Digital marketing advertisement method and system based on ai analysis of eyes and gaze of a user of a smart device equipped with a camera

Abstract

This invention proposes a technology capable of accurately grasping internet users' interests and dynamically adjusting ad display accordingly. Specifically, it proposes an advertising method and system utilizing AI gaze analysis. The method includes the following steps: collecting image or video data, including the user's face, from a user terminal (including a camera) exposed to internet ads, so that it can be processed by AI analysis tools installed on the user terminal or ad server; extracting information related to the user's gaze from the collected image or video data using the AI ​​analysis tools, and determining whether the user's gaze overlaps with the ad display area based on the gaze information; when it is determined that the user's gaze is pointing towards the ad display area, changing the ad's exposure method on the user terminal; furthermore, this invention also proposes a scheme to maximize advertising profits by analyzing customer gaze at ads.

Description

Technical Field

[0001] This invention relates to a digital marketing advertising method, computer recording medium, and system based on artificial intelligence (AI) analysis of the eyes and gaze of users of smart devices. The invention aims to maximize the effectiveness of digital or online advertising by analyzing the gaze of users of camera-equipped smart devices (hereinafter referred to as "AI") and applying the analysis results in real time to digital marketing. Background Technology

[0002] Thanks to advancements in wired and wireless networks and computer technology, the global digital advertising market is projected to reach $700 trillion by 2023 alone. There are various definitions and standards for the concept and types of digital advertising. Generally, internet digital advertising utilizing internet technology operates primarily through three methods: CPI (cost per impression), CPC (cost per click), and CPA (cost per action).

[0003] CPI (Cost Per Thousand Impressions) refers to the method of charging 0.5 to 5 Korean won (hereinafter referred to as "Korean Won") for displaying internet advertisements to users on large internet portals that already have a large number of website users or visitors (end users, customers). Advertisers sign advertising contracts that are valid until a certain number of ad impressions are reached, or they receive a guarantee of ad impressions, such as on a specific platform. The large portal will place the advertisement at a fixed rate for the predetermined period, regardless of the number of impressions. This fixed-term advertising model is also known as CPT (Pay-Per-Time). For example, in some cases, advertisers agree to pay a certain advertising fee to the large portal for every 1,000 impressions. Of course, if the portal entrusts other advertising agencies to place the advertisements, the advertiser may need to pay an additional agency fee. However, in many cases, large portals will directly develop advertising plans for advertisers without the need for third-party advertising agencies.

[0004] In a pay-per-click (CPC) model, advertisers don't pay for ads solely based on impressions; they pay for each ad click. The cost per click is calculated after determining the actual number of clicks on the internet ad. For example, in keyword advertising, which typically uses this CPC model, the cost per click can range from 100 to several thousand Korean won. Different CPC costs may be based on a bidding system related to specific keywords. If an advertiser agrees to pay a higher bid for a specific keyword, their ad will be more likely to be seen when internet users search for that keyword.

[0005] However, it also has some long-standing problems, such as searches or clicks made by machines (e.g., search engine bots), meaningless and persistent clicks by users, and deliberate fraudulent clicks designed to trick competitors into wasting advertising budgets. From the advertiser's perspective, this can lead to wasted advertising funds. Of course, various technological methods are currently being developed to address the problem of fraudulent clicks.

[0006] CPA (Cost Per Action) may be the most advantageous method for advertisers. It's an advertising payment structure where fees are only settled when an ad performs, such as when an internet user visits the advertiser's website through an online ad and registers as a member, applies for a credit card, purchases goods, or participates in an activity. For example, advertisers might sign marketing agreements with multiple website operators and pay fees only when specific advertising events occur. However, if the marketing partnership is unsuccessful, part of the planned advertising budget may be wasted. Furthermore, consumers may subsequently cancel purchases, preventing advertisers who have paid CPA fees from achieving actual advertising results.

[0007] While CPI, CPC, CPA, or combinations thereof may seem to differ significantly in their methods of calculating advertising costs, all of these metrics can be viewed as measures of advertising costs based on whether an advertiser's online ads attract user attention. However, CPI presupposes that an ad might have attracted a user's attention simply by displaying the ad on a website the user visited. Cost per click (CPC) presupposes that if a user clicks on the ad, it indicates that the ad has successfully attracted their attention. Cost per conversion (CPA) is an internet advertising technique that considers an ad to have truly attracted user attention only when an event such as a user making a purchase occurs.

[0008] However, among the various internet advertising technologies currently proposed, there is still no clear answer on how to increase user interest in advertisements. In fact, the criteria for judging whether an advertisement attracts user interest are also unclear. This article will take targeted advertising as an example to introduce another traditional marketing strategy.

[0009] Targeted advertising refers to an advertising strategy where advertisers leverage the characteristics of potential consumers and associate them with the products or services to be promoted. For example, if consumer Z visits search engine website X and then visits a specific advertiser's website Y, the operator of search engine website X may know Z's browsing history on Y and Z's IP address. Therefore, the operator of search engine website X can improve advertising efficiency by targeting the ads requested by advertiser Y to users with similar characteristics or website browsing habits to consumer Z. If consumer Z has visited advertiser Y's website for some reason, targeted advertising is becoming a very important tool, especially in the field of online advertising. This is because providing consumer Z with Y's advertising materials is expected to be effective; for example, potential consumer Z is more likely to visit Y's website again and purchase products after seeing the ads. In the above example, it can be seen that the operator of the search engine website X acts as an advertising agency, while Y becomes the advertiser, paying advertising fees to X in the expectation that X can effectively provide targeted ads to potential consumers who are more likely to be interested in Y's ads.

[0010] Of course, targeted advertising isn't limited to search engines; it can also be used as a targeted advertising platform. Accessibility to channels frequently visited and used by potential buyers, such as IPTV, smartphone apps, and social media platforms, can facilitate targeted advertising. IPTV can understand what content consumers have watched. App developers can understand which types of smartphone users are using their apps. Social media platforms can understand user activity on their platforms. Therefore, IPTV, apps, and social media can all serve as advertising platforms to deliver ads to platform users. In short, depending on the targeted advertising strategy, advertisers can achieve effective advertising results as long as they can select the right audience and successfully attract their attention to products or services relevant to the characteristics or habits of potential users.

[0011] Despite these efforts, current targeted advertising systems still face several problems, broadly categorized into two main types: issues on the part of advertising companies (i.e., advertisers) and issues on the part of internet users and consumers. First, from the advertiser's perspective, low conversion rates (CVR) have always been a problem. Conversion rate refers to the percentage of internet users who see a specific ad and ultimately take action such as clicking or making a purchase. Conversion rates typically range from 2% to 10%. If a conversion rate is higher than 10%, it can be assumed that the advertiser is spending a significant amount of money on popular websites with high pay-per-click rates. However, on the other hand, from the consumer's perspective, most web pages are likely saturated with offensive digital ads, which can be considered a form of web pollution. Regardless of the specific advertising strategy, so many pop-ups, banner ads, or affiliate marketing campaigns can be perceived by consumers as web pollution, almost indistinguishable from their experience of receiving spam.

[0012] As mentioned above, the key to internet advertising lies in accurately assessing user interest in an advertisement using reasonable metrics. In summary, (a) CPI advertising refers to an advertisement being considered to have gained a certain level of user attention simply because it was seen by a user on the internet; (b) CPC advertising, on the other hand, assumes that user clicks must be confirmed to gain internet user attention; (c) CPA advertising judges user interest based on whether the advertising effect or event meets specific criteria, such as actual purchase behavior or participation in a marketing campaign; and (d) targeted advertising is an advertising strategy designed to increase conversion rates by serving ads to internet users who are likely to be interested in the advertising content, rather than serving ads to an unspecified group.

[0013] From an advertiser's perspective, another aspect to consider is profit. For example, a 2016 US survey found that if a seller had a 30% profit margin at full price and now offers a 10% discount, their actual profit margin would drop by more than 50%. In other words, even if a 1% to 2% discount may seem insignificant to buyers, for sellers, promotional marketing can have a significant impact on company profits in certain situations. If promotional marketing is ineffective, both advertisers and sellers will inevitably suffer negative financial consequences due to increased advertising costs and reduced profits.

[0014] In short, we need a suitable metric to reasonably determine whether customers are interested in advertising; in addition, we need to understand how much marketing cost can be invested to successfully attract consumer interest while maximizing advertiser profits. Summary of the Invention

[0015] Technical issues

[0016] As mentioned earlier, advertisements that fail to capture internet users' attention can be considered cyber pollution, while those that do capture attention not only provide users with useful online information but also improve advertising efficiency from the advertiser's perspective by boosting sales. However, whether traditional online advertising techniques can accurately measure user interests is debatable.

[0017] Traditional CPI advertising technology assumes that internet ads will appear on websites that attract users' attention, but internet users do not know whether these displayed ads are useful information or just one of many irrelevant pieces of information in cyberspace.

[0018] Pay-per-click advertising, exemplified by keyword search ads, can generate very high conversion rates for advertisers who rank highly in search results for specific keywords. However, for most advertisers with lower rankings, the advertising effectiveness may be unsatisfactory compared to the cost. For example, if a user searches for a specific keyword in a search engine, they may need to browse multiple web pages or wait a long time to see the ad.

[0019] For example, CPA advertising, also known as affiliate marketing, involves setting up dedicated ad slots on websites or even gaming apps, requiring users to complete tasks set by the advertiser (such as participating in activities) in exchange for rewards like in-game currency. In this case, it's difficult to determine whether users are participating in affiliate marketing out of interest in the rewards (such as in-game currency) for completing the affiliate marketing conditions, or simply because they are interested in the ad itself.

[0020] In the case of online targeted advertising, if too much information is used for customer targeting, consumers may feel averse to targeted advertising rather than interested in it, and may feel that their privacy has been violated.

[0021] Furthermore, neither CPI, CPC, CPA, nor targeted advertising provide definitive answers on how to optimize advertisers' profits. This means advertisers lack information on how much promotion they need to send to internet users to achieve successful digital marketing. Ultimately, advertisers become the sellers, and issues like profit margins are likely to be resolved by the advertisers themselves, rather than by advertising agencies or portals.

[0022] This invention aims to solve the aforementioned problems, but the key lies in the fact that none of the various internet advertising technologies mentioned above attempt to more accurately determine whether a user is interested in internet advertising, nor do they attempt to further stimulate the interest of users who have already expressed interest. If it is possible to accurately grasp the interests of internet users and dynamically adjust the ad display accordingly, thereby making only users interested in internet advertising more interested (and thus improving advertising efficiency), then it can be said that the inspiration for this invention comes from the concept of solving all or at least some of the aforementioned problems.

[0023] The second key technical problem recognized by this invention is how to maximize corporate profits while guiding user purchasing behavior to improve conversion rates, thereby finding a balance point for the benefits that advertising should provide. Therefore, this invention proposes a method to increase advertisers' profits by tracking user reactions to digital advertisements that have attracted internet users' attention but have not yet converted into purchasing behavior. Specifically, this invention will elaborate on this method in more detail.

[0024] Furthermore, the technical problem this invention aims to solve is not limited to internet advertising generated by traditional personal computers (PCs). Instead, this invention focuses on advertising delivery to users of information technology (IT) or information and communication technology (ICT) devices, collectively referred to as "smart devices," such as laptops, smartphones, and tablets equipped with cameras. Considering the high-speed computing power, ample storage capacity, and robust network environment of these smart devices, it is believed that the aforementioned advertising technical problems can be solved to some extent.

[0025] Problem-solving methods

[0026] The present invention aims to solve all or some of the above-mentioned problems. In order to solve the first technical problem, the present invention proposes a technology that can accurately grasp the interests of Internet users and dynamically adjust the display or exposure mode of advertisements accordingly.

[0027] In other words, according to a first aspect of the present invention, a method for transmitting internet advertisements from an advertising server to a user terminal equipped with a camera for advertising includes: a first step of receiving an internet advertisement from an advertiser's terminal; a second step of transmitting the received internet advertisement to the user terminal, causing the internet advertisement to be published and displayed on the display screen of the user terminal; a third step of collecting image or video data containing the user's face through the user terminal and the camera, for processing by an artificial intelligence analysis tool installed on the user terminal or the advertising server; a fourth step of having the artificial intelligence analysis tool extract information related to the user's gaze from the collected image or video data, and determining whether the user's gaze overlaps with the advertising display area where the internet advertisement is located based on the user's gaze information; if it is determined that the user's gaze is pointing towards the advertising display area, then an internet advertisement method using artificial intelligence gaze analysis is proposed, characterized by including a fifth step, namely, changing the display method of the internet advertisement on the user terminal.

[0028] According to the advertising method of the second aspect of the present invention, the improved exposure method or advertising display mode is characterized by: including a method for increasing the size of the internet advertisement publication and exposure portion, or a method for changing the position of the internet advertisement publication and display on the viewing screen.

[0029] According to the advertising method of the third aspect of the present invention, the artificial intelligence analysis tool includes a sophisticated computer vision artificial intelligence analysis tool that can process images, videos or all of the above, and the criteria for determining the overlap between the advertising display area and the user's line of sight in the fourth step are: (a) whether the measured line of sight duration exceeds a predetermined threshold; (b) whether there is an advertising display area on the scanning path; and (c) whether there is a display area within the convex hull created by the scanning path, the convex hull being composed of at least one or more combinations thereof.

[0030] According to the advertising method of the fourth aspect of the present invention, the artificial intelligence analysis tool may include a step in which the artificial intelligence analysis tool generates a user gaze map based on the user's gaze information on the viewing screen, and determines whether the advertising display area and the user's gaze area overlap.

[0031] The fifth aspect of the invention relates to a system embodying the advertising method described in the first aspect above. The fifth aspect is an advertising server system that sends internet advertisements to a user terminal equipped with a camera for placing advertisements, receives internet advertisements from an advertiser's terminal, and then sends the received internet advertisements to the user terminal for display on the user terminal's screen advertising display area. Furthermore, through the user terminal and the camera, the system includes a database for collecting, classifying, and storing image or video data, including user facial data, for processing by an artificial intelligence analysis tool installed on a main server. The AI ​​analysis tool extracts information related to the user's gaze from the collected image or video data and determines whether the user's gaze overlaps with the advertising display area where the internet advertisement is located. When it is determined that the user's gaze is pointing towards the advertising display area, the main server changes the display method of the internet advertisement on the user terminal.

[0032] Next, the sixth aspect of the present invention focuses on the second technical challenge of the present invention mentioned above, which includes: a first stage in which an advertising server receives an internet advertisement from an advertiser's terminal; a second stage in which the received internet advertisement is transmitted to a user terminal equipped with a camera, and the internet advertisement is published and displayed in a predetermined advertising display area on the user terminal's screen; a third stage in which image or video data containing the user's face is collected through the user terminal and the camera, so that it can be processed by an artificial intelligence analysis tool installed on the user terminal or the advertising server; a fourth stage in which the artificial intelligence analysis tool extracts information related to the user's gaze from the collected image or video data, and judges whether the published and displayed internet advertisement has attracted the user's attention based on the user's gaze information according to a preset judgment standard; if it is determined that the published and displayed internet advertisement has attracted the user's attention, then a digital advertising method based on artificial intelligence gaze analysis is proposed, characterized by including a fifth stage, namely, real-time enhancement of the benefit level contained in the internet advertisement.

[0033] A seventh aspect of the present invention relates to a system embodying the sixth aspect described above, comprising: a main server that receives internet advertisements from an advertiser's terminal, transmits the received internet advertisements to a user terminal equipped with a camera, and publishes and displays the internet advertisements in a predetermined advertising display area on the user terminal's screen; and a database that collects and stores image or video data containing the user's face through the user terminal and the camera, wherein the main server interacts with the database to execute an artificial intelligence analysis tool installed on the main server, the artificial intelligence analysis tool extracting information related to the user's gaze from the collected image or video data, and judging, based on the user's gaze information and a preset judgment criterion, whether the published and displayed internet advertisements have attracted the user's attention. If it is determined that the published and displayed internet advertisements have attracted the user's attention, a digital advertising system based on artificial intelligence gaze analysis is proposed, characterized by taking real-time action to enhance the level of benefits contained in the internet advertisements.

[0034] Invention Effects

[0035] First, according to the first to fifth aspects of the present invention described above, the common core of the present invention is to accurately grasp the interests of Internet users and dynamically adjust the display of advertisements accordingly.

[0036] In other words, since smartphones and other smart devices are equipped with cameras, this invention proposes a technology that interprets a user's gaze at an advertisement as their interest in the advertisement by determining which part of the smartphone screen a user is looking at when accessing a website, and whether that part displays an advertisement. In short, this invention provides a reasonable and technical solution for analyzing user interests (the core of internet advertising) by applying artificial intelligence technologies such as eye tracking and gaze detection to internet advertising.

[0037] This invention determines a user's interest by recognizing their gaze, thus understanding whether a user is interested in an advertisement more accurately than the traditional CPI advertising method, and consequently calculating advertising costs more rationally.

[0038] The main difference between this invention and the CPC method lies in that this invention analyzes user interests before the user clicks. Specifically, this invention proposes a method to change the way advertisements are displayed. When it is determined that an advertisement has potential user interest, the display method is altered by enlarging the advertisement size or changing its position. This can be understood as a technical means to guide users to make conversion behaviors such as clicking; in other words, it enhances user interest in this way.

[0039] However, this invention is not intended to create separate ad pages for affiliate marketing (such as CPA advertising). This invention proposes an advertising technology different from traditional affiliate marketing or CPA advertising, namely, using artificial intelligence analysis to push ads that users are interested in based on their gaze direction. Even compared to targeted advertising, this invention can be considered an AI advertising technology, as it does not require pre-collection of internet user information matching the target audience's characteristics, nor does it require pre-selection of the target advertising audience; it can be applied to any internet user using a smart device.

[0040] Specifically, this invention proposes using an artificial intelligence analysis tool called computer vision, capable of processing images, videos, or both, to analyze internet users' interests. As mentioned earlier, this invention focuses on internet advertising targeting users of smart devices equipped with cameras. In other words, this invention utilizes cameras installed on smart devices to analyze the screen area currently being viewed by the user, and by accurately identifying the user's interests, formulates corresponding technical measures to enable advertisers to accept the rationality of the user interest measurement method provided by this invention. Specifically, this invention proposes a method for reasonably judging user interests using artificial intelligence technology, such as the duration of the user's gaze, whether the advertisement appears on the scanning path of the user's gaze, and whether internet advertisements exist within the area formed by the scanning path.

[0041] Furthermore, this invention can dynamically change the display method of internet advertisements based on user gaze analysis using artificial intelligence, and generate or pre-generate a gaze map (gaze map) for each webpage in real time. It can also analyze whether the placement of the advertisement on the smartphone screen is ideal based on the gaze map.

[0042] If the AI-powered gaze analysis according to the present invention is performed in real time on the advertising server, the display method of the advertisement can be changed in real time according to the actual situation, thereby maximizing advertising efficiency. In particular, if by changing the display method of existing online advertisements, it is confirmed that the user has successfully clicked, thereby gaining more user attention, and if a click has indeed occurred, then by storing the successful click history in real time and feeding it back to the advertiser's terminal, an advertising cost strategy that satisfies the needs of both the advertiser and the advertising agency can be formulated.

[0043] In short, the first to fifth aspects of the present invention not only improve the efficiency of advertising cost execution from the advertiser's perspective, but also filter out advertisements that do not attract the attention of Internet users and advertisements that do attract the attention of Internet users through artificial intelligence, so that the advertisements presented according to the present invention are regarded as useful information by users, thereby realizing dynamic advertising exposure and improving the advertising effect of advertising materials that are useful to users.

[0044] According to the sixth and seventh aspects of the present invention, the present invention proposes a digital advertising method that uses artificial intelligence eye recognition technology to select digital advertisements that attract the attention of Internet users, and observes in real time the degree of promotion of the advertisements by Internet users, thereby observing the degree of positive response of Internet users to the advertisements.

[0045] The effects of the sixth and seventh aspects of this invention are as follows: In other words, by observing in real time the increasing benefits of the advertisements displayed to the user's current gaze, the ad server can understand the level of benefit required for the user to transition from a simple gaze-based advertisement to a purchase decision. This accumulated data is expected to greatly help advertisers determine how much benefit should be provided to attract potential consumers to make a purchase decision. Thus, the artificial intelligence eye-tracking analysis according to this invention can be performed in real time on the ad server, thereby enabling real-time adjustment of the benefits offered by the displayed advertisements and more effectively guiding user behavior conversion. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of an internet advertising system utilizing AI gaze analysis according to an embodiment of the present invention.

[0047] Figure 2 This is an exemplary configuration diagram of AI software according to an embodiment of the present invention, which can be used for AI gaze analysis.

[0048] Figure 3This is a schematic diagram of the operation mode of computer vision AI software according to an embodiment of the present invention.

[0049] Figure 4 and Figure 5 This is a schematic diagram illustrating an embodiment of the present invention, showing how computer vision AI software analyzes image data and related line-of-sight data in real time.

[0050] Figure 6 This is an example of how, according to an embodiment of the present invention, AI is used to analyze a user's gaze and synthesize it into a map based on the scanning path (i.e., the movement path of the gaze) of the saccade (i.e., the instantaneous movement of the eyeball) with a specific website.

[0051] Figure 7 This is an example diagram illustrating a heat map generated by analyzing a user's gaze using AI according to an embodiment of the present invention, based on the period of eye focus fixation (the moment when the eyeball is relatively fixed).

[0052] Figure 8 This is a diagram illustrating an example of changing the exposure method of advertisements that a user is interested in after completing AI analysis of the user's gaze, according to an embodiment of the present invention.

[0053] Figure 9 This diagram illustrates another example of how, according to an embodiment of the present invention, the exposure method of advertisements that a user is interested in is changed after AI analysis of the user's gaze is completed.

[0054] Figure 10 This is an example diagram illustrating how, according to an embodiment of the present invention, the offering strength provided to users is increased in real time based on AI analysis results. Detailed Implementation

[0055] The preferred embodiments of this disclosure will now be described in detail with reference to the accompanying drawings. Figures 1 to 4 Examples of various coupons, or objects that can be considered coupons, are shown for artificial intelligence analysis according to the present invention.

[0056] A schematic diagram of an internet advertising system 500 employing AI gaze analysis technology according to an embodiment of the present invention. Figure 1 As shown, an internet advertising system 500 employing AI gaze analysis technology (hereinafter referred to as the "AI advertising system") according to an embodiment of the present invention consists of an AI advertising server and a database 100. Figure 2As shown, the AI ​​software 200 according to the present invention can be integrated with the AI ​​advertising server and database 100. In order for the AI ​​advertising server and database 100 to operate as an advertising agency, it needs to interact with the advertiser's terminal 300 and the client or consumer's terminal 400 (e.g., the smartphone of the Internet user who will be exposed to the advertisement; hereinafter referred to as "client") via wired or wireless networks.

[0057] Meanwhile, the camera 410 is briefly described below. For example, in recent years, the camera 410 of a smartphone 400 (including a front-facing camera and a rear-facing camera (not shown)) typically includes a wide-angle camera 411, a telephoto camera 412, and an ultra-wide-angle camera 413, capable of high-quality shooting at 12 megapixels (MP). Furthermore, these cameras 411, 412, and 413 are often surrounded by other small devices, such as a flash 414, a LiDAR (LiDAR, not shown) scanner, and a microphone for noise and background noise cancellation during video recording. In this invention, for convenience, various types of cameras (411, 412, 413) or infrared cameras that can be connected to smart devices may be used. They are collectively referred to as "camera 410".

[0058] Refer again Figure 1 In step S1, the AI ​​advertising server and database 100 receive online advertisements from the advertiser's terminal 300 according to a pre-agreed advertising fee contract. The display method of the advertisement, such as the publication cycle, publication location, and publication method (e.g., banner ad or pop-up ad), is usually predetermined in the advertising fee contract. However, it should be noted that the AI ​​advertising server and database 100 of this invention may have the authority to dynamically control the advertising display mode, preferably real-time control. Furthermore, the following... Figure 10 As shown, this invention requires the AI ​​advertising server and database 100 to dynamically and gradually enhance user benefits, including promotional benefits. Therefore, it may be necessary to consult with the advertiser 300 in advance so that the AI ​​advertising server and database 100 can adjust promotional benefits in real time, thereby applying the key features of this invention to online advertising.

[0059] For reference, if the AI ​​advertising server and database 100 executes the AI ​​software 200 in real time, the AI ​​advertising server and database 100 will be equipped with a central processing unit (CPU), a graphics processing unit (GPU), a field-programmable gate array (FPGA), and various accelerator cards to improve the hardware's computing speed and performance. Such an advanced AI system is particularly ideal for handling complex AI calculations in real time.

[0060] In step S2, the AI ​​advertising server and database 100 send internet advertisements to the client terminal 400 according to the advertiser's request 300. (Reference) Figure 8 and Figure 9 "Dress on sale! Click here to buy!" (Clothing on sale, click here to buy!) Internet ads will be placed in specific ad slots. Display area 421 on the screen 420 of smartphone 400. Reference Figure 10 The digital banner ad 920 sent from step S2, named “10% Coupon”, will be placed in another specific advertising area 422 within the screen 420 of the smartphone 400.

[0061] exist Figure 1 In step S3, the AI ​​advertising server and database 100 receive image or video data captured by the camera 410 of the client terminal 400 in real time. Of course, to improve advertising efficiency, the user's personal information consent must be obtained. Information regarding camera captures and external data transmission to the AI ​​advertising server and database 100 is pre-obtained from the client terminal 400. Ideally, the image or video data transmitted in step S3 should be facial data, at least including the user's eyes. Although AI can filter out images containing non-facial objects or backgrounds, according to the present invention, such information should not be discarded as long as it may be useful for eye and gaze analysis.

[0062] exist Figure 1 In step S3, in addition to the user's facial information, the AI ​​advertising server and database 100 can also transmit and collect various client-side data. For example, the AI ​​advertising server and database 100 may include personal information provided by the search engine operator and client terminal 400 with the user's consent, such as the user's website visit history and content browsing history. In this way, the AI ​​advertising server and database 100 can collect and store tracking information (hereinafter referred to as "client-side tracking information") of the user's use of the application or online activities on the client terminal 400, thereby more accurately summarizing advertising effectiveness and analyzing whether internet advertising is related to specific user actions. The client-side tracking information collected and stored in this way may also become the object of analysis for the artificial intelligence software 200, which will... Figure 2 Detailed description is provided.

[0063] In step S4, AI gaze analysis is performed, which will... Figure 2 and Figure 3 A detailed explanation follows. Although the physical object being analyzed is... Figure 1The AI ​​software 200 is shown as an AI advertising server and database 100, but it can be installed on another external server (not shown) that can connect to the AI ​​advertising server and database 100 and run step S4, or it can run AI gaze analysis on the client terminal 400 after obtaining user consent and only transmit the analysis results to the AI ​​advertising server and database 100.

[0064] Next, in step S5, the AI ​​advertising server and database 100, with the consent of the advertiser, determine and execute an advertising display method that can attract the attention of more client terminal 400 users, based on the user gaze analysis results performed in step S4.

[0065] Reference Figure 8 and Figure 9 By analyzing user gaze using artificial intelligence software 200, it can be determined that the user is paying attention to the words "Dress on Sale! Click here to buy!" displayed in the advertising display area 421, which is the location of the online advertisement 900 that has already been published. Now, the artificial intelligence advertising server and database 100 can take various measures to attract more attention from users who have already viewed the online advertisement.

[0066] First, in order to attract more user attention, the AI ​​ad server and database 100 may replace the original ad 421 with a newly created amplified ad 900, such as... Figure 8 As shown. This newly created enlarged advertisement 900 can be placed at the exact location of the original advertisement 421, or according to, for example... Figure 6 and Figure 7 The eye-tracking graph is placed in other ad display areas (not displayed) to increase the likelihood of users clicking on the ad.

[0067] In addition, such as Figure 9 As shown, the AI ​​advertising server and database 100 can generate virtual hand images 910 using generative AI technology to increase the likelihood of users clicking on the advertisement. In other words, if a user is very interested in the advertisement, other means of inducing click behavior (such as hand images or animated GIFs related to clicking) can be added to the published advertisement 421. It should be noted that in this invention, if the AI ​​determines that the user has expressed interest, it will provide further means to attract the user's attention (such as 900 or 910). Because the user has already viewed the published advertisement 421, such marketing means (900 or 910) may be more effective in enhancing user interest without affecting the user's internet experience.

[0068] Finally, in step S6, the AI ​​advertising server and database 100 can report the results obtained through steps S1 to S5 to the advertiser's terminal 300. Of course, the user needs to consent to providing personal information to a third-party company in order to generate reports that may be valuable and meaningful to the advertiser. If the user consents, customer tracking information, AI analysis results, and advertising performance data may be reported to the advertiser. Based on these reports, the advertiser's terminal 300 can obtain assistance in developing future advertising strategies, such as creating new online advertisements or developing new marketing strategies, and then return to step S1, where the advertiser's terminal 300 will be able to request online advertisements from the AI ​​advertising server and database 100 again. Furthermore, as will be described later, step S6 includes determining whether sending a new advertisement (e.g., 930) to the advertiser's terminal 300 will lead to a user purchase, such as... Figure 10 As shown.

[0069] Figure 2 This is an exemplary configuration diagram of AI software 200 according to an embodiment of the present invention, which can be used for AI gaze analysis.

[0070] Artificial intelligence software 200 typically involves various processes, such as: (a) problem definition; (b) data collection and preparation; (c) model development and training; (d) model evaluation and optimization; (e) deployment of the AI ​​into a real product; and (f) execution of machine learning operations. Since these processes are not entirely independent but interconnected, to effectively assist the AI ​​computation process, it may be necessary to connect to external servers and AI advertising servers and database 100, rather than limiting the AI ​​analysis of this invention to a physically singular location. Furthermore, as mentioned above, according to other embodiments of the invention, at least a portion of the artificial intelligence software 200 can also be configured to run on a client terminal 400.

[0071] According to one embodiment of the present invention, artificial intelligence software 200 may include a generative artificial intelligence tool 210. The generative artificial intelligence tool 210 is responsible for receiving training data input and creating similar text, images, or media based on the patterns and structures of the input training data. For example, in this invention, it is possible to generate... Figure 9 Various formats (e.g., images, text, and videos) are used to attract user attention (e.g., 910). The generation process can be assisted by generative AI tool 210. In this case, generative AI tool 210 is particularly important for advertisers because, as mentioned above, attracting more user attention (e.g., 910) ultimately aims to improve conversion rate (CVR). On the other hand, for example, as will be discussed later... Figure 10In this context, generative AI tools 210 may be particularly important to advertisers because the new ad creatives created by generative AI tools 210 (e.g., Figure 10 The 930 (or 930) can provide real-time insights into the impact of changes in marketing benefits on user purchasing behavior.

[0072] Furthermore, the present invention also includes a large language model (211, Large Language Model; LLM) as a sub-model. Additionally, since internet advertising frequently appears in videos created by others, the present invention also includes a multimodal base model 212 (Multimodal Base Model; MFM) in the generative artificial intelligence tool 210.

[0073] Figure 2 The AI ​​software 200 is based on machine learning (ML) tools 220. For example, a large amount of selfie data stored in the user's smartphone 400 can serve as important machine learning training data for the AI ​​software 200 to analyze the user's gaze. In addition, by collecting facial data, head posture data, and gaze data of the user in various situations (such as when the user is wearing glasses or a mask, or when the user's perspective is different from usual), the accuracy of user gaze analysis can be improved.

[0074] According to one embodiment of the present invention, the machine learning tool 220 may employ a deep learning model 221, which analyzes given data by dividing it into multiple layers; it may also employ a supervised learning model 222, which performs learning modeling so that the results of user gaze analysis can output the values ​​of the interest-inducing method in the form of reference numbers (900 or 910). In other words, this method does not drive machine learning by pre-inputting expected values ​​to determine the desired results. Alternatively, an unsupervised learning model 223 may be employed, which improves the accuracy of user gaze recognition by simulating training data on a neural network and conducting repeated trials using artificial intelligence software 200, without providing the aforementioned training label data.

[0075] Furthermore, according to one embodiment of the present invention, the artificial intelligence software 200 may include a natural language processing tool 230. Admittedly, data related to the user's face and eyes is most important in the eye analysis of the present invention, but it is also crucial to be able to filter out information such as text and background information on the user's clothing during eye analysis.

[0076] Natural Language Processing Tool 230 is a Natural Language Understanding (NLU) model 231. Natural Language Generation (NLG) model 232 may be particularly useful in artificial intelligence analysis (i.e., step S4 described above) and step S5. For reference, since the AI ​​advertising system 500 of the present invention is not targeted at any specific country, the translation function of the AI ​​software 200 of the present invention may be very important, even though it seems unrelated to facial recognition. For example, if a traveler from South Korea to France uses a smartphone 400 to browse information in front of a museum in France, the images or text information about promotional activities at the local French museum captured by the camera 410 of the smartphone 400 can serve as a target for AI analysis. Similarly, for example, if the new advertisement in Figure 10930 contains a foreign language, Natural Language Processing Tool 230 may be useful.

[0077] like Figure 2 As shown, the artificial intelligence software 200 of this invention must include a computer vision tool 240. This invention employs a technique that analyzes user gaze points and assesses user interest in internet advertising based on the analysis results. Therefore, the computer vision tool 240 is crucial for user gaze point analysis.

[0078] Ideally, object detection model 241 should be used, which extracts only the necessary objects for gaze analysis from images or videos, such as human eyes, nose, mouth, hair, glasses, hands, and other objects in the background. For example, in a physical store, a user might be looking at a smartphone screen 400 while comparing in-store and online prices. In this case, to accurately identify the user's gaze point reflected in the smartphone camera 410, it is necessary not only to identify the user's face but also to distinguish objects in the background, rather than the user themselves.

[0079] Scene understanding model 242 is also one of the AI ​​models available in computer vision tool 240. This model 242 uses AI analysis to determine which objects in an image or video are more important, and which objects have specific importance or priority. From a machine's perspective, a user's face (composed of only pixels) can be subdivided into various objects such as eyes, nose, mouth, pupils, irises, and eyelids using scene understanding model 242, thereby calculating which objects should be considered important in gaze analysis.

[0080] Figure 2The face detection and recognition model 243 shown is essentially a technology for finding and recognizing faces in digital images or videos. Artificial intelligence technology applied to social media, photo management, facial recognition security access control, and even criminal investigations can infer the age (e.g., pupillary reflexes darken with age), gender, and emotion of the analyzed individual based on facial expressions and appearance. In other words, according to the present invention, the user facial data collected by camera 410 is the object of analysis for the aforementioned face detection and recognition model 243. This model combines the analysis results with the user gaze analysis results described below, thereby enabling a deeper analysis of the user's interest in advertising materials. For example, the artificial intelligence advertising server and database 100 can detect changes in a user's facial expressions after viewing an advertisement and provide feedback to the advertiser 300 regarding whether the advertisement content needs to be modified.

[0081] on the other hand, Figure 2 The eye-tracking model 244 shown is a key artificial intelligence component in this invention. This eye-tracking model 244 can be divided into two sub-domains: determining the position of the eyes (“eye localization”) and determining the direction of the gaze (“gaze estimation”). For reference, in AI-based eye analysis, “eye” mainly refers to the pupil (including dark and bright pupils) and iris. In addition to pixel data of the pupil and iris, eye analysis also uses image or video information related to corneal reflection (iris reflection), limbus, pupil outline, and eyelids. In other words, eye localization focuses on accurately determining the presence and position of the human eye in a given image or video, while gaze estimation focuses on analyzing the position of the eyes in each frame of the image or video and finding the current direction of the gaze and the direction of movement of the gaze in three-dimensional space.

[0082] Due to the aforementioned technical differences, it is difficult to equate eye-tracking models with eye-tracking technology in principle. However, from the perspective of oculometry, these two models can be combined and termed an eye-tracking and gaze-tracking model244, such as... Figure 2 As shown. In this invention, it is considered that the integrated representation of the eye-tracking and gaze-tracking model 244 is sufficient.

[0083] For reference, during the eye-tracking stage, artificial intelligence can also consider the position and posture information of the human head. Furthermore, even within the eye area, color can vary slightly due to the surrounding environment; the influence of the surrounding environment on eye color in images is currently being studied. Therefore, this invention proposes a comprehensive analysis of information such as the face, facial expressions, body and head posture, glasses (especially requiring correction of eye pixel data distortion caused by thick magnifying glasses), and the background screen in the camera, rather than simply performing artificial intelligence analysis on the eyes. As I have continued to emphasize, because this invention uses artificial intelligence gaze analysis to determine whether a user is interested in internet advertising, it can... Figure 2The various model combinations shown are used to more accurately determine whether a user is interested.

[0084] Not hard to understand Figure 2 The computer vision tool 240 includes a motion analysis model 245. For example, depending on the head position, eyelids may sometimes appear as a straight line and sometimes as an ellipse in the same image; therefore, defining parameters such as the eyelid shape must be a certain specific shape can lead to errors in gaze analysis. The motion analysis model 245 is a technique that determines the type of action performed in the captured images based on two or more consecutive image sequences taken by a camera or high-speed camera. Such motion analysis data may be crucial for the accurate calculation of the eye-tracking model 244.

[0085] Finally, computer vision tools 240 can be used to analyze text recognition model 246; for example, optical character recognition (OCR) model 246 can also serve as an auxiliary means. If the background of a user video captured by a smartphone camera 400 contains a large amount of text, the OCR model can serve as an auxiliary means to the aforementioned scene understanding model 242, providing some assistance for artificial intelligence analysis.

[0086] Enter Figure 3 Before proceeding, it is necessary to clarify the difference between the artificial intelligence software 200 of this invention and the recently popular so-called "Chat GPT". ™ The difference between "". For example Figure 2 As shown, according to the present invention, the component 200 for analyzing user gaze may include a generative artificial intelligence tool 210. However, this does not mean that the artificial intelligence advertising technology of the present invention is completely unrelated to systems like Chat GPT (e.g., systems where users ask questions to artificial intelligence in an AI chat window and Chat GPT provides the result value). Unlike Chat GPT, the present invention does not include the process by which users think about the best question to ask the artificial intelligence to obtain the desired result (i.e., the optimization process) as a component. Therefore, changes in artificial intelligence gaze analysis technology or corresponding advertising display methods are unrelated to Chat GPT.

[0087] Figure 3 It is Figure 1 The operation mode of AI software 200 equipped with computer vision tool 240 according to an embodiment of the present invention is shown.

[0088] like Figure 3As shown, the image or video data captured by the camera 410 of the smartphone 400 is used as the initial input value of the computer vision tool 240 through step SS1. The image or video data is input to the eye tracking module 240a and the eye detection module 240d through steps SS2 and SS4, respectively. In step SS3, the eye tracking module 240a identifies pixel data related to the pupil and iris in the image or video, and collects pixel data related to corneal reflection, limbus, pupil contour, and eyelid, performing eye localization to accurately determine the presence and position of the human eye in a given image or video. The eye position information obtained from step SS3 can be transmitted to the eye detection module 240d along with the camera data transmission from step SS4. The eye detection module 240d feeds back the initial (and subsequent) eye position detection results identified by the eye detection module 240d in step SS5 to the eye tracking module 240a, thereby enabling real-time cyclic correction of the continuously input image or video data from the camera 410.

[0089] On the other hand, the head pose module 240b analyzes the head pose information in the image or video from the camera 410 (step SS7) and sends the analysis results to the gaze point determination module 240c in step SS8. The eye tracking module 240a also transmits eye position-related data to the gaze point determination module 240c from step SS6. The gaze point determination module 240c combines the eye position information received from the eye tracking module 240a and the head pose information received from the head pose module 240b to perform gaze point analysis to determine the position of the user's pupil in three-dimensional space (step SS10). The final analysis results of the gaze point detection module 240c and the eye detection module 240d (i.e., the gaze point coordinates of the gaze point detection module 240c and the eye position coordinates of the eye detection module 240d) are transmitted to the application module 100a, which exists in the AI ​​advertising server and database 100, and then the application module 100a executes... Figure 1 Step S4. In other words, in step S5, application module 100a is responsible for organizing, classifying, and transmitting the user gaze point analysis results to the AI ​​advertising server and database 100 before the AI ​​advertising server and database 100 determine and execute the advertisement. The display scheme aims to attract the attention of more users on the client terminal 400. In addition, application module 100a will determine whether the user is currently looking at the existing advertisement (e.g., based on the analysis results of artificial intelligence software 200) Figure 10 (Number 920 in the text), and based on this, decide whether to generate a new advertisement (e.g., Figure 10 (Number 930 in the text). In order to objectively determine whether a user is viewing an existing advertisement (e.g., ... Figure 10 (Number 920 in the table below) It is recommended to establish the mathematical standards shown in Table 1 below.

[0090] This invention provides a schematic diagram illustrating real-time analysis of image data and related gaze data by computer vision AI software, according to one embodiment of the present invention. Figure 4 and Figure 5 In this context, reference numeral 600 refers to a camera video screen 600, for example, which contains a user being identified in real time by the camera 410 of a smartphone 400. Figure 3 Application module 100a will utilize, for example Figure 4 and Figure 5 The results of the line-of-sight analysis are shown.

[0091] Combination Figure 4 and Figure 5 The camera image screen 600 virtually overlaps with the area 610 used for recognizing the user's face. Within this virtual face recognition area 610, as... Figure 4 and Figure 5 As shown, virtual points called keymarks 611 are generated on the facial image. These keymarks 611 move in real time according to the user's facial or eye movements. Then, using a preset calculation program, the pitch angle (the angle of head nodding up and down) and yaw angle (also known as purlin motion, i.e., the angle of head turning left and right) are calculated. For reference, the pitch angle is measured from -90 degrees to +90 degrees, based on an imaginary horizontal line passing through both ears; the yaw angle is measured from -180 degrees to +180 degrees, based on an imaginary vertical line passing perpendicularly through the top of the head when the head is upright. These pitch and yaw angle values ​​also fluctuate in real time with human movement and can be superimposed on the camera image 600 in the same form as reference number 620, as shown. Figure 4 and Figure 5 As shown.

[0092] like Figure 4 and Figure 5 As shown, the positions of the virtual overlapping arrow 630 and point 640 change with the pitch and yaw angles. Arrow 630 represents the direction of the user's gaze, and point 640 represents a point within the user's smart device 400. Therefore, if the user is not looking at the screen 420 of the smartphone 400 at all, but turns their head in another direction, then point 640, which is within the user's line of sight, may not be displayed.

[0093] Next, Figure 6 An example of a screen 700 is shown, which, according to an embodiment of the present invention, is generated using artificial intelligence analysis of the user's gaze point and presented in the form of a scan path map, which is generated based on the instantaneous movement of the eyeballs (eye jumps). In other words, Figure 4 and Figure 5 The image projected by the camera is 600, while Figure 6 It can be viewed as a line-of-sight map, based on Figure 4 and Figure 5 The gaze analysis results shown tracked the user's gaze on the smartphone screen at a depth of 400 degrees (see, for example, see...). Figure 8 The movement trajectory on the ). In short, Figure 6 Screen 700 in the image represents the first example of a line-of-sight diagram.

[0094] like Figure 6 As shown, the user did not see the area where internet ad 710 was located at all. In other words, according to... Figure 6 In the eye-tracking diagram 600, the user's gaze starts from point P1 and repeatedly skips between points P2, P3, ..., P13 without continuing to focus on the webpage. In this case, marketing methods that force users to pay attention to internet advertisements 710, such as forced pop-ups, are not very meaningful in this invention. This is because internet advertisements 710 that fail to attract user attention may be considered web pollution.

[0095] Figure 7 This is an example of a heatmap 800 generated by AI analysis of the user's gaze point and the duration of eye focus (line of sight, i.e., the time when the eyeball is relatively fixed) according to an embodiment of the present invention. It is called a "heatmap" because if the user gazes at the area for a long time, it will show a more intense color, just like a thermal imaging camera.

[0096] in short, Figure 7 Screen 800 in the diagram can be seen as a second example of a gaze map created according to the present invention to improve advertising efficiency. Figure 7 In this context, the deeper the gaze point (i.e., the higher the heat), the longer the user's gaze lasts. For example, suppose... Figure 7 The areas with relatively high popularity are H1 and H2. We can see that area H2 overlaps with internet ad 810. In other words, Figure 7 Internet ads in 810 have already attracted users' attention. Figure 1 The AI ​​advertising server and database 100 in step S4 will be ready to proceed to step S5.

[0097] Of course, for example in Figure 7 In this context, the threshold for determining user interest can be used as a threshold for determining whether an internet advertisement has reached a certain level of visual intensity (or popularity). Figure 6 In the example, identifying an internet ad 710 as something the user is interested in simply because it appears on the user's scanning path may depend on system settings rather than the existence of an absolute standard for interest.

[0098] The table below lists the eye-related variables that can be used by the artificial intelligence software 200 of this invention, including the user's eyes and eye-tracking model 244. These variables and data can be collectively referred to as "gaze information." In this invention, not all of these variables are used as criteria for judging whether an advertisement attracts the user's attention; rather, thresholds for one or more combinations of variables can be set.

[0099] [Table 1] Examples of eye measurement variable values

[0100] For example, as shown in Table 1 above, internet advertisements (e.g., "Dress promotion! Click here to buy!"). If the AI ​​ad server or the main server and database 100 determine whether the advertisement is in the ad display area (e.g., below) Figure 8 If the reference number 421 overlaps with the scan path, the judgment criteria include: (a) whether the measured gaze duration exceeds a preset threshold; (b) whether the advertisement display area exists on the scan path; and (c) whether there is a display area in the convex hull generated by the scan path, thereby objectively judging the user's interest in the published internet advertisement. For reference, the convex hull refers to the size of the area formed by connecting the outermost vertices of the scan path. Figure 6 In this context, it will be a polygon connected in the order of P1, P13, P4, P5, P8, P2, and P3. Figure 6 In the examples shown, "Click here to buy!" In the three examples above, even according to standard (c), the "Buy Now! For Sale!" ad 710 did not attract the user's attention. The AI ​​ad server and database 100 tracked whether the user's gaze was on the ad display area 422 ( Figure 10 Overlapping, in which published Figure 10 The internet advertisement shown is for a "10% discount coupon" (920). If the user's line of sight overlaps with the area where the advertisement is displayed, one or more of the criteria described in Table 1 may apply.

[0101] Figure 8 An example of a network advertisement showing modified display size is shown. According to one embodiment of the invention, after completing AI analysis of the user's gaze, the content the user is currently interested in is determined. For example... Figure 1 As shown, in step S3, the client tracking information and the image or video data captured by the camera 410 are sent to the AI ​​advertising server and database 100 for storage. Figure 8 In the example, the advertising display area 421 displays an advertising banner that reads "Dress on sale! Click here to buy!", and the AI ​​advertising server and database 100 preferably analyze the user's gaze point on the smartphone 400 display screen 420 in real time.

[0102] For example, based on the AI ​​analysis results, if the AI ​​determines that the advertisement meets the user interest threshold defined by the variables shown in Table 1, the process can proceed to step S5. At this point, the AI ​​advertising server and database 100 will determine, based on the user gaze analysis results in step S4, the advertising display method or presentation mode that can attract more attention from the client terminal 400 users, and execute the plan in real time as much as possible. For example, if... Figure 8 As shown, if a user views the "Dress Sale! Click here to buy!" banner ad for a certain period of time, the AI ​​ad server and database 100 will immediately increase the size of the "Dress Sale! Click here to buy!" banner ad to attract more user attention. In this case, the first display mode is the original size banner ad 421, and the second display mode is the enlarged banner ad 900.

[0103] Figure 9 This is another example of how, according to an embodiment of the present invention, the way a user interacts with advertisements of interest is changed after AI analysis of the user's gaze is completed. Figure 8 The difference is that the "Dress on Sale! Click here to buy!" banner ad now includes a finger icon 910 indicating "Click here," and the AI ​​ad server and database 100 run step S5. On the other hand, as... Figure 8 and Figure 9 The database (110) shown is used to classify and store image or video data collected by camera 410 from the AI ​​advertising server and database 100; the billing server 120 is a server for billing. For example, according to the present invention, if a user changes the advertising display mode (i.e., Figure 9 The newly added icon (910) indicates that a successful click on an ad will aggregate detailed information on such pattern changes and user behavior shifts, and provide the statistics to the advertiser (300). Here, because the AI ​​system, based on complex eye-tracking and gaze analysis, increases the likelihood of users successfully clicking on ads, additional advertising fees can be set based on marketing effectiveness.

[0104] An example diagram from one embodiment of the present invention illustrates how, based on the gaze analysis results of artificial intelligence, the user benefits provided in the original internet advertisement are enhanced to a higher level in real time. (See diagram for example.) Figure 10 As shown, the user is currently viewing an ad for a "10% coupon" (advertisement 920). For reference, this conclusion is reached in real time after step S3 when the AI ​​analysis meets the predetermined judgment criteria shown in Table 1.

[0105] Subsequently, the AI ​​ad server and database 100 will send new ads 930 in real time and display them on the screen 420 of the client terminal 400. For example, using the aforementioned AI model related to language processing, the ad content will display: "Thank you for viewing this coupon! As a token of our appreciation, we are offering you a 20% discount coupon valid for 1 minute," to help users understand why the coupon discount has suddenly increased. It is important to note that when a user views the 10% discount coupon ad 920, for example, if the gaze time exceeds 3 seconds, the 10% discount coupon ad 920 will immediately switch to the 20% discount coupon ad 930. Any variable in Table 1 can be used as a criterion for determining whether the 10% discount coupon 920 should be immediately switched to the 20% (or even higher) discount coupon 930, provided that the user shows sufficient interest in the published ad. This invention anticipates that real-time changes in marketing benefits based on gaze analysis will significantly improve conversion rates.

[0106] For reference, such as Figure 10 As shown, creating a new ad 930 can be accomplished using the generative artificial intelligence tool 210 described above, or by using an ad sent and stored by the advertiser's terminal 300 as a new ad 930. Regardless of the method used, the increase in ad creative revenue may involve sensitive issues such as ad costs and profit margins, and therefore requires prior approval from the advertiser 300.

[0107] If customer behavior changes (i.e., an actual purchase of the advertised product) while a new advertisement is being displayed (930), then the above applies. Figure 1 In step S6, the additional advertising costs are billed and transmitted as success fees along with customer tracking information, and the billing server 120 will be responsible for the work related to these costs.

[0108] At this point, for reference Figures 1 to 10 We have examined in detail the advertising method, system, and related AI programs based on AI gaze analysis of this invention. As mentioned above, this invention proposes the use of an AI analysis tool called Complex Computer Vision (200, 240), which can process images, videos, or both, and is particularly useful for analyzing the interests of internet users. As previously stated, this invention focuses on internet advertising (e.g., 710, 810) targeting users of smart devices 400 equipped with cameras 410. In other words, this invention utilizes the camera installed on the smart device to analyze the screen area 420 currently being viewed by the user and thereby accurately confirms the user's interests, enabling the advertiser 300 to accept the rationality of the user interest measurement method of this invention. It is particularly noteworthy that by objectifying and quantifying the criteria for judging the existence of user interests based on gaze analysis results and introducing thresholds, the advertiser terminal 300 can highly accept the reliability of advertising efficiency, and the advertising exposure method is also more easily obtained by the consumer terminal 400, thus providing useful information.

[0109] Furthermore, the present invention is also configured to be based on user gaze points analyzed by artificial intelligence (e.g., such as...). Figure 8 or Figure 9 (As shown) Dynamically change the way internet ads are displayed, and use Figure 6 or Figure 7 The line-of-sight diagram shown serves as a standard for judging such dynamic changes, thereby enabling the determination of the line-of-sight diagram (e.g.) Figure 6 or Figure 7 The system analyzes the ideal placement of ads on the smartphone screen. The results of this analysis will be included in the feedback provided to advertiser 300 in step S6, and are expected to provide advertiser 300 with better insights into subsequent ad placements.

[0110] Furthermore, if the AI ​​gaze analysis according to the present invention is executed in real time on the AI ​​advertising server and database 100, it is expected that the display method of advertisements can be changed in real time according to the actual situation, thereby maximizing advertising efficiency. In particular, since the present invention judges the user's interest based on the user's constantly changing web browsing behavior and corresponding pupil movements, it will... Figure 1 A series of operations from steps S3 to S5 in the middle or Figure 1 The entire process shown is ideally configured to be executed in real time.

[0111] As described above, this invention uses artificial intelligence to filter out advertisements that do not attract the attention of internet users and advertisements that do attract the attention of internet users, thereby achieving dynamic advertisement exposure, improving the advertising effect of internet information that is useful to users, improving the efficiency of advertising cost execution from the advertiser's perspective, and the advertisements presented according to this invention can be identified as useful information from the user's perspective.

[0112] On the other hand, for example, in this invention, not only the display method of internet advertisements, but also the promotional effect of the advertisements themselves are adjusted in real time (e.g. Figure 10 As shown in the diagram, step S6 is expected to provide advertisers 300 with market insights that other existing advertising technologies cannot provide. In other words, if it is determined that a 20% discount rate will reduce... Figure 10 If the profit margin in the example is lower, but it can bring better sales growth, then we can return to step S1 and replace the existing ad 920 with the new ad 930 as the desired ad.

[0113] Furthermore, this invention can provide a new advertisement 930 with a higher promotional effect to all consumers who view the existing advertisement 920, or it can be piloted only with a select group of consumers, subject to restrictions set by the advertiser. Since the purpose of this invention is to provide advertiser 300 with the most appropriate market insights, therefore, although... Figure 10Although not shown in the diagram, the AI ​​advertising server and database 100 can determine whether a new advertisement can be sent before performing step S5.

[0114] As described above, this invention not only changes the way advertisements are displayed, but also proposes a method using artificial intelligence to filter advertisements that do not attract internet users' attention from those that do, thereby improving the advertising effectiveness of advertisements that provide users with useful internet information and actively guiding users to click or make purchases. This will enable advertisers to assess how much promotional offer is needed to drive a purchase.

[0115] The accompanying drawings and detailed description of this invention are not intended to limit the scope of the invention. Those skilled in the art to which this invention pertains can refer to the drawings and detailed description and make various design changes without departing from the technical concept of this invention. For example, the drawings show that the artificial intelligence software 200 performs computational operations within the artificial intelligence advertising server and database 100; however, as mentioned above, with the customer's consent, all or part of the artificial intelligence software 200 can be installed on the client terminal 400, or supported by an external artificial intelligence server 200.

Claims

1. An internet advertising method utilizing AI gaze analysis, wherein the method delivers internet advertisements to user terminals equipped with cameras via an advertising server, characterized in that, include: The first step is to receive internet advertisements from the advertiser's terminal; The second step is to send the received internet advertisement to the user's terminal so that it can be published and exposed in the user's terminal screen display area; The third step involves collecting image or video data containing the user's face through the user terminal and camera, so that it can be processed by AI analysis tools installed on the user terminal or advertising server. The fourth step involves using AI analysis tools to extract information related to the user's gaze from the collected image or video data, and then determining whether the user's gaze overlaps with the display area of ​​the internet advertisement based on the user's gaze information. as well as The fifth step is to change the way internet ads are displayed on the user's device when it is determined that the user's gaze is directed at the ad display area.

2. The internet advertising method utilizing AI gaze analysis according to claim 1, wherein, The altered exposure methods include increasing the size of the internet ad placement and exposure area, or changing the placement and exposure position of the internet ad on the viewing screen.

3. The internet advertising method utilizing AI gaze analysis according to claim 1, wherein, The AI ​​analysis tools include complex computer vision AI analysis tools capable of processing images, videos, or both. In the fourth step, the criteria used to determine whether the advertising display area overlaps with the user's gaze point consist of at least one or more of the following: (a) whether the measured gaze duration exceeds a predetermined threshold; (b) whether the advertising display area exists on the scanning path; and (c) whether the display area exists within the convex hull created by the scanning path.

4. The internet advertising method utilizing AI gaze analysis according to claim 1, wherein, AI analysis tools generate user gaze maps based on the user's gaze information on the screen, and determine whether the advertising display area overlaps with the user's gaze area based on the gaze map.

5. An internet advertising system utilizing AI gaze analysis, wherein the system sends internet advertisements to user terminals equipped with cameras for advertising delivery, characterized in that, include: The main server receives internet advertisements from advertisers' terminals and transmits the received internet advertisements to user terminals so that the internet advertisements can be published and exposed in the advertising display area of ​​the user terminal's viewing screen; as well as A database that collects, categorizes, and stores image or video data containing users’ faces through user terminals and cameras, so that AI analysis tools installed on the main server can process this data; AI analysis tools are used to extract information related to the user's gaze from collected image or video data. Based on this gaze information, the main server determines whether the user's gaze point overlaps with the display area of ​​the internet advertisement. If the server determines that the user's gaze is directed towards the ad display area, it will change the way the internet ad is displayed on the user's device.

6. A digital advertising method utilizing AI gaze analysis, characterized in that, include: The first step is for the ad server to receive internet ads from the advertiser's terminal; The second step is to transmit the received internet advertisement to a user terminal equipped with a camera, so that the internet advertisement is published and exposed in the predetermined advertising exposure area on the user terminal display screen; The third step involves collecting image or video data containing the user's face through the user terminal and camera, so that this data can be processed by AI analysis tools installed on the user terminal or advertising server. The fourth step involves using AI analysis tools to extract information related to the user's gaze from the collected image or video data, and based on the user's gaze information, determining whether the published and exposed online advertisements have attracted the user's attention according to predetermined judgment criteria. as well as The fifth step is to increase the discounts included in the online advertisement in real time once it is determined that the published and exposed online advertisement has attracted the user's attention.

7. A digital advertising system utilizing AI gaze analysis, characterized in that, include: The main server receives online advertisements from advertisers' terminals, transmits the received online advertisements to user terminals equipped with cameras, and publishes and exposes the online advertisements within a predetermined advertising exposure area on the user terminal's display screen. A database that collects and stores image or video data containing a user's face through user terminals and cameras; The main server interacts with the database to execute the AI ​​analysis tool installed on the main server. This AI analysis tool extracts information related to the user's gaze from the collected image or video data. Based on the pre-set judgment criteria, it judges whether the published and exposed online advertisements have attracted the user's attention. If it is determined that the published and exposed online advertisements have attracted the user's attention, it performs an operation to increase the discount included in the online advertisements in real time.

Images