A method, system and storage medium for intelligent advertising delivery

Abstract

The present application discloses an intelligent advertising delivery method, system and storage medium, and belongs to the field of computer technology. Through the technical solution provided by the embodiment of the present application, in response to the target video being selected by the target object, the video information of the target video and the object information of the target object are obtained. The target video, video information and object information are used to determine the target advertisement and the target display time from multiple candidate advertisements, thereby realizing the screening of advertisements and the intelligent determination of the advertising delivery time. When the video is played to the reference time before the target display time, the object image of the target object and the playback mode of the target video are used to determine whether to deliver the target advertisement at the target display time, thereby realizing the intelligent control of advertising delivery and improving the efficiency of advertising delivery.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an intelligent advertising delivery method, system and storage medium. Background Art

[0002] With the rapid development of Internet technology, advertising business has also been widely spread. Advertising has gradually shifted from traditional mass communication based on radio and television to interpersonal communication based on social media and entertainment platforms.

[0003] In the related art, the delivery of advertisements relies on manual configuration, and advertisers are required to configure the scope and content of advertisement delivery by themselves, resulting in low efficiency of advertisement delivery.

[0004] Therefore, how to improve the efficiency of advertising delivery is a hot topic of research. Summary of the invention

[0005] The embodiments of the present application provide an intelligent advertising delivery method, system and storage medium, which can improve the efficiency of advertising delivery. The technical solution is as follows:

[0006] In one aspect, a method for intelligent advertisement delivery is provided, the method comprising:

[0007] In response to a target video being selected by a target object, acquiring video information of the target video and object information of the target object;

[0008] Determine a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, object information of the target object, and video information of the target video, wherein the plurality of candidate advertisements are advertisements that match the target object and the target video;

[0009] In response to the target video being played to a reference time before the target presentation time, acquiring an object image of the target object and a play mode of the target video, the play mode including accelerated play, normal play, and slow play;

[0010] Based on the object image and the play mode, it is determined whether to deliver the target advertisement at the target display time.

[0011] In a possible implementation, determining a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, the object information of the target object, and the video information of the target video includes:

[0012] Extracting features from the target object information to obtain the target object features;

[0013] Extracting features from the video information of the target video to obtain video information features of the target video;

[0014] Using the object features of the target object and the video information features of the target video to perform advertisement recall, to obtain the plurality of candidate advertisements;

[0015] Based on the target video and the multiple candidate advertisements, the target advertisement and the target display time of the target advertisement in the target video are determined from the multiple candidate advertisements.

[0016] In a possible implementation, the determining the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements based on the target video and the multiple candidate advertisements includes:

[0017] Determining a plurality of candidate commodities corresponding to the plurality of candidate advertisements respectively;

[0018] Based on the multiple candidate commodities, the target video is subjected to commodity identification and commodity screening to obtain multiple reference commodities, wherein commodity types of the reference commodities belong to a commodity type set, and the commodity type set is a set of commodity types of the multiple candidate commodities;

[0019] Based on the target video and the multiple reference products, determine a target product from the multiple reference products and determine a candidate advertisement corresponding to the target product as the target advertisement;

[0020] Based on the target video and the target product, a target display time of the target advertisement in the target video is determined.

[0021] In a possible implementation, the determining the target product from the multiple reference products based on the target video and the multiple reference products includes:

[0022] Determine a set of video frames corresponding to each of the reference products in the target video, wherein the set of video frames includes a plurality of video frames in which the corresponding reference products appear;

[0023] Determine, based on the video frame set, the importance of each of the reference products in the target video and the interference degree of advertisement placement corresponding to each of the reference products, wherein the interference degree of advertisement placement is used to indicate the interference degree of placing candidate advertisements corresponding to the reference products on viewing the target video;

[0024] A target product is determined from the multiple reference products based on the importance of each of the reference products in the target video and the interference degree of advertisement placement corresponding to each of the reference products.

[0025] In a possible implementation, determining a target display time of the target advertisement in the target video based on the target video and the target product includes:

[0026] Determining a plurality of candidate display times based on a target video frame set corresponding to the target advertisement in the target video;

[0027] Using the multiple candidate presentation times to segment the target video frame set to obtain multiple sub-video frame sets;

[0028] Based on the multiple sub-video frame sets, the target presentation time is determined from the multiple candidate presentation times.

[0029] In a possible implementation, the determining the target presentation time from the multiple candidate presentation times based on the multiple sub-video frame sets includes:

[0030] Extracting features from each of the sub-video frame sets to obtain sub-video frame set features of each of the sub-video frame sets;

[0031] Determining an advertisement display index of each of the sub-video frame sets based on the sub-video frame set features of each of the sub-video frame sets, wherein the advertisement display index is used to indicate the time suitability of the candidate display time corresponding to the sub-video frame set;

[0032] The candidate display time having the highest corresponding advertisement display index among the multiple candidate display times is determined as the target display time.

[0033] In a possible implementation manner, determining whether to place the target advertisement at the target display time based on the object image and the playback mode includes:

[0034] Based on the object image, determining the expression and body movement of the target object;

[0035] Based on the playback mode, the expression and body movements of the target object, it is determined whether to place the target advertisement at the target display time.

[0036] In a possible implementation, determining whether to place the target advertisement at the target display time based on the playback mode and the expression and body movement of the target object includes:

[0037] Based on the playback mode, determining a first delivery judgment parameter, wherein the first delivery judgment parameter is used to indicate a delivery suitability degree associated with the playback mode;

[0038] Based on the expression of the target object, determining a second delivery judgment parameter, wherein the second delivery judgment parameter is used to indicate a delivery suitability degree associated with the expression;

[0039] Based on the body movement of the target object, determining a third delivery judgment parameter, wherein the third delivery judgment parameter is used to indicate a delivery suitability degree associated with the body movement;

[0040] The first delivery judgment parameter, the second delivery judgment parameter and the third delivery judgment parameter are integrated to obtain an integrated judgment parameter;

[0041] When the fusion judgment parameter is greater than or equal to the parameter threshold, determining to place the target advertisement at the target display time;

[0042] When the fusion judgment parameter is less than the parameter threshold, it is determined not to deliver the target advertisement at the target display time.

[0043] In one aspect, a system for intelligent advertising delivery is provided, the system comprising:

[0044] A first acquisition module, configured to acquire video information of the target video and object information of the target object in response to the target video being selected by the target object;

[0045] A determination module, configured to determine a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, object information of the target object, and video information of the target video;

[0046] A second acquisition module, in response to a reference time before the target video is played to the target display time, acquires the object image of the target object and the play mode of the target video, wherein the play mode includes accelerated play, normal play and slow play;

[0047] A judgment module is used to determine whether to deliver the target advertisement at the target display time based on the object image and the playback mode.

[0048] In a possible implementation, the determination module is used to perform feature extraction on the object information of the target object to obtain the object features of the target object; perform feature extraction on the video information of the target video to obtain the video information features of the target video; use the object features of the target object and the video information features of the target video to perform advertisement recall to obtain the multiple candidate advertisements; and based on the target video and the multiple candidate advertisements, determine the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements.

[0049] In a possible implementation, the determination module is used to determine multiple candidate products corresponding to the multiple candidate advertisements respectively; based on the multiple candidate products, the target video is subjected to product identification and product screening to obtain multiple reference products, wherein the product type of the reference products belongs to a product type set, and the product type set is a set of product types of the multiple candidate products; based on the target video and the multiple reference products, a target product is determined from the multiple reference products and the candidate advertisement corresponding to the target product is determined as the target advertisement; based on the target video and the target product, a target display time of the target advertisement in the target video is determined.

[0050] In a possible implementation, the determination module is used to determine a set of video frames corresponding to each of the reference products in the target video, the set of video frames including a plurality of video frames in which the corresponding reference products appear; based on the set of video frames, determine the importance of each of the reference products in the target video and the degree of interference of advertising placement corresponding to each of the reference products, the degree of interference of advertising placement being used to indicate the degree of interference of candidate advertisements corresponding to the reference products with viewing the target video; determine the target product from the plurality of reference products based on the importance of each of the reference products in the target video and the degree of interference of advertising placement corresponding to each of the reference products.

[0051] In one possible implementation, the determination module is used to determine a plurality of candidate display times based on a target video frame set corresponding to the target advertisement in the target video; segment the target video frame set using the plurality of candidate display times to obtain a plurality of sub-video frame sets; and determine the target display time from the plurality of candidate display times based on the plurality of sub-video frame sets.

[0052] In a possible implementation, the determination module is used to extract features from each of the sub-video frame sets to obtain sub-video frame set features of each of the sub-video frame sets; determine an advertising display index for each of the sub-video frame sets based on the sub-video frame set features of each of the sub-video frame sets, wherein the advertising display index is used to indicate the time suitability of a candidate display time corresponding to the sub-video frame set; and determine the candidate display time having the highest corresponding advertising display index among multiple candidate display times as the target display time.

[0053] In a possible implementation, the judgment module is used to determine the expression and body movements of the target object based on the object image; and determine whether to place the target advertisement at the target display time based on the playback mode, the expression and body movements of the target object.

[0054] In a possible implementation, the judgment module is used to determine a first delivery judgment parameter based on the playback mode, the first delivery judgment parameter being used to indicate a delivery suitability associated with the playback mode; determine a second delivery judgment parameter based on an expression of the target object, the second delivery judgment parameter being used to indicate a delivery suitability associated with the expression; determine a third delivery judgment parameter based on a body movement of the target object, the third delivery judgment parameter being used to indicate a delivery suitability associated with the body movement; fuse the first delivery judgment parameter, the second delivery judgment parameter and the third delivery judgment parameter to obtain a fused judgment parameter; determine to deliver the target advertisement at the target display time when the fused judgment parameter is greater than or equal to a parameter threshold; and determine not to deliver the target advertisement at the target display time when the fused judgment parameter is less than the parameter threshold.

[0055] On the one hand, a computer device is provided, which includes one or more processors and one or more memories, wherein at least one computer program is stored in the one or more memories, and the computer program is loaded and executed by the one or more processors to implement the intelligent advertising delivery method.

[0056] On the one hand, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, and the computer program is loaded and executed by a processor to implement the intelligent advertising delivery method.

[0057] On the one hand, a computer program product or a computer program is provided, which includes a program code, and the program code is stored in a computer-readable storage medium. A processor of a computer device reads the program code from the computer-readable storage medium, and the processor executes the program code, so that the computer device executes the above-mentioned intelligent advertising delivery method.

[0058] Through the technical solution provided by the embodiment of the present application, in response to the target video being selected by the target object, the video information of the target video and the object information of the target object are obtained. The target video, video information and object information are used to determine the target advertisement and the target display time from multiple candidate advertisements, thereby realizing the screening of advertisements and the intelligent determination of the advertisement delivery time. When the video is played to the reference time before the target display time, the object image of the target object and the playback mode of the target video are used to determine whether to deliver the target advertisement at the target display time, thereby realizing the intelligent control of advertisement delivery and improving the efficiency of advertisement delivery. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0060] Figure 1 It is a schematic diagram of an implementation environment of an intelligent advertising delivery method provided in an embodiment of the present application;

[0061] Figure 2 This is a flow chart of an intelligent advertising delivery method provided by an embodiment of the present application;

[0062] Figure 3 This is a flow chart of another method for intelligent advertising delivery provided by an embodiment of the present application;

[0063] Figure 4 It is a structural diagram of an intelligent advertising delivery system provided by an embodiment of the present application;

[0064] Figure 5 It is a structural diagram of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0065] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0066] In this application, the terms "first", "second", etc. are used to distinguish identical or similar items with basically the same effects and functions. It should be understood that there is no logical or temporal dependency between "first", "second", and "nth", nor is there any limitation on quantity and execution order.

[0067] Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain better results. In other words, artificial intelligence is a comprehensive technology of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence.

[0068] Machine Learning (ML) is a multi-disciplinary subject that involves probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, and other disciplines. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge sub-models to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence.

[0069] Normalization: Mapping sequences with different value ranges to the interval (0, 1) makes data processing easier. In some cases, the normalized values ​​can be directly implemented as probabilities.

[0070] Embedded Coding: Embedded coding mathematically represents a correspondence, that is, mapping data in X space to Y space through a function F, where the function F is an injective function, and the mapping result is structural preservation. The injective function indicates that the data after mapping is uniquely corresponding to the data before mapping, and structural preservation indicates that the size relationship of the data before mapping is the same as the size relationship of the data after mapping. For example, there are data X1 and X2 before mapping, and after mapping, Y1 corresponding to X1 and Y2 corresponding to X2 are obtained. If the data X1 before mapping is greater than X2, then correspondingly, the data Y1 after mapping is greater than Y2. For words, it means mapping the words to another space to facilitate subsequent machine learning and processing.

[0071] Attention weight: It can indicate the importance of a certain data in the training or prediction process. Importance indicates the influence of input data on output data. Data with high importance has a higher corresponding attention weight value, and data with low importance has a lower corresponding attention weight value. In different scenarios, the importance of data is different. The process of training the attention weight of the model is also the process of determining the importance of data.

[0072] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions. For example, the object images and object information in the embodiments of this application are obtained only after the target object has given permission.

[0073] Figure 1 This is a schematic diagram of an implementation environment of an intelligent advertising delivery method provided in an embodiment of the present application, see Figure 1 , the implementation environment may include a terminal 110 and a server 140.

[0074] The terminal 110 is connected to the server 140 via a wireless network or a wired network. Optionally, the terminal 110 is a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto. The terminal 110 is installed and runs an application program that supports video playback and advertisement display.

[0075] The server 140 is an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, distribution networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms. The server 140 can provide video playback and advertising background services for applications running on the terminal 110.

[0076] The following is a description of the intelligent advertising delivery method provided in the embodiments of the present application. Figure 2 This is a flowchart of an intelligent advertising delivery method provided by an embodiment of the present application. Figure 2 Taking the execution subject as a server as an example, the method includes the following steps.

[0077] 201. In response to a target video being selected by a target object, the server obtains video information of the target video and object information of the target object.

[0078] The target object is the user who watches the video, that is, the target of the advertisement. The target video is the video selected by the target object, that is, the video that the target object wants to watch. The video information of the target video is used to describe the target video, and the object information of the target object is used to describe the target object.

[0079] 202. The server determines a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, object information of the target object, and video information of the target video, wherein the plurality of candidate advertisements are advertisements that match the target object and the target video.

[0080] The target advertisement is the advertisement that is finally determined to be delivered. The target display time is a video playback time, which refers to the time when the target advertisement is displayed during the playback of the target video, also known as the advertisement delivery time.

[0081] 203. In response to the target video being played to a reference time before the target display time, the server obtains the object image of the target object and the play mode of the target video, where the play mode includes accelerated play, normal play, and slow play.

[0082] Among them, the reference time is also a video playback time, and the time interval between the reference time and the target display time is a preset time interval, which is set by the technical personnel according to the actual situation, and the embodiment of the present application does not limit this. The object image of the target object refers to the image of the target object collected when the target video is played to the reference time, and the object image can reflect the state of the target object watching the target video. Accelerated playback refers to a mode of playing at a first speed, and the first speed is a speed greater than 1; accelerated playback refers to a mode of playing at a second speed, and the second speed is a speed less than 1; normal playback refers to a mode of playing at a playback speed of 1.

[0083] 204. The server determines whether to place the target advertisement at the target display time based on the object image and the playback mode.

[0084] Through the technical solution provided by the embodiment of the present application, in response to the target video being selected by the target object, the video information of the target video and the object information of the target object are obtained. The target video, video information and object information are used to determine the target advertisement and the target display time from multiple candidate advertisements, thereby realizing the screening of advertisements and the intelligent determination of the advertisement delivery time. When the video is played to the reference time before the target display time, the object image of the target object and the playback mode of the target video are used to determine whether to deliver the target advertisement at the target display time, thereby realizing the intelligent control of advertisement delivery and improving the efficiency of advertisement delivery.

[0085] The above steps 201-204 are a brief introduction to the intelligent advertising delivery method provided in the embodiment of the present application. The following will combine some examples to more clearly illustrate the intelligent advertising delivery method provided in the embodiment of the present application. Figure 3 Taking the execution subject as a server as an example, the method includes the following steps.

[0086] 301. In response to a target video being selected by a target object, the server obtains video information of the target video and object information of the target object.

[0087] The target object is the user who watches the video, that is, the target of the advertisement. The target video is the video selected by the target object, that is, the video that the target object wants to watch. The video information of the target video is used to describe the target video, and the object information of the target object is used to describe the target object. It should be noted that the acquisition of the object information of the target object is carried out only after the permission of the target object.

[0088] In a possible implementation, in response to the target object clicking the target video, the server determines that the target video is selected by the target object. In response to the target video being selected by the target object, the server obtains video information of the target video from a video database and obtains object information of the target object from an object database.

[0089] The video database stores a plurality of videos and video information of each video, and the object database stores object information of a plurality of objects.

[0090] For example, in response to the target object's click operation on the target video, the server determines that the target video is selected by the target object. In response to the target video being selected by the target object, the server uses the video identifier of the target video to query in the video database to obtain the video information of the target video. The server uses the object identifier of the target object to query in the object database to obtain the object information of the target object.

[0091] 302. The server determines a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, object information of the target object, and video information of the target video. The plurality of candidate advertisements are advertisements that match the target object and the target video.

[0092] The target advertisement is the advertisement that is finally determined to be delivered. The target display time is a video playback time, which refers to the time when the target advertisement is displayed during the playback of the target video, also known as the advertisement delivery time.

[0093] In a possible implementation, the server extracts features from the object information of the target object to obtain the object features of the target object. The server extracts features from the video information of the target video to obtain the video information features of the target video. The server uses the object features of the target object and the video information features of the target video to perform advertisement recall to obtain the multiple candidate advertisements. Based on the target video and the multiple candidate advertisements, the server determines the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements.

[0094] The object feature is an abstract expression of the object information, and the video information feature is an abstract expression of the video information. In some embodiments, the object information includes the target object's gender, age, hobbies, region, and preferred video type. The video information includes the video type and video content introduction. Ad recall refers to recalling candidate ads that match the target video and the target object from the ad database.

[0095] In order to explain the above implementation more clearly, the above implementation is explained in several parts below.

[0096] In the first part, the server extracts the feature of the object information of the target object to obtain the object feature of the target object.

[0097] In a possible implementation, the server inputs the object information of the target object into an object feature extraction model, and performs feature extraction on the object information of the target object through the object feature extraction model to obtain the object features of the target object.

[0098] Among them, the object feature extraction model is a feature extractor that can extract input object information into object features. The embodiment of the present application does not limit the structure and type of the object feature extraction model.

[0099] For example, the server inputs the object information of the target object into the object feature extraction model, and through the object feature extraction model, encodes the object information based on the attention mechanism to obtain the object features of the target object.

[0100] In the second part, the server extracts features of the video information of the target video to obtain the video information features of the target video.

[0101] In a possible implementation, the server inputs the video information of the target video into a video information feature extraction model, and performs feature extraction on the video information of the target video through the video information feature extraction model to obtain the video information features of the target video.

[0102] Among them, the video information feature extraction model is a feature extractor, which can extract the input video information into video information features. The embodiment of the present application does not limit the structure and type of the video information feature extraction model.

[0103] For example, the server inputs the video information of the target video into a video information feature extraction model, and through the video information feature extraction model, encodes the video information based on the attention mechanism to obtain the video information features of the target video.

[0104] In the third part, the server uses the object features of the target object and the video information features of the target video to recall advertisements to obtain the multiple candidate advertisements.

[0105] In a possible implementation, the server splices the object feature of the target object and the video information feature of the target video to obtain a spliced ​​recall feature. The server matches the spliced ​​recall feature in an advertisement database to obtain multiple candidate advertisements, and the feature similarity between the advertisement feature of the candidate advertisement and the spliced ​​recall feature is greater than or equal to a similarity threshold.

[0106] The advertisement database stores multiple advertisements and advertisement features of each advertisement. The similarity threshold is set by a technician according to actual conditions, and the embodiment of the present application does not limit this.

[0107] For example, the server splices the object feature of the target object and the video information feature of the target video to obtain a spliced ​​recall feature. The server determines the feature similarity between the spliced ​​recall feature and multiple advertisement features stored in the advertisement database. The server determines the advertisement corresponding to the advertisement feature whose feature similarity with the spliced ​​recall feature is greater than or equal to the similarity threshold among the multiple advertisement features as the multiple candidate advertisements.

[0108] Part 4: The server determines the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements based on the target video and the multiple candidate advertisements.

[0109] In a possible implementation, the server determines multiple candidate commodities corresponding to the multiple candidate advertisements. Based on the multiple candidate commodities, the server performs commodity identification and commodity screening on the target video to obtain multiple reference commodities, wherein the commodity type of the reference commodities belongs to a commodity type set, and the commodity type set is a set of commodity types of the multiple candidate commodities. Based on the target video and the multiple reference commodities, the server determines the target commodity from the multiple reference commodities and determines the candidate advertisement corresponding to the target commodity as the target advertisement. Based on the target video and the target commodity, the server determines the target display time of the target advertisement in the target video.

[0110] Among them, one candidate advertisement corresponds to one candidate product. Product identification is to find products that have appeared in the target video, and product screening is to find products whose product types belong to the product type set.

[0111] In order to explain the above implementation more clearly, the above implementation will be further explained in several parts below.

[0112] A. The server determines multiple candidate commodities corresponding to the multiple candidate advertisements.

[0113] In a possible implementation, the server uses the advertisement identifiers of multiple candidate advertisements to query a commodity database to obtain candidate commodities corresponding to each candidate advertisement.

[0114] There is a correlation between advertisements and commodities, so the corresponding commodities can be found by using advertisement identifiers.

[0115] B. The server performs commodity identification and commodity screening on the target video based on the multiple candidate commodities to obtain multiple reference commodities.

[0116] In a possible implementation, the server performs commodity recognition on multiple video frames of the target video to obtain multiple recognized commodities in the target video, and the server determines the multiple reference commodities based on the multiple recognized commodities and the multiple candidate commodities.

[0117] For example, the server inputs multiple video frames of the target video into a product recognition model, performs target detection on each video frame through the product recognition model, and obtains the initial recognized products in each video frame. The server aggregates and removes duplicates of the initial recognized products in the multiple video frames to obtain the multiple recognized products. The server determines the product types of the multiple recognized products and the product types of multiple candidate products. The server aggregates the product types of the multiple candidate products to obtain a product type set. The server determines the recognized products whose product types belong to the product type set among the multiple recognized products as reference products, thereby obtaining multiple reference products.

[0118] C. The server determines a target product from the multiple reference products based on the target video and the multiple reference products and determines a candidate advertisement corresponding to the target product as the target advertisement.

[0119] In a possible implementation, the server determines a set of video frames corresponding to each reference product in the target video, the set of video frames including a plurality of video frames in which the corresponding reference product appears. Based on the set of video frames, the server determines the importance of each reference product in the target video and the degree of interference of advertising placement corresponding to each reference product, the degree of interference of advertising placement being used to indicate the degree of interference of candidate advertisements corresponding to the reference product with viewing the target video. Based on the importance of each reference product in the target video and the degree of interference of advertising placement corresponding to each reference product, the server determines the target product from the plurality of reference products.

[0120] For any reference product among the multiple reference products, the set of video frames corresponding to the reference product in the target video refers to the set of video frames in which the reference product appears in the target video.

[0121] For example, the server determines a plurality of video frames corresponding to each reference product in the target video. The server aggregates the plurality of video frames corresponding to each reference product in the target video to obtain a set of video frames corresponding to each reference product. The server extracts features from the set of video frames corresponding to each reference product to obtain video frame set features of the set of video frames corresponding to each reference product. Based on the video frame set features of the set of video frames corresponding to each reference product, the server determines the importance of the set of video frames corresponding to each reference product in the target video and the coherence of the set of video frames corresponding to each reference product. The server determines the importance of each reference product in the target video based on the importance of the set of video frames corresponding to each reference product in the target video. The server determines the degree of interference of advertising placement corresponding to each reference product based on the coherence of the set of video frames corresponding to each reference product. The server determines the reference product with the lowest degree of interference of advertising placement among the multiple reference products as the target product.

[0122] For example, the server determines the multiple video frames corresponding to each reference product in the target video. The server aggregates the multiple video frames corresponding to each reference product in the target video to obtain the video frame set corresponding to each reference product. The server encodes the video frame set corresponding to each reference product based on the attention mechanism to obtain the video frame set features of the video frame set corresponding to each reference product. The server uses the first classification head to fully connect and normalize the video frame set features of the video frame set corresponding to each reference product to obtain the importance of the video frame set corresponding to each reference product in the target video. The server uses the second classification head to fully connect and normalize the video frame set features of the video frame set corresponding to each reference product to obtain the coherence of the video frame set corresponding to each reference product. The server performs weighted fusion on the coherence of the video frame set corresponding to each reference product to obtain the advertising interference degree corresponding to each reference product. The server determines the reference product with the lowest advertising interference degree among the multiple reference products as the target product.

[0123] The weight of weighted fusion is set by the technician according to the actual situation, and the embodiment of the present application does not limit this. The first classification head is a classification head for determining the importance, and the second classification head is a classification head for determining the coherence. The first classification head and the second classification head can adopt the classification heads in the related technology, and the embodiment of the present application does not limit this.

[0124] D. The server determines the target display time of the target advertisement in the target video based on the target video and the target product.

[0125] In a possible implementation, the server determines multiple candidate display times based on a target video frame set corresponding to the target advertisement in the target video. The server uses the multiple candidate display times to segment the target video frame set to obtain multiple sub-video frame sets. The server determines the target display time from the multiple candidate display times based on the multiple sub-video frame sets.

[0126] In order to explain the above implementation more clearly, the above implementation will be further explained in several parts below.

[0127] A. The server determines multiple candidate display times based on a target video frame set corresponding to the target advertisement in the target video.

[0128] In a possible implementation, the server performs key frame identification on the target video frame set to obtain multiple key video frames in the target video frame set. The server determines the play times of the multiple key video frames in the target video as multiple candidate presentation times, where one key video frame corresponds to one candidate presentation time.

[0129] B. The server uses the multiple candidate presentation times to segment the target video frame set to obtain multiple sub-video frame sets.

[0130] In a possible implementation, the server determines the segmented video frames corresponding to each candidate display time in the target video. The server segments the target video frame set based on the multiple segmented video frames to obtain multiple sub-video frame sets, where one sub-video frame set corresponds to one segmented video frame.

[0131] The segmented video frame is the key video frame mentioned above.

[0132] C. The server determines the target presentation time from the multiple candidate presentation times based on the multiple sub-video frame sets.

[0133] In a possible implementation, the server extracts features from each sub-video frame set to obtain sub-video frame set features of each sub-video frame set. The server determines an advertisement display index for each sub-video frame set based on the sub-video frame set features of each sub-video frame set, and the advertisement display index is used to indicate the time suitability of the candidate display time corresponding to the sub-video frame set. The server determines the candidate display time with the highest corresponding advertisement display index among multiple candidate display times as the target display time.

[0134] For example, the server encodes each sub-video frame set based on the attention mechanism to obtain the sub-video frame set features of each sub-video frame set. The server fully connects and normalizes the sub-video frame set features of each sub-video frame set to obtain the advertising display index of each sub-video frame set. The server determines the candidate display time with the highest corresponding advertising display index among multiple candidate display times as the target display time.

[0135] 303. In response to the target video being played to a reference time before the target display time, the server obtains the object image of the target object and the play mode of the target video, where the play mode includes accelerated play, normal play, and slow play.

[0136] Among them, the reference time is also a video playback time, and the time interval between the reference time and the target display time is a preset time interval, which is set by the technical personnel according to the actual situation, and the embodiment of the present application does not limit this. The object image of the target object refers to the image of the target object collected when the target video is played to the reference time, and the object image can reflect the state of the target object watching the target video. Accelerated playback refers to a mode of playback at a first speed, and the first speed is a speed greater than 1. Speed ​​playback refers to a mode of playback at a second multiple, and the second speed is a speed less than 1. Normal playback refers to a mode of playback at a playback speed of 1.

[0137] In a possible implementation, in response to the target video being played to a reference time before the target display time, the server obtains an object image of the target object through an image acquisition component of the terminal, and the server obtains a playback mode of the target video through an application that plays the target video on the terminal.

[0138] Among them, the terminal is the terminal used by the target object.

[0139] 304. The server determines the expression and body movement of the target object based on the object image.

[0140] In a possible implementation, the server performs expression recognition and action recognition on the object image to obtain the expression and body movement of the target object.

[0141] Among them, the methods of expression recognition and action recognition can adopt the methods in the relevant technology, and the embodiments of the present application are not limited to this.

[0142] 305. The server determines whether to place the target advertisement at the target display time based on the playback mode, the expression and body movements of the target object.

[0143] In a possible implementation, the server determines a first delivery judgment parameter based on the playback mode, and the first delivery judgment parameter is used to indicate the delivery suitability associated with the playback mode. The server determines a second delivery judgment parameter based on the expression of the target object, and the second delivery judgment parameter is used to indicate the delivery suitability associated with the expression. The server determines a third delivery judgment parameter based on the body movement of the target object, and the third delivery judgment parameter is used to indicate the delivery suitability associated with the body movement. The server fuses the first delivery judgment parameter, the second delivery judgment parameter, and the third delivery judgment parameter to obtain a fused judgment parameter. When the fused judgment parameter is greater than or equal to the parameter threshold, the server determines to deliver the target advertisement at the target display time. When the fused judgment parameter is less than the parameter threshold, the server determines not to deliver the target advertisement at the target display time.

[0144] Among them, the parameter threshold is set by the technicians according to the actual situation, and the embodiments of the present application do not limit this.

[0145] For example, the server uses the playback mode to query in the first relationship table to obtain the first delivery judgment parameter. The server uses the expression of the target object to query in the second relationship table to obtain the second delivery judgment parameter. The server uses the body movement of the target object to query in the third relationship table to obtain the third delivery judgment parameter. The server weightedly fuses the first delivery judgment parameter, the second delivery judgment parameter, and the third delivery judgment parameter to obtain a fusion judgment parameter. When the fusion judgment parameter is greater than or equal to the parameter threshold, the server determines to deliver the target advertisement at the target display time. When the fusion judgment parameter is less than the parameter threshold, the server determines not to deliver the target advertisement at the target display time.

[0146] The first relationship table stores the correspondence between multiple play modes and multiple first delivery judgment parameters; the second relationship table stores the correspondence between multiple expressions and multiple second delivery judgment parameters; the third relationship table stores the correspondence between multiple body movements and multiple third judgment parameters. The weights of weighted fusion are set by technicians according to actual conditions, and the embodiments of this application do not limit this.

[0147] Optionally, after step 305, the following step 306 or 307 can also be performed.

[0148] 306. When it is determined that the target advertisement is to be delivered at the target display time, the server sends a first message to the terminal. The first message carries the advertisement display information of the target advertisement and the target display time, so that the terminal displays the target advertisement based on the advertisement display information in the first message and the target display time, thereby achieving delivery of the target advertisement.

[0149] 307. When it is determined that the target advertisement will not be delivered at the target display time, the server sends a second message to the terminal. The second message carries the advertisement display information and the end display flag of the target advertisement, so that the terminal displays the target advertisement based on the advertisement display information in the second message when the target video is finished playing, so as to achieve the delivery of the target advertisement.

[0150] The end display flag is used to indicate that the advertisement will be displayed after the target video is finished playing.

[0151] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.

[0152] Through the technical solution provided by the embodiment of the present application, in response to the target video being selected by the target object, the video information of the target video and the object information of the target object are obtained. The target video, video information and object information are used to determine the target advertisement and the target display time from multiple candidate advertisements, thereby realizing the screening of advertisements and the intelligent determination of the advertisement delivery time. When the video is played to the reference time before the target display time, the object image of the target object and the playback mode of the target video are used to determine whether to deliver the target advertisement at the target display time, thereby realizing the intelligent control of advertisement delivery and improving the efficiency of advertisement delivery.

[0153] Figure 4 This is a schematic diagram of the structure of an intelligent advertising delivery system provided by an embodiment of the present application. Figure 4 The system includes: a first acquisition module 401, a determination module 402, a second acquisition module 403 and a judgment module 404.

[0154] The first acquisition module 401 is configured to acquire video information of the target video and object information of the target object in response to the target video being selected by the target object.

[0155] The determination module 402 is used to determine a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, object information of the target object and video information of the target video.

[0156] The second acquisition module 403 acquires the object image of the target object and the play mode of the target video in response to the reference time before the target display time when the target video is played, and the play mode includes accelerated play, normal play and slow play.

[0157] The determination module 404 is used to determine whether to deliver the target advertisement at the target display time based on the object image and the play mode.

[0158] In a possible implementation, the determination module 402 is used to extract features from the object information of the target object to obtain the object features of the target object. Extract features from the video information of the target video to obtain the video information features of the target video. Use the object features of the target object and the video information features of the target video to perform advertisement recall to obtain the multiple candidate advertisements. Based on the target video and the multiple candidate advertisements, determine the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements.

[0159] In a possible implementation, the determination module 402 is used to determine multiple candidate commodities corresponding to the multiple candidate advertisements. Based on the multiple candidate commodities, commodity identification and commodity screening are performed on the target video to obtain multiple reference commodities, the commodity type of the reference commodities belongs to a commodity type set, and the commodity type set is a set of commodity types of the multiple candidate commodities. Based on the target video and the multiple reference commodities, a target commodity is determined from the multiple reference commodities and the candidate advertisement corresponding to the target commodity is determined as the target advertisement. Based on the target video and the target commodity, a target display time of the target advertisement in the target video is determined.

[0160] In a possible implementation, the determination module 402 is used to determine a set of video frames corresponding to each of the reference products in the target video, the set of video frames including a plurality of video frames in which the corresponding reference products appear. Based on the set of video frames, the importance of each of the reference products in the target video and the degree of interference with advertising placement corresponding to each of the reference products are determined, the degree of interference with advertising placement being used to indicate the degree of interference with viewing the target video caused by placing candidate advertisements corresponding to the reference products. Based on the importance of each of the reference products in the target video and the degree of interference with advertising placement corresponding to each of the reference products, a target product is determined from the plurality of reference products.

[0161] In a possible implementation, the determination module 402 is configured to determine a plurality of candidate display times based on a target video frame set corresponding to the target advertisement in the target video, segment the target video frame set using the plurality of candidate display times to obtain a plurality of sub-video frame sets, and determine the target display time from the plurality of candidate display times based on the plurality of sub-video frame sets.

[0162] In a possible implementation, the determination module 402 is used to extract features from each of the sub-video frame sets to obtain sub-video frame set features of each of the sub-video frame sets. Based on the sub-video frame set features of each of the sub-video frame sets, an advertisement display index of each of the sub-video frame sets is determined, and the advertisement display index is used to indicate the time suitability of the candidate display time corresponding to the sub-video frame set. The candidate display time with the highest corresponding advertisement display index among the multiple candidate display times is determined as the target display time.

[0163] In a possible implementation, the judgment module 404 is used to determine the expression and body movements of the target object based on the object image, and determine whether to place the target advertisement at the target display time based on the playback mode, the expression and body movements of the target object.

[0164] In a possible implementation, the judgment module 404 is used to determine a first delivery judgment parameter based on the playback mode, and the first delivery judgment parameter is used to indicate the delivery suitability associated with the playback mode. Based on the expression of the target object, a second delivery judgment parameter is determined, and the second delivery judgment parameter is used to indicate the delivery suitability associated with the expression. Based on the body movements of the target object, a third delivery judgment parameter is determined, and the third delivery judgment parameter is used to indicate the delivery suitability associated with the body movements. The first delivery judgment parameter, the second delivery judgment parameter, and the third delivery judgment parameter are fused to obtain a fused judgment parameter. When the fused judgment parameter is greater than or equal to the parameter threshold, it is determined to deliver the target advertisement at the target display time. When the fused judgment parameter is less than the parameter threshold, it is determined not to deliver the target advertisement at the target display time.

[0165] It should be noted that: the advertising intelligent delivery system provided in the above embodiment only uses the division of the above functional modules as an example when delivering advertisements. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the advertising intelligent delivery system provided in the above embodiment and the advertising intelligent delivery method embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0166] Through the technical solution provided by the embodiment of the present application, in response to the target video being selected by the target object, the video information of the target video and the object information of the target object are obtained. The target video, video information and object information are used to determine the target advertisement and the target display time from multiple candidate advertisements, thereby realizing the screening of advertisements and the intelligent determination of the advertisement delivery time. When the video is played to the reference time before the target display time, the object image of the target object and the playback mode of the target video are used to determine whether to deliver the target advertisement at the target display time, thereby realizing the intelligent control of advertisement delivery and improving the efficiency of advertisement delivery.

[0167] Figure 5 This is a schematic diagram of the structure of a server provided in an embodiment of the present application. The server 500 may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) 501 and one or more memories 502, wherein at least one computer program is stored in the one or more memories 502, and the at least one computer program is loaded and executed by the one or more processors 501 to implement the methods provided in the above-mentioned various method embodiments. Of course, the server 500 may also have components such as a wired or wireless network interface, a keyboard, and an input and output interface for input and output. The server 500 may also include other components for implementing device functions, which will not be described in detail here.

[0168] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including a computer program, and the computer program can be executed by a processor to complete the intelligent advertising delivery method in the above embodiment. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0169] In an exemplary embodiment, a computer program product or a computer program is also provided, which includes a program code, and the program code is stored in a computer-readable storage medium. A processor of a computer device reads the program code from the computer-readable storage medium, and the processor executes the program code, so that the computer device executes the above-mentioned intelligent advertising delivery method.

[0170] In some embodiments, the computer program involved in the embodiments of the present application may be deployed and executed on a computer device, or on multiple computer devices located at one location, or on multiple computer devices distributed at multiple locations and interconnected by a communication network. Multiple computer devices distributed at multiple locations and interconnected by a communication network may constitute a blockchain system.

[0171] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0172] The above are only optional embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for intelligent advertising delivery, characterized in that: The method comprises: In response to a target video being selected by a target object, acquiring video information of the target video and object information of the target object; Determine a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, object information of the target object, and video information of the target video, wherein the plurality of candidate advertisements are advertisements that match the target object and the target video; The step of determining a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, the object information of the target object, and the video information of the target video includes: Extracting features from the object information of the target object to obtain the object features of the target object; extracting features from the video information of the target video to obtain the video information features of the target video; recalling advertisements using the object features of the target object and the video information features of the target video to obtain the multiple candidate advertisements; determining the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements based on the target video and the multiple candidate advertisements; The step of determining the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements based on the target video and the multiple candidate advertisements includes: Determine a plurality of candidate commodities corresponding to the plurality of candidate advertisements respectively; based on the plurality of candidate commodities, perform commodity identification and commodity screening on the target video to obtain a plurality of reference commodities, wherein the commodity type of the reference commodities belongs to a commodity type set, and the commodity type set is a set of commodity types of the plurality of candidate commodities; based on the target video and the plurality of reference commodities, determine a target commodity from the plurality of reference commodities and determine the candidate advertisement corresponding to the target commodity as the target advertisement; based on the target video and the target commodity, determine a target display time of the target advertisement in the target video; The determining a target product from the multiple reference products based on the target video and the multiple reference products includes: Determine a set of video frames corresponding to each of the reference products in the target video, the set of video frames including a plurality of video frames in which the corresponding reference products appear; determine, based on the set of video frames, the importance of each of the reference products in the target video and the degree of interference with advertisement placement corresponding to each of the reference products, the degree of interference with advertisement placement being used to indicate the degree of interference with viewing the target video caused by placing candidate advertisements corresponding to the reference products; determine a target product from the plurality of reference products based on the importance of each of the reference products in the target video and the degree of interference with advertisement placement corresponding to each of the reference products; In response to the target video being played to a reference time before the target presentation time, acquiring an object image of the target object and a play mode of the target video, the play mode including accelerated play, normal play, and slow play; Based on the object image and the play mode, it is determined whether to deliver the target advertisement at the target display time.

2. The method according to claim 1, characterized in that The step of determining a target display time of the target advertisement in the target video based on the target video and the target product includes: Determining a plurality of candidate display times based on a target video frame set corresponding to the target advertisement in the target video; Using the multiple candidate presentation times to segment the target video frame set to obtain multiple sub-video frame sets; Based on the multiple sub-video frame sets, the target presentation time is determined from the multiple candidate presentation times.

3. The method according to claim 2, characterized in that The determining the target presentation time from the multiple candidate presentation times based on the multiple sub-video frame sets includes: Extracting features from each of the sub-video frame sets to obtain sub-video frame set features of each of the sub-video frame sets; Determining an advertisement display index of each of the sub-video frame sets based on the sub-video frame set features of each of the sub-video frame sets, wherein the advertisement display index is used to indicate the time suitability of the candidate display time corresponding to the sub-video frame set; The candidate display time having the highest corresponding advertisement display index among the multiple candidate display times is determined as the target display time.

4. The method according to claim 1, characterized in that: The determining, based on the object image and the playback mode, whether to deliver the target advertisement at the target display time includes: Based on the object image, determining the expression and body movement of the target object; Based on the playback mode, the expression and body movements of the target object, it is determined whether to place the target advertisement at the target display time.

5. The method according to claim 4, characterized in that The determining whether to place the target advertisement at the target display time based on the playback mode, the expression and body movement of the target object includes: Based on the playback mode, determining a first delivery judgment parameter, wherein the first delivery judgment parameter is used to indicate a delivery suitability degree associated with the playback mode; Based on the expression of the target object, determining a second delivery judgment parameter, wherein the second delivery judgment parameter is used to indicate a delivery suitability degree associated with the expression; Based on the body movement of the target object, determining a third delivery judgment parameter, wherein the third delivery judgment parameter is used to indicate a delivery suitability degree associated with the body movement; The first delivery judgment parameter, the second delivery judgment parameter and the third delivery judgment parameter are integrated to obtain an integrated judgment parameter; When the fusion judgment parameter is greater than or equal to the parameter threshold, determining to place the target advertisement at the target display time; When the fusion judgment parameter is less than the parameter threshold, it is determined not to deliver the target advertisement at the target display time.

6. An intelligent advertising delivery system, characterized in that: The system comprises: A first acquisition module, configured to acquire video information of the target video and object information of the target object in response to the target video being selected by the target object; A determination module, configured to determine a target advertisement and a target display time of the target advertisement in the target video from a plurality of candidate advertisements based on the target video, object information of the target object, and video information of the target video; The determination module is used to extract features from the object information of the target object to obtain the object features of the target object; extract features from the video information of the target video to obtain the video information features of the target video; use the object features of the target object and the video information features of the target video to perform advertisement recall to obtain the multiple candidate advertisements; and determine the target advertisement and the target display time of the target advertisement in the target video from the multiple candidate advertisements based on the target video and the multiple candidate advertisements; The determination module is used to determine a plurality of candidate commodities corresponding to the plurality of candidate advertisements respectively; based on the plurality of candidate commodities, perform commodity identification and commodity screening on the target video to obtain a plurality of reference commodities, wherein the commodity type of the reference commodities belongs to a commodity type set, and the commodity type set is a set of commodity types of the plurality of candidate commodities; based on the target video and the plurality of reference commodities, determine a target commodity from the plurality of reference commodities and determine the candidate advertisement corresponding to the target commodity as the target advertisement; based on the target video and the target commodity, determine a target display time of the target advertisement in the target video; The determination module is used to determine a set of video frames corresponding to each of the reference products in the target video, wherein the set of video frames includes a plurality of video frames in which the corresponding reference products appear; based on the set of video frames, determine the importance of each of the reference products in the target video and the degree of interference with advertisement placement corresponding to each of the reference products, wherein the degree of interference with advertisement placement is used to indicate the degree of interference with viewing the target video caused by placing candidate advertisements corresponding to the reference products; and determine a target product from the plurality of reference products based on the importance of each of the reference products in the target video and the degree of interference with advertisement placement corresponding to each of the reference products; A second acquisition module, in response to a reference time before the target video is played to the target display time, acquires the object image of the target object and the play mode of the target video, wherein the play mode includes accelerated play, normal play and slow play; A judgment module is used to determine whether to deliver the target advertisement at the target display time based on the object image and the playback mode.

7. A computer-readable storage medium, characterized in that: At least one computer program is stored in the computer-readable storage medium, and the computer program is loaded and executed by the processor to implement the intelligent advertising delivery method according to any one of claims 1 to 5.

Images