and measure advertising effectiveness based on the AI ​​model.

[0109] The commonly known facial image-based gender and age classification model loses accuracy at distances of 5m or more because the image information (pixels) shown by a face is too small. Since people often view outdoor advertising media from distances of 5m or more, it is important to improve the accuracy of analysis of objects at distances of 5m or more.

[0110] The model applied to this invention learns and distinguishes the characteristics of the entire human body, and because it learns more than 10 times more information than a face-image-based model, it has high analysis accuracy for people at a distance of more than 5 meters and can classify gender and age from just a person's back view.To achieve this, it divides the human body into three parts (e.g., upper body, lower body, and legs) and applies an image analysis AI model (e.g., Transformer (ViT)) that learns the characteristics of each part and improves inference performance.

[0111] ViT is a technology for image recognition in the field of computer vision that works well on very large datasets, allows for the expansion of model size, and can be applied to a variety of image recognition tasks.

[0112] In addition, conventional whole-body-based models were trained on people all over the world, and the training weight for Asians and Koreans was relatively low, resulting in low accuracy for Asians and Koreans. To solve this problem, we learned the human body characteristics of people by region or country, and in particular, we labeled and trained approximately 1 million images of Asians and Koreans, raising the accuracy to 85% to 95%. This accuracy rate is much higher than other models.

[0113] The processing speed of computing devices and servers is important for improving the efficiency of advertising effectiveness analysis. In particular, model lightweighting is necessary for computing devices to analyze information on-device in advertising media. Among video analysis deep learning models, models that analyze gender, age, etc. based on the overall characteristics of the human body require a large amount of calculation, which requires a long time for analysis. In order to reduce analysis time, it is necessary to increase processing speed. To achieve this, deep learning models can apply quantization calculations when calculating weights, reducing the amount of calculation by more than two times.

[0114] Quantization is a technique for reducing the amount of calculations. The main purpose of quantization is to reduce the complexity of calculations, save storage space, and reduce power consumption. Quantization can greatly improve the efficiency of machine learning models without reducing their performance, helping AI models work faster and with fewer resources.

[0115] In particular, smartphones and small devices have limited resources available for operation, so it is necessary to reduce the size and computational complexity of artificial intelligence models through quantization technology.

[0116] Even with this weight reduction, the accuracy of the analysis must be maintained compared to before the weight reduction. The present invention's parallel processing between the model weight reduction and the analysis model, as well as improvements to the process flow, can reduce processing time by more than 50%. When analyzing a single image for a large space (e.g., at least 50 people), it takes 78.9 ms for the processor to perform all of the detection, gender and age classification, attention, tracking, and re-identification steps. This allows for the analysis of 12 images per second, more than three times faster than conventional methods.

[0117] JPEG2026031363000002.jpg19170

[0118] FIG. 9 is a diagram illustrating the configuration of an advertisement effect analysis device according to an embodiment.

[0119] The advertising effectiveness analysis device includes an output unit 901, a sensor unit 903, a control unit 905, an analysis unit 907, a storage unit 909, and a communication unit 911. The output unit outputs advertisements through advertising media 19 such as a display device. The sensor unit includes various sensors 13 such as vision sensors and environmental information collection sensors, and collects information around the advertising media. The control unit controls the output unit, sensor unit, analysis unit, storage unit, communication unit, etc. The analysis unit includes a computing device 50 and performs the function of analyzing collected data and measuring advertising effectiveness. The storage unit can store collected information, analyzed data, advertising effectiveness measurement results, and information necessary for advertising presentation. The communication unit communicates with various devices such as the sensor 13, terminal 15, advertising media 19, driving device 20, and server 30. However, the various units (901 to 909) mentioned above can be eliminated and / or installed and operated elsewhere depending on the advertising media operation method and situation.

[0120] Although the present invention has been described above through limited examples and drawings, those skilled in the art will recognize that various modifications, variations, applications, and combinations may be made from the above description without departing from the essential features of the technical ideas and forms of the present invention. For example, the described techniques may be performed in a different order than described, and / or the components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different manner than described, and / or may be replaced and / or substituted with other components or equivalents, and still achieve suitable results. Therefore, the content associated with such modifications, variations, applications, and combinations, other embodiments, and equivalents to the claims, should be construed as falling within the scope of the following claims.

Claims

1. An apparatus for measuring the advertising effectiveness of an advertising medium, one or more processors; Memory and one or more programs stored in the memory and configured to be executed by the processor; The one or more programs: A step of photographing a person or vehicle located within the field of view of the advertising medium with a vision sensor; receiving and analyzing the captured video / images on a computing device; transmitting the analyzed data or the advertising effectiveness measurement results to a terminal or a server by the communication device; and analyzing, by the computing device, information related to people, space, or vehicles based on the received video / image. Advertising effectiveness measurement device.

2. The program a human state analysis program that analyzes at least one of a person's advertisement exposure state, viewing state, and attention state; a program for counting the number of people that counts at least one of the number of people exposed to an advertisement, the number of people viewing the advertisement, and the number of people paying attention to the advertisement; a human distribution analysis program that analyzes at least one of a human gender distribution and an age distribution; a human behavior analysis program that analyzes at least one of a person's movement route, residence time, and inflow and outflow population; a spatial analysis program that divides areas within the shooting space and analyzes at least one of population density and resident population; a vehicle analysis program that analyzes at least one of the number, type, and movement speed of vehicles; and a potential audience counting program that counts people in vehicles within the field of view of the advertising medium as potential audiences. The advertising effectiveness measurement device according to claim 1 .

3. the computing device, Collecting advertisement display information displayed on the advertising medium, and collecting advertising medium environment information such as weather at the location where the advertising medium is located through a sensor; The method further includes a data combining program that analyzes the collected information through the program, combines the collected information with generated data, and generates metadata; The metadata includes the exposed population, the viewing population, the attention population, the gender distribution, the age distribution, the viewing rate, the attention rate, the maximum exposure day, the maximum exposure time, the average number of appearances, the average viewing time, the average staying time, etc. The advertising effectiveness measurement device according to claim 1 .

4. The vision sensor The function of the vision sensor can be adjusted according to the size of the space being photographed and the measurement range that indicates the number of people in the image. The performance of the vision sensor and the computing device is differentiated and used depending on the space within the field of view of the advertising medium. The advertising effectiveness measurement device according to claim 1 .

5. the computing device, Quantifying advertising media performance based on the metadata and transmitting the resulting data to the terminal or the server; 4. The advertising effectiveness measuring device according to claim 3, wherein the advertising effectiveness measuring device quantifies the advertising performance of each advertisement by combining the advertising performance of the advertising media with the advertising performance of the advertising media in a time series manner and transmits the quantified advertising performance to the terminal or the server.

6. the computing device, It includes an AI model for image analysis that learns the characteristics of the entire human body and classifies people's gender and age even from a distance. The accuracy of the gender and age analysis is improved by additionally learning the physical characteristics of people by region or country. The advertising effectiveness measurement device according to claim 1 .

7. Data collected by multiple advertising media sensors is shared between the advertising media to complement the data and / or improve the accuracy of analysis; Based on the shared data, we predict the flow of the floating population and prepare and display advertisements that match gender, age, and situation in advance. The shared data includes at least one of personal data including gender and age, vehicle data, advertisement display information, and advertisement media environment information; The advertising effectiveness measurement device according to claim 1 .

8. the computing device, When calculating the weights of the AI ​​model, quantization calculation is applied to reduce the amount of calculation and improve the analysis processing speed. The data collected from the sensor is analyzed using a lightweight artificial intelligence model, and the results are transmitted to the terminal or the server. The advertising effectiveness measurement device according to claim 5.

9. the computing device, Detection ensemble technology is applied to more precisely detect the outline of a person using the intersection of the coordinates output by object detection and the coordinates output by body pose estimation. The object detection outputs the outline coordinates of a person, and the body pose estimation outputs the coordinates of five or more main points of the person's body. The advertising effectiveness measurement device according to claim 1 .

10. the computing device, It is characterized by using an object tracking technology that determines whether or not the same object exists in consecutive images, and a ReID matching technology that determines whether or not the same object exists based on the overlapping area between the outline boxes of objects detected in consecutive images. The advertising effectiveness measurement device according to claim 1 .

11. A method for measuring advertising effectiveness of an advertising medium, comprising: by one or more processors executing one or more programs stored in memory, A step of photographing a person or vehicle located within the field of view of the advertising medium through a vision sensor; receiving and analyzing the captured video / images on a computing device; transmitting the analyzed data or the advertising effectiveness measurement results to a terminal or a server by the communication device; The analyzing step is characterized in that the step of analyzing person, space, or vehicle-related information based on the received video / image using the program. How to measure advertising effectiveness.

12. The program Analyze at least one of a person's advertising exposure state, viewing state, and attention state, Count at least one of the following: the number of people exposed to the ad, the number of people watching, and the number of people paying attention to the ad. Analyze at least one of the gender distribution and age distribution of people, Analyze at least one of the following behaviors: people's movement routes, residence times, and inflow and outflow populations. Divide the area within the shooting space and analyze at least one of the population density and the remaining population. Analyzing at least one of the number, type, and speed of vehicles; and In counting people in vehicles within the field of view of advertising media as potential audiences, Contains at least one or more The method for measuring advertising effectiveness according to claim 11.

13. In the analyzing step, the server analyzes the captured video / image, or the computing device and the server analyze the captured video / image together; transmitting the analyzed data or advertising effectiveness measurement results to the terminal or another server; The method for measuring advertising effectiveness according to claim 11.

14. The computing device converting the analyzed data into non-identifiable data and storing it in a storage device or on the server; The method for measuring advertising effectiveness according to claim 11.

15. Transmitting the analyzed data or advertising effectiveness measurement results to the terminal or the server; The data transmitted at this time includes at least one of the following: human status, number of people, human distribution, human behavior, spatial analysis, vehicle analysis, and latent human data. The method for measuring advertising effectiveness according to claim 11.

16. Collecting advertisement display information displayed on the advertising medium and collecting advertising medium environment information such as weather at the location where the advertising medium is located through a sensor; The collected information is analyzed through the program, and combined with the generated data to generate metadata; Transmitting the metadata to the terminal or the server; The metadata includes the exposed population, the viewing population, the attention population, the gender distribution, the age distribution, the viewing rate, the attention rate, the maximum exposure day, the maximum exposure time, the average number of appearances, the average viewing time, the average staying time, etc. The method for measuring advertising effectiveness according to claim 11.

17. quantifies advertising media performance based on the metadata and transmits the quantified results to the terminal or the server; The advertising performance of each advertisement is quantified by combining the advertisement display time of the advertising medium with the time series and transmitted to the terminal or server. The method for measuring advertising effectiveness according to claim 16.

18. Data collected by multiple advertising media sensors is shared between the advertising media to complement the data and / or improve the accuracy of analysis, Based on the shared data, we predict the flow of the floating population and prepare and display advertisements that match gender, age, and situation. The shared data includes at least one of personal data including gender and age, vehicle data, advertisement display information, and advertisement media environment information; The method for measuring advertising effectiveness according to claim 11.

19. The analyzing step learns the characteristics of the entire human body through an image analysis artificial intelligence model and classifies the gender and age of a person even at a long distance. and additionally learning the physical characteristics of people by region or country to improve the accuracy of the gender and age analysis. The method for measuring advertising effectiveness according to claim 11.

20. When calculating the weights of the AI ​​model, quantization calculation is applied to reduce the amount of calculation and increase the analysis processing speed. Analyzing the data collected from the sensors using a lightweight artificial intelligence model and transmitting the results to the terminal or the server; The method for measuring advertising effectiveness according to claim 17.

Images