Image delivery method and device, equipment, storage medium and program product

Abstract

The invention relates to the technical field of artificial intelligence, and particularly provides an image delivery method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring a plurality of to-be-delivered images and historical published contents of a plurality of users in a to-be-delivered area within a preset historical time period; constructing an image and keyword graph based on a plurality of target keywords determined by the plurality of to-be-released images; the edges of the images and the keyword graph have image association weights of the to-be-projected images and the target keywords; constructing a user and keyword graph based on the historical published content and the plurality of target keywords; the edge of the user and keyword graph has the user association weight of each user and each target keyword; calculating a fitness score of each user for each to-be-released image based on the image association weight and the user association weight by using a fitness scoring model; and putting each to-be-put image to the user whose fitness score meets a preset score requirement. According to the method and the device, the delivery cost can be reduced while accurate recommendation is realized.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to an image projection method, apparatus, device, storage medium, and program product. Background Technology

[0002] With the widespread adoption of the internet and smart mobile devices, the advertising market is expanding and competition is intensifying. The advertising industry is gradually shifting from traditional traffic-driven advertising to intelligent advertising. Leveraging big data and machine learning technologies, intelligent advertising systems can deeply analyze user behavior and preferences to deliver more personalized ads, thereby improving user satisfaction and ad conversion rates.

[0003] In related technologies, traditional advertising methods typically cast a wide net, failing to handle the relationship between users and ads, and thus unable to accurately target potential customer groups, resulting in wasted resources. Furthermore, traditional models face significant training challenges when dealing with large-scale, highly sparse, and noisy user-ad interaction data, making them prone to overfitting and negatively impacting the final advertising performance. Summary of the Invention

[0004] This disclosure is made in view of the above-mentioned problems. This disclosure provides an image projection method, apparatus, device, storage medium, and program product.

[0005] According to one aspect of this disclosure, an image projection method is provided, comprising: Acquire multiple images to be displayed, as well as the historical content posted by multiple users in the display area within a preset historical time period; Based on multiple target keywords determined from the multiple images to be deployed, an image-keyword graph is constructed; wherein, the edges of the image-keyword graph have image association weights between each of the images to be deployed and each of the target keywords; Based on the historical published content and the multiple target keywords, a user-keyword graph is constructed; wherein, the edges of the user-keyword graph have user association weights between each user and each target keyword; Using a pre-built suitability scoring model based on the image association weights and the user association weights, calculate the suitability score for each user for each image to be displayed; Each of the images to be displayed is delivered to the users whose suitability scores meet the preset score requirements.

[0006] According to another aspect of this disclosure, an image projection device is provided, comprising: The acquisition module is used to acquire multiple images to be delivered, as well as the historical content published by multiple users in the delivery area within a preset historical time period. A construction module is used to construct an image-keyword graph based on multiple target keywords determined from the multiple images to be deployed; wherein the edges of the image-keyword graph have image association weights between each image to be deployed and each target keyword; The construction module is further configured to construct a user-keyword graph based on the historical published content and the multiple target keywords; wherein, the edges of the user-keyword graph have user association weights between each user and each target keyword; The calculation module is used to calculate the suitability score of each user for each image to be delivered based on the image association weight and the user association weight using a pre-built suitability scoring model. The delivery module is used to deliver each of the images to be delivered to the users whose suitability scores meet the preset score requirements.

[0007] In another aspect of exemplary embodiments of this disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, the processor executing the computer program to implement the methods described in exemplary embodiments of this disclosure.

[0008] In another aspect of exemplary embodiments of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the methods described in exemplary embodiments of the present disclosure.

[0009] In another aspect of the exemplary embodiments of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the methods described in the exemplary embodiments of this disclosure.

[0010] As will be described in detail below, the image delivery method according to embodiments of this disclosure involves acquiring multiple images to be delivered and historical content published by multiple users in the delivery area within a preset historical time period; constructing an image-keyword graph based on multiple target keywords determined from the multiple images to be delivered; wherein the edges of the image-keyword graph have image association weights between each image to be delivered and each target keyword; constructing a user-keyword graph based on historical content and multiple target keywords; wherein the edges of the user-keyword graph have user association weights between each user and each target keyword; calculating the suitability score for each user for each image to be delivered using a pre-constructed suitability scoring model based on image association weights and user association weights; and delivering each image to be delivered to users whose suitability scores meet preset scoring requirements. This method can achieve accurate recommendations to multiple users in the delivery area, avoid users receiving irrelevant information, and reduce delivery costs.

[0011] It should be understood that both the foregoing general description and the following detailed description are exemplary and intended to provide further illustration of the claimed technology. Attached Figure Description

[0012] The above and other objects, features, and advantages of this disclosure will become more apparent from the more detailed description of the embodiments thereof in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this disclosure and form part of the specification. They are used together with the embodiments of this disclosure to explain the disclosure and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0013] Figure 1 A flowchart illustrating an exemplary embodiment of the present disclosure of an image projection method is shown. Figure 2 This illustration shows a structural diagram of the image and keyword graph and the user and keyword graph provided in an exemplary embodiment of this disclosure; Figure 3 A schematic diagram of the structure of the image projection apparatus provided in an exemplary embodiment of this disclosure is shown; Figure 4 A schematic diagram of the structure of an electronic device provided in an exemplary embodiment of this disclosure is shown; Figure 5 A schematic diagram of the structure of a computer system provided in an exemplary embodiment of this disclosure is shown. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this disclosure more apparent, exemplary embodiments according to this disclosure will now be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this disclosure, and not all embodiments of this disclosure. It should be understood that this disclosure is not limited to the exemplary embodiments described herein.

[0015] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0016] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below. It should be noted that the concepts of "first", "second", etc., used in this disclosure are only used to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0017] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0018] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0019] Traditional advertising methods often fail to accurately reach the target audience, resulting in unsatisfactory advertising effects. Intelligent advertising delivery systems can utilize big data and machine learning technologies to analyze user characteristics, gain a deeper understanding of user behavior and needs, and accurately push advertisements to the target audience, thereby improving ad exposure, click-through rates, and conversion rates, and ultimately optimizing advertising effectiveness.

[0020] With the diversification of consumer demands and information overload, users struggle to find the ads they need from a sea of ​​information, and the advertising industry needs to adapt to this personalized demand. Intelligent ad delivery systems can understand users' unique needs based on their historical behavior and interests, providing them with personalized, more relevant, and higher-value advertising services, thereby increasing user acceptance and engagement. Furthermore, intelligent ad delivery systems can help advertisers achieve more precise ad targeting and better implement ad placements on social media to reach a wider audience. Therefore, in today's technologically advanced world, ad delivery systems remain crucial and have significant potential for improvement.

[0021] In related technologies, advertising delivery methods mainly include the following two types: The first type is collaborative filtering: This method is mainly divided into ad-based collaborative filtering and user-based collaborative filtering. These systems rely on explicit user votes and implicit interactions with ads. In the past, memory-based K-nearest neighbor algorithms and more recently matrix factorization methods have been used to build collaborative filtering recommendation systems. Matrix factorization directly extracts the latent vectors of ads and users from the ad-user matrix, capturing the communication information between items and users. However, matrix factorization cannot handle the complex relationships between users and ads.

[0022] The second method is based on predictive models for ad delivery: This involves acquiring ad delivery target data and environmental data; inputting the target data into a delivery parameter prediction model to obtain delivery parameter prediction data; inputting the delivery parameter prediction data and environmental data into an environmental learning model to obtain delivery target prediction data; inputting the delivery target prediction data into a delivery parameter optimization model to obtain delivery parameter optimization data; and finally, delivering ads based on the optimized delivery parameter data. While this method incorporates multi-factor analysis to some extent, it still struggles to fully integrate the deep semantic relationships between ad content features and user interests, resulting in insufficient accuracy and adaptability in real-world scenarios.

[0023] Both of the above methods have the following disadvantages: (1) Inaccurate targeting: Traditional advertising methods usually cast a wide net, which cannot handle the relationship between users and advertisements, and cannot accurately target potential customer groups, thus causing waste; (2) There are problems such as large data volume, data sparsity, and data anomalies; In terms of model selection, it is difficult to train complex models, and it is easy to produce overfitting phenomenon. The trained model generally performs poorly.

[0024] Therefore, in order to solve the above problems, this disclosure provides an image delivery method, which aims to achieve personalized advertising recommendations, avoid users receiving irrelevant information, facilitate advertisers to identify user preferences, accurately delineate the areas where advertisements need to be delivered, and support multi-area selection and delivery; moreover, it can push accurate and suitable advertising content to different users with high delivery accuracy.

[0025] The image projection method provided in this disclosure can be executed by a terminal or by a chip applied to the terminal.

[0026] For example, the terminal may include one or more of the following: mobile phone, tablet computer, wearable device, in-vehicle device, laptop computer, ultra-mobile personal computer (UMPC), netbook, PDA, and wearable device based on augmented reality (AR) and / or virtual reality (VR) technology. The exemplary embodiments disclosed herein do not impose specific limitations on these.

[0027] Figure 1 A flowchart illustrating an exemplary embodiment of the image projection method provided in this disclosure is shown, such as... Figure 1 As shown, the image projection method includes: S101, acquire multiple images to be delivered, as well as the historical content published by multiple users in the area to be delivered within a preset historical time period; S102, construct an image-keyword graph based on multiple target keywords determined from multiple images to be deployed; wherein, the edges of the image-keyword graph have image association weights between each image to be deployed and each target keyword; S103, construct a user-keyword graph based on historical published content and multiple target keywords; wherein, the edges of the user-keyword graph have user association weights between each user and each target keyword; S104. Using a pre-built suitability scoring model based on image association weights and user association weights, calculate the suitability score for each user for each image to be delivered. S105, deliver each image to be delivered to users whose suitability scores meet the preset score requirements.

[0028] Specifically, the image to be displayed can be an image used for publishing advertisements. It can be an advertisement material uploaded by the advertiser and supports multiple image formats. The format of the image to be displayed can include, but is not limited to, JPG, PNG, GIF, etc., and this disclosure does not specifically limit it.

[0029] The aforementioned target area can be the region on the map selected by the advertiser according to their actual needs, as per this embodiment of the disclosure. The number of target areas can be one or more, depending on actual needs, and this embodiment of the disclosure does not impose specific limitations on this. The target area can include, but is not limited to, areas such as shopping malls, residential communities, and companies. The selection method for the target area can include, but is not limited to, circular, rectangular, and polygonal selection methods.

[0030] After the target area is selected, this embodiment of the disclosure can obtain multiple users located in that area. These users are considered to be a group of users more likely to accept the advertisement for the image to be displayed. Since users' interests can be detected from the content they share on social networking platforms such as Weibo and Douyin, the historical posts of these users within a preset historical time period can be obtained from the social networking platforms. Subsequently, based on these historical posts, it can be determined whether to display advertisements to multiple users in the target area.

[0031] Here, the aforementioned preset historical time period can be the most recent year, six months, a quarter, a month, a week, or a day, but is not limited to these. It is determined specifically according to actual needs, and this embodiment of the disclosure does not impose any specific limitations on it. The aforementioned historical published content can be obtained through platform API interfaces or data cooperation methods, and undergoes anonymization and structuring processing. It can include, but is not limited to, user-posted text, images, videos, comments, etc., and this embodiment of the disclosure does not impose any specific limitations on it.

[0032] The set of images to be deployed is represented as I = { i 1, i 2, ..., i a , ..., i x},in, I This represents the set of images to be displayed. i a Indicates the first a One image to be displayed. a greater than or equal to 1 and less than or equal to 1 x Integers; x This indicates the total number of images to be displayed.

[0033] The set of target keywords determined by the set of images to be delivered is represented as follows: K = { k 1, k 2, k 3, ..., k b , ..., k m},in, K Represents the target keyword set; k b Indicates the first b One target keyword, b greater than or equal to 1 and less than or equal to 1 m Integers; m This indicates the total number of target keywords.

[0034] The user set corresponding to multiple users in the region to be targeted is represented as follows: U = { u 1, u 2, ..., u c , ..., u z},in, U Represents a set of users; k b Indicates the first c One user, c greater than or equal to 1 and less than or equal to 1 z Integers; z This indicates the total number of users in the region to be targeted.

[0035] Figure 2 This illustration shows a structural diagram of the image and keyword graph and the user and keyword graph provided in an exemplary embodiment of this disclosure, such as... Figure 2 As shown, the set of images to be deployed I There are 4 images to be displayed, namely i 1. i 2. i 3 and i 4; The set of target keywords associated with these 4 images to be displayed K There are 7 target keywords, namely k 1. k 2. k 3. k 4. k 5. k 6 and k 7; User set U There are 4 users, namely u 1. u 2. u 3 and u 4.

[0036] like Figure 2 As shown, in the construction of the image-keyword graph 201, multiple images to be deployed are used as image nodes, and multiple target keywords determined by the multiple images to be deployed are used as target keyword nodes. Edges are formed by connecting each image node to be deployed and each target keyword node. The edges in the image-keyword graph have image association weights between each image to be deployed and each target keyword. These image association weights characterize the degree of association between each image to be deployed and each target keyword. The larger the image association weight, the higher the degree of association between each image to be deployed and each target keyword; the smaller the image association weight, the lower the degree of association between each image to be deployed and each target keyword.

[0037] Image and keyword graph 201 can be a weighted bipartite graph. G (( I , K ), E , w ( i a , k b )),in, I = { i 1, i 2, ..., i a , ..., i x The multiple target keywords determined from multiple images to be delivered can be represented as follows: K = U x a=1 K ( i a Here, U represents the meaning of a set. K ( i a () indicates the image to be projected. i a Target keywords E This represents a set of edges in the image and keyword graph 201, where each edge connects to a... K The target keywords and a I Connect the images to be displayed in the image. w ( i a , k b The image association weights of the edges determine the quality and strength of the connection between each pair.

[0038] like Figure 2 As shown, in the construction process of the user-keyword graph 202, multiple target keywords in the image-keyword graph 201 are used as target keyword nodes, and multiple users are used as user nodes. Edges are obtained by connecting each target keyword node and each user node. The edges of the user-keyword graph have user association weights between each user and each target keyword. These user association weights represent the degree of association between each user's historical published content and each target keyword. The larger the user association weight, the higher the degree of association between each user's historical published content and each target keyword; the smaller the user association weight, the lower the degree of association between each user's historical published content and each target keyword.

[0039] The user and keyword graph 202 can be represented by a weighted bipartite graph. G ′(( U , K ),E ′, w ′( u c , k b )),in, E ' represents a set of edges in graph 202 connecting users and keywords, where each edge connects a user and a keyword. K The target keywords and a U Connecting users in the middle, w ′( u c , k b The edge weight represents the user association weight, indicating the connection level between the user and the target keyword. It can be calculated based on the frequency with which the user used the target keyword in their historical content within a preset historical time period. For example, ... Figure 2 As shown, the user u 1. I previously used 3 target keywords ( k 1. k 3 and k 6) User u 1. Each of these three target keywords has a user association weight, representing the user's... u 1. The degree of use and relevance of these 3 target keywords.

[0040] Based on this, this embodiment of the disclosure constructs a two-layer graph by connecting the image and keyword graph 201 and the user and keyword graph 202. This two-layer graph establishes a connection between the image to be displayed and the users, thereby identifying potential customer groups for the image to be displayed. This connection is achieved by connecting two defined graphs and constructing a two-layer graph; the common element connecting these two layers is the target keyword. Therefore, the interconnection between the two graph networks is based on the shared target keyword between these two graph networks. Subsequently, potential customers for each image to be displayed are identified from the user set.

[0041] For example, in this embodiment of the disclosure, after determining the image association weight through the image and keyword diagram 201 and the user association weight through the user and keyword diagram 202, a pre-constructed suitability scoring model can be obtained. Based on the image association weight and the user association weight, the suitability scoring model is used to calculate the suitability score of each user for each image to be delivered. The images to be delivered are then delivered to users whose suitability scores meet the preset scoring requirements.

[0042] In the method of this embodiment, the preset scoring requirement can be that the fitness score is greater than a preset fitness score. The preset fitness score can be determined according to actual needs, and this embodiment does not specifically limit it. The preset scoring requirement can also be that the fitness scores are sorted in descending order, and the top few scores are selected. Of course, the above preset scoring requirement can also be determined according to actual needs, and this embodiment does not specifically limit it.

[0043] For example, such as Figure 2 As shown, in the user u 2 and users u 4. For each image to be displayed i When the suitability scores of user 3 are all higher than the preset suitability scores, it indicates that the user... u 2 and users u 4 is the image to be displayed. i 3. Potential customers, then to users u 2 and users u 4. Project the image to be projected i 3.

[0044] Here, the suitability score can be used to measure the degree of dependence between each user and each image to be displayed. Displaying images to users whose suitability scores meet preset requirements allows for precise matching of images to potential audiences, improving user satisfaction and campaign performance. This abandons the broad-based approach and achieves accurate recommendations. Simultaneously, targeting the potential audience within the designated area avoids wasting advertising costs due to large or abnormal user data volumes, thus reducing advertising costs for advertisers.

[0045] According to the technical solution of the example embodiment of this disclosure, multiple images to be delivered and historical content published by multiple users in the delivery area within a preset historical time period are obtained; an image-keyword graph is constructed based on multiple target keywords determined by the multiple images to be delivered; wherein the edges of the image-keyword graph have image association weights between each image to be delivered and each target keyword; a user-keyword graph is constructed based on the historical content and multiple target keywords; wherein the edges of the user-keyword graph have user association weights between each user and each target keyword; a pre-built suitability scoring model is used to calculate the suitability score of each user for each image to be delivered based on the image association weights and user association weights; and each image to be delivered is delivered to users whose suitability scores meet the preset scoring requirements. This can achieve accurate recommendations to multiple users in the delivery area, avoid users receiving irrelevant information, and reduce delivery costs.

[0046] In some embodiments, constructing an image-keyword graph based on multiple target keywords determined from multiple images to be deployed may include: Multiple image features are extracted from multiple images to be deployed, and multiple candidate keywords associated with the multiple image features are determined; wherein each image to be deployed has at least one image feature, and each image feature is associated with at least one candidate keyword; Obtain a pre-constructed image association weight calculation model, and use the image association weight calculation model to calculate the image association weight between each image to be deployed and each candidate keyword based on at least one candidate keyword associated with each image feature in each image to be deployed; Multiple candidate keywords whose image association weight is greater than the dynamic weight threshold of each image to be delivered are identified as multiple target keywords for each image to be delivered. Based on multiple target keywords for multiple images to be deployed, construct an image-keyword graph.

[0047] Specifically, for each image to be deployed, it can be decomposed into at least one image feature (feature extraction process), and each image feature is associated with at least one candidate keyword. Based on this, multiple image features corresponding to multiple images to be deployed, as well as multiple candidate keywords associated with multiple image features, can be obtained. Multiple image features and multiple candidate keywords facilitate understanding different aspects of information about the given multiple images to be deployed.

[0048] Here, in the feature extraction process, techniques such as image annotation, object recognition, brand recognition, landmark detection, optical character recognition, sentiment detection, and content detection can be used for feature detection. Multiple image features can cover a large number of entities, including information such as objects, content, text, activities, sentiment, landmarks, and brands in the image to be displayed, but are not limited to these.

[0049] Assuming the image to be displayed i a Images to be projected i a The corresponding image feature set is represented as F ( i a ) = {( f 1, w ( i a , f 1), ( f 2, w ( i a , f 2), ..., ( f j , w ( i a , fj )), ..., ( f n , w ( i a , f n ))},in, F ( i a () indicates the image to be projected. i a The corresponding set of image features; f j Indicates the first j Image features, j greater than or equal to 1 and less than or equal to 1 n Integers; n Indicates the image to be displayed i a The total number of corresponding image features; w ( i a , f j () indicates the image to be projected. i a Image features f j Feature weights.

[0050] like Figure 2 As shown, in the method of this embodiment, the four images to be projected have six image features, namely... f 1. f 2. f 3. f 4. f 5 and f 6. Using these 6 image features, connect the 4 images to be delivered and multiple candidate keywords, and filter the multiple candidate keywords to obtain 7 target keywords. Then, using the 4 images to be delivered as image nodes and the 7 target keywords as target keyword nodes, connect each image node to be delivered and each target keyword node to obtain edges, thereby constructing the image and keyword graph 201.

[0051] In other words, in the image and keyword graph 201, firstly, by identifying the image features extracted from the images to be delivered, multiple target keywords are created and the association between multiple images to be delivered and multiple target keywords is established; subsequently, these keywords are connected to the corresponding images to be delivered through edges with image association weights.

[0052] When filtering multiple candidate keywords, this embodiment of the present disclosure can pre-construct an image association weight calculation model, and use the image association weight calculation model to calculate the image association weight between each image to be delivered and each candidate keyword based on at least one candidate keyword associated with each image feature in each image to be delivered; and determine multiple candidate keywords whose image association weight is greater than the dynamic weight threshold of each image to be delivered as multiple target keywords of each image to be delivered.

[0053] For example, if the image features associated with the candidate keywords are measurable features (measurable features can be image features that occupy a specific and computable region in the image to be displayed and have measurable physical properties), then the image association weight calculation model is as follows:

[0054] in, Indicates the image to be displayed With keywords Image association weights; Representation and keywords Associated image features; Representation and keywords The total number of associated image features; Representation and keywords Associated image features Image to be displayed The proportion of the area occupied by the symbol ranges from 0 to 1. Indicates the image to be displayed Chinese identification and keywords Associated image features confidence level The value of is greater than or equal to 0 and less than or equal to 1; r Indicates the image to be displayed Chinese and keywords Associated image features The overall coverage and importance.

[0055] Here, confidence scores can be extracted along with image features during the feature extraction stage. Generally, an image feature may appear in multiple locations within the image to be projected. In this case, r Used to calculate the total coverage and importance of the image features.

[0056] If the image features associated with the candidate keywords are unmeasurable features (unmeasurable features can be conceptual image features that do not exist in a specific region of the image to be delivered and are extracted from the entire image to be delivered, such as content features), then the image association weight calculation model is as follows:

[0057] in, Representation and keywords Associated image features and images to be displayed The correlation value ranges from 0 to 1. In other words, Representation and keywords Associated image features Describe the image to be displayed The degree of relevance.

[0058] In some embodiments, the method may further include: Based on the image association weights between each image to be delivered and each candidate keyword, the mean and standard deviation of the image association weights of multiple candidate keywords in each image to be delivered are calculated. Obtain a pre-built dynamic weight threshold calculation model, and use the dynamic weight threshold calculation model to calculate the dynamic weight threshold of each image to be deployed based on the mean of image association weights and the standard deviation of image association weights.

[0059] Specifically, in this embodiment, the average and standard deviation of the image association weights of multiple candidate keywords in each image to be deployed can be calculated based on the image association weights between each image to be deployed and each candidate keyword. The average and standard deviation of the image association weights are then substituted into a pre-constructed dynamic weight threshold calculation model to calculate the dynamic weight threshold for each image to be deployed. Here, the dynamic weight threshold for each image to be deployed can be adjusted in real time. Therefore, multiple target keywords can be dynamically determined from multiple candidate keywords to update the image and keyword graph 201 and the user and keyword graph 202, thereby adjusting the potential customers for the images to be deployed.

[0060] For example, the dynamic weight threshold calculation model can be:

[0061]

[0062] in, Indicates the image to be displayed Dynamic weight threshold; Indicates the image to be displayed With keywords Image association weights; Indicates the image to be displayed Related keywords The total quantity; Indicates the image to be displayed With multiple keywords The mean of image association weights; Indicates the image to be displayed With multiple keywords The standard deviation of the image association weights.

[0063] In some embodiments, the formula for calculating the user association weight between each user and each target keyword can be:

[0064] in, Indicates user With keywords User association weight; Indicates user The total number of times multiple target keywords are used in historical content published within a preset historical time period; Indicates user Use keywords in historical posts within a preset historical time period The number of times.

[0065] In some embodiments, the fitness scoring model is:

[0066] in, Indicates user For the image to be displayed Suitability rating; Represents a collection of multiple users; This represents a set of multiple images to be displayed. Indicates the image to be displayed With keywords Image association weights; Indicates user With keywords User association weight; Indicates the image to be displayed Related keywords The total number.

[0067] Based on the above image projection method, this disclosure also provides an image projection system, including a multi-area map selection module, a projection strategy formulation module, an advertising projection module, a user feedback module, and a data analysis module.

[0068] In the map selection module, advertisers can select the areas where they want to display the images. It supports multiple areas, such as shopping malls, residential areas, and companies. It also supports circular, rectangular, and polygonal selection methods. After the selection is completed, the system backend can calculate and display the estimated number of people in the selected area, which helps advertisers decide whether to place ads on the people in that area.

[0069] In the campaign strategy formulation module, advertisers upload images to be advertised, including the ad creatives to be displayed. Multiple image formats are supported. Successful uploads are then used for training within the campaign strategy. Advertisers select social network platforms where users frequently post content, such as Weibo and Douyin. The backend retrieves the users and uploaded files within the target region and runs a recommendation algorithm program based on a two-layer graph structure (image and keyword graph and user and keyword graph) to achieve the following steps: 1. Construct the first-layer graph network, namely the image and keyword graph. Create candidate keywords and establish associations by identifying image features extracted from the image to be delivered. Connect these candidate keywords to the corresponding image to be delivered through edges with image association weights. Calculate the image association weights of all candidate keywords related to the image to be delivered. Select those candidate keywords whose image association weights are higher than the dynamic weight threshold of the image to be delivered as target keywords in order to construct the image and keyword graph.

[0070] 2. Construct the second-layer graph network, namely the user and keyword graph. By connecting users with target keywords to form a network graph, this network is modeled as a weighted bipartite graph. Based on the frequency of keyword usage in the historical content published by such users within a preset historical time period, the user association weight of the edge between the user and the target keyword is calculated.

[0071] 3. Establish the connection between the image to be delivered and the user by connecting two defined graphs and constructing a two-layer graph. Calculate the suitability score for the image to be delivered to all users. Then, sort and classify multiple users according to the suitability score. Finally, identify the top-ranked users as potential customers for the image to be delivered and determine the delivery strategy accordingly.

[0072] In the ad delivery module, the generated delivery strategy allows for precise delivery of the advertisers' uploaded ad creatives to the selected delivery area. Different types of ads will be delivered to the target audience. Here, you can select the corresponding social network delivery channels chosen in the previous module. During the ad delivery process, the ad performance can be monitored in real time on the visual interface, and the delivery strategy can be automatically adjusted based on the performance feedback.

[0073] The user feedback module collects user feedback data on advertisements, including metrics such as click-through rate, conversion rate, and user satisfaction, and displays it in real time on the visualization interface.

[0074] The data analysis module provides in-depth analysis of ad performance, offering visualized data reports to help users understand ad performance and optimize ad delivery strategies.

[0075] The foregoing mainly describes the solutions provided by the embodiments of this disclosure. It is understood that, in order to achieve the above functions, the electronic device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this disclosure can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0076] This disclosure embodiment can divide the electronic device into functional units according to the above method example. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this disclosure embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0077] In the case of dividing each functional module according to its corresponding function, an exemplary embodiment of this disclosure provides an image projection device, which can be a terminal or a chip applied to a terminal. Figure 3 A schematic diagram of the structure of an image projection apparatus provided in an exemplary embodiment of this disclosure is shown. Figure 3 As shown, the device 300 includes: The acquisition module 301 is used to acquire multiple images to be delivered, as well as the historical content published by multiple users in the delivery area within a preset historical time period. Construction module 302 is used to construct an image-keyword graph based on multiple target keywords determined from the multiple images to be deployed; wherein the edges of the image-keyword graph have image association weights between each image to be deployed and each target keyword; The construction module 302 is further configured to construct a user-keyword graph based on the historical published content and the multiple target keywords; wherein, the edges of the user-keyword graph have user association weights between each user and each target keyword; The calculation module 303 is used to calculate the suitability score of each user for each image to be delivered based on the image association weight and the user association weight using a pre-built suitability scoring model. The delivery module 304 is used to deliver each of the images to be delivered to the users whose suitability scores meet the preset score requirements.

[0078] In some embodiments, the construction module 302 is further configured to extract multiple image features from the plurality of images to be deployed, and determine multiple candidate keywords associated with the plurality of image features; wherein each of the images to be deployed has at least one of the image features, and each of the image features is associated with at least one of the candidate keywords; obtain a pre-constructed image association weight calculation model, and use the image association weight calculation model to calculate the image association weight between each of the images to be deployed and each of the candidate keywords based on at least one of the candidate keywords associated with each of the image features in each of the images to be deployed; determine multiple candidate keywords whose image association weight is greater than the dynamic weight threshold of each of the images to be deployed as multiple target keywords of each of the images to be deployed; and construct an image and keyword graph based on the multiple target keywords of the plurality of images to be deployed.

[0079] In some embodiments, if the image feature associated with the candidate keyword is a measurable feature, then the image association weight calculation model is as follows:

[0080] in, Indicates the image to be displayed With keywords Image association weights; Representation and keywords Associated image features; Representation and keywords The total number of associated image features; Representation and keywords Associated image features Image to be displayed The proportion of the area occupied by the symbol ranges from 0 to 1. Indicates the image to be displayed Chinese identification and keywords Associated image features confidence level The value of is greater than or equal to 0 and less than or equal to 1; r Indicates the image to be displayed Chinese and keywords Associated image features The overall coverage and importance; If the image feature associated with the candidate keyword is an unmeasurable feature, then the image association weight calculation model is as follows:

[0081] in, Representation and keywords Associated image features and images to be displayed The correlation value ranges from 0 to 1.

[0082] In some embodiments, the calculation module 303 is further configured to calculate the mean and standard deviation of the image association weights of multiple candidate keywords in each image to be delivered based on the image association weights between each image to be delivered and each candidate keyword; obtain a pre-built dynamic weight threshold calculation model; and use the dynamic weight threshold calculation model to calculate the dynamic weight threshold of each image to be delivered based on the mean and standard deviation of the image association weights. The dynamic weight threshold calculation model is as follows:

[0083]

[0084] in, Indicates the image to be displayed Dynamic weight threshold; Indicates the image to be displayed With keywords Image association weights; Indicates the image to be displayed Related keywords The total quantity; Indicates the image to be displayed With multiple keywords The mean of image association weights; Indicates the image to be displayed With multiple keywords The standard deviation of the image association weights.

[0085] In some embodiments, the formula for calculating the user association weight between each user and each target keyword is as follows:

[0086] in, Indicates user With keywords User association weight; Indicates user The total number of times the multiple target keywords are used in the historical published content within the preset historical time period; Indicates user Keywords used in the historical published content within the preset historical time period The number of times.

[0087] In some embodiments, the fitness scoring model is:

[0088] in, Indicates user For the image to be displayed Suitability rating; Represents a collection of multiple users; This represents a set of multiple images to be displayed. Indicates the image to be displayed With keywords Image association weights; Indicates user With keywords User association weight; Indicates the image to be displayed Related keywords The total number.

[0089] This disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the methods disclosed in this disclosure.

[0090] Figure 4 A schematic diagram of the structure of an electronic device provided in an exemplary embodiment of this disclosure is shown. For example... Figure 4 As shown, the electronic device 400 includes at least one processor 401 and a memory 402 coupled to the processor 401, which can perform the corresponding steps in the methods disclosed in the embodiments of this disclosure.

[0091] The processor 401 described above can also be called a Central Processing Unit (CPU), which can be an integrated circuit chip with signal processing capabilities. Each step in the method disclosed in this embodiment can be implemented by the integrated logic circuitry in the hardware of the processor 401 or by instructions in software form. The processor 401 can be a general-purpose processor, a digital signal processor (DSP), an ASIC, a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this embodiment can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules can be located in the memory 402, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The processor 401 reads information from the memory 402 and, in conjunction with its hardware, completes the steps of the method described above.

[0092] Furthermore, various operations / processes according to this disclosure, implemented via software and / or firmware, can be transmitted from a storage medium or network to a computer system with a dedicated hardware architecture, for example, Figure 5 The computer system 500 shown is equipped with the programs that constitute the software. When various programs are installed, the computer system is able to perform various functions, including functions such as those described above. Figure 5 A schematic diagram of the structure of a computer system provided in an exemplary embodiment of this disclosure is shown.

[0093] Computer system 500 is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of this disclosure described and / or claimed herein.

[0094] like Figure 5As shown, the computer system 500 includes a computing unit 501, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. The RAM 503 may also store various programs and data required for the operation of the computer system 500. The computing unit 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0095] Multiple components in the computer system 500 are connected to the I / O interface 505, including: an input unit 506, an output unit 507, a storage unit 508, and a communication unit 509. The input unit 506 can be any type of device capable of inputting information into the computer system 500. The input unit 506 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of the electronic device. The output unit 507 can be any type of device capable of presenting information and may include, but is not limited to, a monitor, speaker, video / audio output terminal, vibrator, and / or printer. The storage unit 508 may include, but is not limited to, a hard disk and an optical disk. The communication unit 509 allows the computer system 500 to exchange information / data with other devices via a network such as the Internet, and may include, but is not limited to, a modem, network card, infrared communication device, wireless communication transceiver, and / or chipset, such as Bluetooth™ devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0096] The computing unit 501 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above. For example, in some embodiments, the methods disclosed in this disclosure can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed on an electronic device via ROM 502 and / or communication unit 509. In some embodiments, the computing unit 501 can be configured to perform the methods disclosed in this disclosure by any other suitable means (e.g., by means of firmware).

[0097] This disclosure also provides a computer-readable storage medium, wherein when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is able to perform the methods disclosed in this disclosure.

[0098] The computer-readable storage medium in this disclosure can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. The aforementioned computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specifically, the aforementioned computer-readable storage medium may include electrical connections based on one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0099] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0100] This disclosure also provides a computer program product, including a computer program, wherein when the computer program is executed by a processor, it implements the methods disclosed in the embodiments of this disclosure.

[0101] In embodiments of this disclosure, computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include, but are not limited to, object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network (including a local area network (LAN) or a wide area network (WAN)), or it can be connected to an external computer.

[0102] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0103] The modules, components, or units described in the embodiments of this disclosure can be implemented in software or hardware. The names of the modules, components, or units do not necessarily constitute a limitation on the module, component, or unit itself.

[0104] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary hardware logic components that can be used include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0105] The above description is merely an illustration of some embodiments of this disclosure and the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0106] While specific embodiments of this disclosure have been described in detail by way of example, those skilled in the art should understand that the examples are for illustrative purposes only and not intended to limit the scope of this disclosure. Those skilled in the art should understand that modifications can be made to the above embodiments without departing from the scope and spirit of this disclosure. The scope of this disclosure is defined by the appended claims.

Claims

1. An image projection method, characterized in that, include: Acquire multiple images to be displayed, as well as the historical content posted by multiple users in the display area within a preset historical time period; Based on multiple target keywords determined from the multiple images to be deployed, an image-keyword graph is constructed; wherein, the edges of the image-keyword graph have image association weights between each of the images to be deployed and each of the target keywords; Based on the historical published content and the multiple target keywords, a user-keyword graph is constructed; wherein, the edges of the user-keyword graph have user association weights between each user and each target keyword; Using a pre-built suitability scoring model based on the image association weights and the user association weights, calculate the suitability score for each user for each image to be displayed; Each of the images to be displayed is delivered to the users whose suitability scores meet the preset score requirements.

2. The method as described in claim 1, characterized in that, The step of constructing an image-keyword graph based on multiple target keywords determined from the multiple images to be deployed includes: Multiple image features are extracted from the multiple images to be delivered, and multiple candidate keywords associated with the multiple image features are determined; wherein each of the multiple images to be delivered has at least one of the image features, and each of the image features is associated with at least one of the candidate keywords; A pre-constructed image association weight calculation model is obtained, and the image association weight calculation model is used to calculate the image association weight between each image to be deployed and each candidate keyword based on at least one candidate keyword associated with each image feature in each image to be deployed; The candidate keywords whose image association weight is greater than the dynamic weight threshold of each image to be delivered are determined as the target keywords of each image to be delivered. Based on the multiple target keywords of the multiple images to be deployed, an image-keyword graph is constructed.

3. The method as described in claim 2, characterized in that, If the image feature associated with the candidate keyword is a measurable feature, then the image association weight calculation model is as follows: in, Indicates the image to be displayed With keywords Image association weights; Representation and keywords Associated image features; Representation and keywords The total number of associated image features; Representation and keywords Associated image features Image to be displayed The proportion of the area occupied by the symbol ranges from 0 to 1. Indicates the image to be displayed Chinese identification and keywords Associated image features confidence level The value of is greater than or equal to 0 and less than or equal to 1; r Indicates the image to be displayed Chinese and keywords Associated image features The overall coverage and importance; If the image feature associated with the candidate keyword is an unmeasurable feature, then the image association weight calculation model is as follows: in, Representation and keywords Associated image features and images to be displayed The correlation value ranges from 0 to 1.

4. The method as described in claim 2, characterized in that, The method further includes: Based on the image association weights between each of the images to be deployed and each of the candidate keywords, the mean and standard deviation of the image association weights of multiple candidate keywords in each of the images to be deployed are calculated. Obtain a pre-constructed dynamic weight threshold calculation model, and use the dynamic weight threshold calculation model to calculate the dynamic weight threshold of each image to be deployed based on the mean of the image association weights and the standard deviation of the image association weights; The dynamic weight threshold calculation model is as follows: in, Indicates the image to be displayed Dynamic weight threshold; Indicates the image to be displayed With keywords Image association weights; Indicates the image to be displayed Related keywords The total quantity; Indicates the image to be displayed With multiple keywords The mean of image association weights; Indicates the image to be displayed With multiple keywords The standard deviation of the image association weights.

5. The method as described in claim 1, characterized in that, The formula for calculating the user association weight between each user and each target keyword is as follows: in, Indicates user With keywords User association weight; Indicates user The total number of times the multiple target keywords are used in the historical published content within the preset historical time period; Indicates user Keywords used in the historical published content within the preset historical time period The number of times.

6. The method according to any one of claims 1 to 5, characterized in that, The fitness scoring model is as follows: in, Indicates user For the image to be displayed Suitability rating; Represents a collection of multiple users; This represents a set of multiple images to be displayed. Indicates the image to be displayed With keywords Image association weights; Indicates user With keywords User association weight; Indicates the image to be displayed Related keywords The total number.

7. An image projection device, characterized in that, include: The acquisition module is used to acquire multiple images to be delivered, as well as the historical content published by multiple users in the delivery area within a preset historical time period. A construction module is used to construct an image-keyword graph based on multiple target keywords determined from the multiple images to be deployed; wherein the edges of the image-keyword graph have image association weights between each image to be deployed and each target keyword; The construction module is further configured to construct a user-keyword graph based on the historical published content and the multiple target keywords; wherein, the edges of the user-keyword graph have user association weights between each user and each target keyword; The calculation module is used to calculate the suitability score of each user for each image to be delivered based on the image association weight and the user association weight using a pre-built suitability scoring model. The delivery module is used to deliver each of the images to be delivered to the users whose suitability scores meet the preset score requirements.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method described in any one of claims 1 to 6.

Images