An integrated intelligent advertisement delivery method

Abstract

Embodiments of the present disclosure provide an integrated intelligent advertisement delivery method, applied to the technical field of digital advertisement delivery. The method comprises: configuring delivery basic parameters, establishing a binding relationship between traffic rules and bottom price strategies, and binding a creative template containing a semantic macro placeholder to an advertisement position; after receiving the original material of an advertiser, intelligently processing and generating standard material conforming to the template specification according to the push form; in response to an advertisement filling request, analyzing the advertisement position identifier and matching the winning advertiser according to the traffic rules, and calling the standard material thereof; meanwhile, obtaining the real-time features of the audio-visual content in the play, analyzing the semantic label of the macro placeholder, filling the matched feature value into the macro placeholder, and dynamically synthesizing the final advertisement with the standard material, and pushing the final advertisement to the target advertisement position for rendering and display. In this way, the technical problems of high threshold for advertisement material production, redundant multi-size adaptation and review process, and split of advertisement content and play scene in the prior art can be solved.

Description

Technical Field

[0001] This disclosure relates to the field of digital advertising delivery technology, and in particular to an integrated intelligent advertising delivery method. Background Technology

[0002] Currently, advertising trading platforms generally adopt a model where the demand-side platform submits creative content, and the media outlet reviews and matches ad placements. In practice, on the one hand, because ad placements come in various sizes, the demand-side platform needs to create and push multiple creative materials of different sizes for the same ad content, and the media outlet also needs to review them one by one, resulting in low efficiency for both parties. On the other hand, many small and medium-sized advertisers or individual promoters lack professional design capabilities and find it difficult to independently complete the production of ad creatives that meet the platform's specifications, increasing additional outsourcing costs.

[0003] Furthermore, in video media scenarios, traditional pre-roll ads and corner ads lack intelligent linkage with the content being played. The ad information is presented independently of the plot and scene, resulting in low user acceptance. Moreover, it is impossible to use the show's metadata (such as show title, episode number, and scene objects) to achieve precise contextualized reach.

[0004] While existing technologies include server-side ad insertion and dynamic ad insertion protocols, neither has solved the end-to-end technical challenges, from fragmented material input to intelligent generation, and from automatic multi-size adaptation to real-time semantic integration with playback content. Summary of the Invention

[0005] This disclosure provides an integrated intelligent advertising delivery method that solves the technical problems of high threshold for advertising material production, redundant multi-size adaptation and review processes, and disconnect between advertising content and playback scenarios in the existing technology.

[0006] According to a first aspect of this disclosure, an integrated intelligent advertising delivery method is provided. The method includes: Receive basic configuration instructions for ad placement, establish a binding relationship between traffic rules and base price strategy, and bind creative templates to target ad slots; wherein, the creative templates include basic display parameters and macro placeholders with predefined semantic tags, the macro placeholders being used to identify the data type of audiovisual content associated parameters; Receive the original advertising materials pushed by the advertiser, identify the push format of the original advertising materials, start the corresponding intelligent processing flow according to the identification result, and generate standard advertising materials that conform to the creative template specifications; In response to an ad fill request sent by a media client, the ad slot identifier carried in the ad fill request is parsed, and matching logic is executed based on the traffic rules bound to the ad slot identifier to determine the winning advertiser and retrieve the standard ad creative corresponding to the winning advertiser. The system obtains real-time feature information of the audio-visual content stream currently being played by the media client, parses the predefined semantic tags of the macro placeholders in the creative template, fills the feature values ​​in the real-time feature information that match the predefined semantic tags into the corresponding macro placeholders, and dynamically synthesizes them with the standard advertising materials to generate the final advertising content. The final advertising content is pushed to the target advertising space in the media client for rendering and display.

[0007] In addition to the aspects and any possible implementations described above, a further implementation is provided in which establishing the binding relationship between traffic rules and the base price strategy, and binding the creative template to the target ad slot, includes: Receive at least one base price strategy configured for a single ad slot, the base price strategy including a minimum bid threshold and an applicable subject range, and generate a unique strategy identifier for each base price strategy; Based on the unique policy identifier, a traffic rule is created, and the policy identifier is written into the data structure of the traffic rule, thereby establishing a one-to-one binding relationship between the traffic rule and the base price policy at the data level; The traffic rules also include bidding mode parameters, targeting condition parameters, and validity period parameters. The bidding mode parameters are used to identify one of the following: real-time bidding mode, private programmatic buying mode, or priority trading mode. Configure a unique creative template for the target ad slot, and obtain the unique template identifier of the creative template and the unique rule identifier of the traffic rule; Establish a three-way mapping relationship between the target ad placement identifier, the unique rule identifier, and the unique template identifier; The macro placeholders in the creative template are configured with corresponding predefined semantic tags, which correspond to the field names in the predefined standard data model.

[0008] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, when the bidding mode parameter in the traffic rule is identified as a private programmatic buying mode, the method further includes: Generate a unique transaction agreement identifier; Establish a unique mapping relationship between the transaction agreement identifier, the target ad slot identifier, and the specific demand-side platform identifier, and store this mapping relationship in the rule configuration database for subsequent ad fill request retrieval.

[0009] As described above and in any possible implementation, a further implementation is provided in which, prior to identifying the push format of the original advertising creative, the method further includes: Identify the media type of the original advertising creative; If the media type is video data, there is no need to identify its push format. The video data is uploaded to the transcoding server, encrypted and hashed, and transcoded using a preset video encoding standard. A multi-bitrate playback address list containing multiple bitrates is generated, and an anti-leeching token based on timestamp signature is added to each playback address. The processed video data is then identified as standard advertising material. If the media type is a static image, then its push format is further identified.

[0010] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the step of initiating the corresponding intelligent processing flow based on the identification result to generate standard advertising materials conforming to the creative template specifications includes: When the push notification is identified as a single-format ad creative, the pixel size of the original material and multiple adaptation sizes of the target ad space are obtained. If the pixel size of the original material is larger than the target adaptation size, multiple sizes of material are generated according to a preset center cropping strategy or proportional scaling strategy. If the pixel size of the original material is smaller than the target adaptation size, a solid color background layer is generated, the difference between the original material and the target adaptation size is calculated, the original material is centered and superimposed on the solid color background layer to fill the target adaptation size, and adaptive code adapted to different terminal resolutions is generated. The generated multiple sizes of material and adaptive code are uniformly defined as standard ad creatives. When the push notification is identified as a collection of scattered material packages combined with content descriptions, the image elements and text descriptions in the scattered material packages are parsed and input into a pre-trained material generation model, which outputs multiple candidate ad creatives that conform to the creative template specifications. Automated compliance checks and manual reviews are performed on the candidate ad creatives, and the approved candidate creatives are pushed to the advertiser's terminal. Only after receiving confirmation instructions from the advertiser are the corresponding candidate creatives determined as standard ad materials.

[0011] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the step of performing matching logic based on traffic rules bound to the ad slot identifier to determine the winning advertiser includes: The ad fill request is parsed to extract the ad placement identifier and terminal environment parameters, and the traffic rules bound to the ad placement identifier are retrieved. If the bidding mode indicated by the traffic rule is the real-time bidding mode, then bidding invitations are sent to multiple demand-side platforms, and bidding requests carrying traffic rule identifiers and bid values ​​are received from each demand-side platform; bidding requests with bid values ​​lower than the minimum bid threshold in the bound base price strategy are filtered out, and the remaining bidding requests are sorted according to bid values, and the demand-side platform with the second highest bid value is selected as the winning advertiser according to the second highest price transaction rule; If the bidding mode indicated by the traffic rule is a private programmatic buying mode, then retrieve the transaction protocol identifier associated with the ad slot identifier, and directly select the specific demand-side platform mapped to the transaction protocol identifier as the winning advertiser. If the bidding mode indicated by the traffic rule is the priority transaction mode, then a price inquiry request is first sent to the preset priority demand-side platform; if a valid bid is received from the priority demand-side platform and the valid bid is not lower than the minimum bid threshold in the base price strategy, then the priority demand-side platform is directly selected as the winning advertiser; if no valid bid is received or the bid is lower than the minimum bid threshold, then the real-time bidding mode is switched to execute the above real-time bidding logic.

[0012] In addition to the aspects and any possible implementations described above, a further implementation is provided in which obtaining the real-time feature information of the audiovisual content stream currently being played by the media client includes: Listen for playback status events from media clients or receive playback metadata periodically reported by media clients; The playback metadata or the current playback frame is parsed: if it is structured metadata, the show name, episode number, character tags and scene description text are directly extracted; if it is an unstructured video stream, the image recognition interface is called to extract the scene object features of the current frame, the speech recognition interface is called to extract the audio text features, and the natural language processing interface is called to extract key entity words from the scene description text or audio text. The extracted drama titles, episode numbers, character tags, scene object features, and key entity words are mapped to the corresponding fields in the predefined standard data model. If no valid information is extracted from a certain field, it is left empty, generating a structured feature information dataset with a uniform format. In this context, the field names in the standard data model correspond one-to-one with the predefined semantic tags of the macro placeholders in the creative template.

[0013] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the step of mapping and filling the real-time feature information into the corresponding macro placeholders and dynamically synthesizing it with the standard advertising material includes: Iterate through the macro placeholders in the creative template and parse the predefined semantic tags carried by each macro placeholder; Find the field corresponding to the predefined semantic label in the structured feature information dataset, extract the feature value in the field as the basic feature value, replace the corresponding macro placeholder with the basic feature value, and complete the basic initialization of text information or speech synthesis parameters. Extract scene object features from the structured feature information dataset, convert them into semantic representations, and calculate the semantic similarity between the semantic representation and each keyword in the preset keyword library; Based on the comparison results between the semantic similarity and the preset threshold, the current ad synthesis strategy is dynamically determined: If the maximum similarity exceeds the preset threshold, the ad style associated with the keyword is selected, and the standard ad material is overlaid with the information flow text or badge graphic corresponding to the ad style to generate contextual ad content containing contextual elements. If the maximum similarity does not exceed the preset threshold, the speech synthesis engine is invoked to generate a customized speech stream based on the speech synthesis parameters after basic initialization. The customized speech stream is then mixed with the video track of the standard advertising material, or the text information after basic initialization is rendered to the screen layer of the standard advertising material to generate standard advertising content.

[0014] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the method further includes: When the final advertisement content is rendered and displayed, the exposure monitoring address bound in the creative template is triggered to record the exposure timestamp, device identifier, and ad placement identifier; Monitor user clicks on the final advertisement content. When a click event is detected, trigger the click monitoring address and record the click timestamp and click coordinates. The recorded exposure and click data are structured and encapsulated to generate an ad performance dataset, which is then sent asynchronously to the data analysis server.

[0015] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the method further includes: Receive statistical results returned by the data analysis server, including exposure conversion rate and click-through rate under different traffic rules; Based on the statistical results, parameter adjustment instructions are automatically generated to update the targeting condition parameters in the traffic rules or the minimum bid threshold in the base price strategy; or, Based on the statistical results, template optimization instructions are automatically generated to update the default fill values ​​of macro placeholders or the associated scene keyword weights in the creative template.

[0016] According to a second aspect of this disclosure, an integrated intelligent advertising delivery device is provided. The device includes: The configuration management module is used to receive basic configuration instructions for ad placement, establish the binding relationship between traffic rules and base price strategies, and bind creative templates to target ad slots; wherein, the creative template includes basic display parameters and macro placeholders with predefined semantic tags, and the macro placeholders are used to identify the data type of audiovisual content associated parameters; The material processing module is used to receive the original advertising materials pushed by the advertiser, identify the push format of the original advertising materials, start the corresponding intelligent processing flow according to the identification result, and generate standard advertising materials that conform to the creative template specifications. The bidding matching module is used to respond to the ad fill request sent by the media client, parse the ad slot identifier carried in the ad fill request, execute matching logic based on the traffic rules bound to the ad slot identifier, determine the winning advertiser, and retrieve the standard ad creative corresponding to the winning advertiser. The dynamic synthesis module is used to obtain real-time feature information of the audio-visual content stream currently being played by the media client, parse the predefined semantic tags of the macro placeholders in the creative template, fill the feature values ​​in the real-time feature information that match the predefined semantic tags into the corresponding macro placeholders, and dynamically synthesize them with the standard advertising materials to generate the final advertising content. The push module is used to push the final advertising content to the target advertising position of the media client for rendering and display.

[0017] According to a third aspect of this disclosure, an electronic device is provided. The electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described above.

[0018] According to a fourth aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the methods according to the first and / or second aspects of this disclosure.

[0019] In this disclosure, firstly, by receiving the basic configuration instructions for ad placement, a binding relationship between traffic rules and the base price strategy is established, and creative templates containing predefined semantic tag macro placeholders are bound to the target ad slots. This establishes a clear rule foundation for subsequent automated bidding and dynamic content synthesis, enabling the ad placement system to distinguish between different bidding modes and to pre-set the types of audiovisual content data that need to be dynamically filled for each ad slot.

[0020] Secondly, by identifying the format of the original creative materials pushed by advertisers, the system initiates a corresponding intelligent processing flow—for complete creatives, it directly performs format and size verification and adaptive cropping; for fragmented materials with descriptive text, it calls the generation model to automatically create multiple candidate creatives. This dual processing path significantly lowers the barrier to entry for advertisers while ensuring that the final standard materials meet the specifications of the creative template, avoiding multiple manual cropping and repeated reviews.

[0021] Furthermore, when the media client initiates an ad fill request, the system parses the ad placement identifier and executes matching logic based on the bound traffic rules to determine the winning advertiser and retrieve their standard creative materials. This matching logic can fairly select the optimal advertiser based on the preset minimum price strategy and the second-highest price rule, ensuring the return on investment.

[0022] Building upon this foundation, by further acquiring real-time feature information of the audiovisual content stream currently playing on the media client, and utilizing predefined semantic tags in the macro placeholders of the creative template, the feature values ​​are precisely filled into the corresponding placeholders, dynamically synthesizing the final advertising content with standard materials. This synthesis process is no longer limited to simple size adaptation, but achieves semantic-level fusion of advertising text, voice, and even visual elements with real-time features such as show titles, episode numbers, characters, and scene objects. For example, it can automatically generate a title audio-visual message like "Brand X invites you to watch episode X of show X," or automatically trigger a medicine icon when a cold scene is detected.

[0023] Finally, the synthesized ads are pushed to the target ad placements for rendering and display, and monitoring data is collected to form a closed-loop feedback loop. In summary, through the synergistic effect of multiple stages such as rule binding, intelligent material processing, and semantic-driven dynamic synthesis, not only is the automation level and material processing efficiency of the ad delivery system significantly improved, but the ad content can also be deeply integrated with the playback scenario, improving user experience and ad reach.

[0024] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0025] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. The drawings are provided for a better understanding of the invention and are not intended to limit the scope of this disclosure. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein: Figure 1 A flowchart illustrating an integrated intelligent advertising delivery method provided by an embodiment of this disclosure is shown; Figure 2A structural diagram of an integrated intelligent advertising delivery device provided by an embodiment of this disclosure is shown; Figure 3 A structural diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0027] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0028] In this disclosure, firstly, by receiving the basic configuration instructions for ad placement, a binding relationship between traffic rules and the base price strategy is established, and creative templates containing predefined semantic tag macro placeholders are bound to the target ad slots. This establishes a clear rule foundation for subsequent automated bidding and dynamic content synthesis, enabling the ad placement system to distinguish between different bidding modes and to pre-set the types of audiovisual content data that need to be dynamically filled for each ad slot.

[0029] Secondly, by identifying the format of the original creative materials pushed by advertisers, the system initiates a corresponding intelligent processing flow—for complete creatives, it directly performs format and size verification and adaptive cropping; for fragmented materials with descriptive text, it calls the generation model to automatically create multiple candidate creatives. This dual processing path significantly lowers the barrier to entry for advertisers while ensuring that the final standard materials meet the specifications of the creative template, avoiding multiple manual cropping and repeated reviews.

[0030] Furthermore, when the media client initiates an ad fill request, the system parses the ad placement identifier and executes matching logic based on the bound traffic rules to determine the winning advertiser and retrieve their standard creative materials. This matching logic can fairly select the optimal advertiser based on the preset minimum price strategy and the second-highest price rule, ensuring the return on investment.

[0031] Building upon this foundation, by further acquiring real-time feature information of the audiovisual content stream currently playing on the media client, and utilizing predefined semantic tags in the macro placeholders of the creative template, the feature values ​​are precisely filled into the corresponding placeholders, dynamically synthesizing the final advertising content with standard materials. This synthesis process is no longer limited to simple size adaptation, but achieves semantic-level fusion of advertising text, voice, and even visual elements with real-time features such as show titles, episode numbers, characters, and scene objects. For example, it can automatically generate a title audio-visual message like "Brand X invites you to watch episode X of show X," or automatically trigger a medicine icon when a cold scene is detected.

[0032] Finally, the synthesized ads are pushed to the target ad placements for rendering and display, and monitoring data is collected to form a closed-loop feedback loop. In summary, through the synergistic effect of multiple stages such as rule binding, intelligent material processing, and semantic-driven dynamic synthesis, not only is the automation level and material processing efficiency of the ad delivery system significantly improved, but the ad content can also be deeply integrated with the playback scenario, improving user experience and ad reach.

[0033] Figure 1 A flowchart illustrating an integrated intelligent advertising delivery method provided by an embodiment of this disclosure is shown, such as... Figure 1 As shown, an integrated intelligent advertising delivery method 100 may include the following steps: S110: Receive basic configuration instructions for ad placement, establish a binding relationship between traffic rules and base price strategy, and bind the creative template to the target ad slot; wherein, the creative template includes basic display parameters and macro placeholders with predefined semantic tags, the macro placeholders being used to identify the data type of the audiovisual content associated parameters.

[0034] In some embodiments, establishing the binding relationship between traffic rules and the base price strategy, and binding the creative template to the target ad placement includes: Receive at least one base price strategy configured for a single ad slot, the base price strategy including a minimum bid threshold and an applicable subject range, and generate a unique strategy identifier for each base price strategy; Based on the unique policy identifier, a traffic rule is created, and the policy identifier is written into the data structure of the traffic rule, thereby establishing a one-to-one binding relationship between the traffic rule and the base price policy at the data level; The traffic rules also include bidding mode parameters, targeting condition parameters, and validity period parameters. The bidding mode parameters are used to identify one of the following: real-time bidding mode, private programmatic buying mode, or priority trading mode. Configure a unique creative template for the target ad slot, and obtain the unique template identifier of the creative template and the unique rule identifier of the traffic rule; Establish a three-way mapping relationship between the target ad placement identifier, the unique rule identifier, and the unique template identifier; The macro placeholders in the creative template are configured with corresponding predefined semantic tags, which correspond to the field names in the predefined standard data model.

[0035] Specifically, the ad delivery system first receives basic configuration instructions from operations personnel. For each ad placement to be delivered, the ad delivery system allows the configuration of one or more base price strategies. Each base price strategy includes a minimum bid threshold (in cents) and the scope of the target audience for which the strategy applies, such as specifying whether it applies to a specific demand-side platform or a specific advertiser. The ad delivery system automatically generates a unique strategy identifier for each base price strategy and stores this identifier, along with the ad placement identifier, target audience, and minimum bid threshold, in a price management data structure.

[0036] Subsequently, the ad delivery system creates traffic rules based on strategy identifiers: each traffic rule's data structure is forcibly written with a strategy identifier, thus establishing a one-to-one binding relationship between traffic rules and base price strategies at the data level. This means a traffic rule can only reference one base price strategy, while an ad placement can have multiple base price strategies bound to different traffic rules. Traffic rules also include bidding mode parameters, targeting condition parameters, and validity period parameters. The bidding mode parameter identifies one of the following: real-time bidding mode, private programmatic buying mode, or priority transaction mode. For private programmatic buying mode, the ad delivery system additionally generates a unique transaction agreement identifier and establishes a unique mapping relationship between this transaction agreement identifier, the target ad placement identifier, and the specific demand-side platform identifier, storing this information in the rule configuration database for retrieval by subsequent ad population requests.

[0037] In some embodiments, when the bidding mode parameter in the traffic rule is identified as a private programmatic buying mode, the method further includes: Generate a unique transaction agreement identifier; Establish a unique mapping relationship between the transaction agreement identifier, the target ad slot identifier, and the specific demand-side platform identifier, and store this mapping relationship in the rule configuration database for subsequent ad fill request retrieval.

[0038] Specifically, the ad delivery system configures a unique creative template for each target ad placement. The creative template includes basic display parameters such as the creative address, size, dimensions, landing page address, and monitoring addresses (including impression and click monitoring addresses), as well as extended parameters such as whether it is associated with a TV series, the number of episodes of the associated series, and the audio type (male or female). In particular, the creative template embeds one or more macro placeholders with predefined semantic tags. These macro placeholders are used to identify the data type of the audiovisual content associated parameters. Each predefined semantic tag corresponds to a field name in a predefined standard data model; for example, %vidname corresponds to the TV series name field, %vidnumber corresponds to the episode number field, and %scene_object corresponds to the scene object field. The ad delivery system obtains the unique template identifier of the creative template and the unique rule identifier of the previously created traffic rule, establishing a three-way mapping relationship between the target ad placement identifier, the unique rule identifier, and the unique template identifier, and persisting this mapping. This mapping relationship ensures that when an ad fill request arrives from a media client, the ad delivery system can quickly retrieve the corresponding traffic rule (and thus obtain the base price strategy and bidding mode) and the required creative template based on the ad placement identifier. For image-based content, the ad delivery system supports bmp, jpg, png, gif, and webp formats with a maximum size of 2MB. For video formats, it supports wmv, rm, rmvb, and mov formats with a maximum size of 5MB. These limitations take effect during subsequent content verification. Through the above configuration, the ad delivery system completes the rule preparation before delivery, laying the data foundation for subsequent intelligent content processing, bidding matching, and dynamic composition.

[0039] S120: Receive the original advertising material pushed by the advertiser, identify the push format of the original advertising material, and start the corresponding intelligent processing flow according to the identification result to generate standard advertising material that conforms to the creative template specifications.

[0040] In some embodiments, prior to identifying the delivery format of the original advertising creative, the method further includes: Identify the media type of the original advertising creative; If the media type is video data, there is no need to identify its push format. The video data is uploaded to the transcoding server, encrypted and hashed, and transcoded using a preset video encoding standard. A multi-bitrate playback address list containing multiple bitrates is generated, and an anti-leeching token based on timestamp signature is added to each playback address. The processed video data is then identified as standard advertising material. If the media type is a static image, then its push format is further identified.

[0041] Specifically, after receiving the original ad creative from the advertiser, the ad delivery system first identifies the media type of the creative. If it is determined to be video data, there is no need to further identify its delivery format; the video file is directly uploaded to the transcoding server. The ad delivery system performs encrypted hash verification on the video data (e.g., using the MD5 algorithm to generate a file fingerprint for integrity verification and deduplication), and then transcodes it according to a preset video encoding standard (such as H.265), generating a multi-bitrate playback address list containing various bitrates. For example, it simultaneously outputs CDN URLs, HD URLs, smooth URLs, and the corresponding H.265 version URLs to adapt to the playback needs of clients under different network conditions. After transcoding, the ad delivery system adds a timestamp-signed anti-leeching token to each playback address, raising the anti-leeching level to level 2 to prevent unauthorized use of resources. The video data processed in the above way is directly identified as standard ad creative, without needing to undergo subsequent intelligent processing.

[0042] In some embodiments, the step of initiating a corresponding intelligent processing flow based on the recognition result to generate standard advertising materials that conform to the creative template specifications includes: When the push notification is identified as a single-format ad creative, the pixel size of the original material and multiple adaptation sizes of the target ad space are obtained. If the pixel size of the original material is larger than the target adaptation size, multiple sizes of material are generated according to a preset center cropping strategy or proportional scaling strategy. If the pixel size of the original material is smaller than the target adaptation size, a solid color background layer is generated, the difference between the original material and the target adaptation size is calculated, the original material is centered and superimposed on the solid color background layer to fill the target adaptation size, and adaptive code adapted to different terminal resolutions is generated. The generated multiple sizes of material and adaptive code are uniformly defined as standard ad creatives. When the push notification is identified as a collection of scattered material packages combined with content descriptions, the image elements and text descriptions in the scattered material packages are parsed and input into a pre-trained material generation model, which outputs multiple candidate ad creatives that conform to the creative template specifications. Automated compliance checks and manual reviews are performed on the candidate ad creatives, and the approved candidate creatives are pushed to the advertiser's terminal. Only after receiving confirmation instructions from the advertiser are the corresponding candidate creatives determined as standard ad materials.

[0043] Specifically, if the original ad creative is a static image, the ad delivery system further identifies its delivery format. When the delivery format is identified as a single-format ad creative, meaning the advertiser has directly submitted the complete image creative, the ad delivery system first obtains the pixel size of the original creative and compares it with various adaptation sizes required by the target ad placement (such as 640×100, 320×50, 640×360, 480×320, 640×960, 320×100, etc., commonly used on mobile devices). If the pixel size of the original creative is larger than the target adaptation size, a preset center cropping strategy is used—that is, retaining the central area of ​​the image and discarding the surrounding edges—or a proportional scaling strategy is used to scale the image to the adaptation size while maintaining the aspect ratio. If the pixel size of the original creative is smaller than the target adaptation size, a solid color background layer is automatically generated (a transparent white background is used by default). By calculating the difference in width and height between the original creative and the target adaptation size, the original creative is centered and superimposed on the background layer, so that the background layer fills the target adaptation size. After completing the size adaptation, the ad delivery system also generates adaptive code to suit different terminal resolutions, such as responsive CSS rules or viewport configurations, to ensure that the final output materials can be displayed correctly on different screen sizes. The multi-size materials and adaptive code generated above are uniformly defined as standard ad materials.

[0044] Specifically, when the system identifies a push notification consisting of a fragmented package of creative materials along with a description—meaning the advertiser has only provided a few icons, brand logos, product images, and a short text description of the ad's creative intent (e.g., "A certain lipstick helps you shine in the workplace")—the ad delivery system first parses the image elements and text descriptions within the fragmented material package and inputs them into a pre-trained material generation model. This model, based on a generative adversarial network or diffusion model architecture, can automatically generate multiple candidate ad creatives conforming to the creative template specifications, including images and short videos, based on the input visual elements and semantic descriptions. Typically, the ad delivery system outputs 10 candidate creatives at a time. Subsequently, the system performs automated compliance checks on these candidate ad creatives, including format verification (images must be bmp, jpg, png, gif, webp and ≤2MB in size; videos must be wmv, rm, rmvb, mov and ≤5MB in size) and content sensitivity analysis (e.g., whether they contain prohibited words or violent content). Candidate creatives that pass the initial review are then pushed to media operations personnel for manual review, focusing on confirming whether the content complies with advertising laws and clarity requirements. Once approved, the ad delivery system pushes a list of candidate creatives to the advertiser's terminal. Only after receiving confirmation from the advertiser for one or more creatives will the corresponding candidate creative be finalized as standard ad creative and associated with the corresponding transaction agreement identifier or demand-side platform identifier for subsequent use. This complete intelligent processing flow enables advertisers without professional design capabilities to quickly obtain compliant, high-quality ad creatives, significantly lowering the creative production threshold for ad delivery.

[0045] S130, in response to the ad fill request sent by the media client, the ad slot identifier carried in the ad fill request is parsed, the matching logic is executed based on the traffic rules bound to the ad slot identifier, the winning advertiser is determined, and the standard ad material corresponding to the winning advertiser is retrieved.

[0046] In some embodiments, the step of performing matching logic based on traffic rules bound to the ad placement identifier to determine the winning advertiser includes: The ad fill request is parsed to extract the ad placement identifier and terminal environment parameters, and the traffic rules bound to the ad placement identifier are retrieved. If the bidding mode indicated by the traffic rule is the real-time bidding mode, then bidding invitations are sent to multiple demand-side platforms, and bidding requests carrying traffic rule identifiers and bid values ​​are received from each demand-side platform; bidding requests with bid values ​​lower than the minimum bid threshold in the bound base price strategy are filtered out, and the remaining bidding requests are sorted according to bid values, and the demand-side platform with the second highest bid value is selected as the winning advertiser according to the second highest price transaction rule; If the bidding mode indicated by the traffic rule is a private programmatic buying mode, then retrieve the transaction protocol identifier associated with the ad slot identifier, and directly select the specific demand-side platform mapped to the transaction protocol identifier as the winning advertiser. If the bidding mode indicated by the traffic rule is the priority transaction mode, then a price inquiry request is first sent to the preset priority demand-side platform; if a valid bid is received from the priority demand-side platform and the valid bid is not lower than the minimum bid threshold in the base price strategy, then the priority demand-side platform is directly selected as the winning advertiser; if no valid bid is received or the bid is lower than the minimum bid threshold, then the real-time bidding mode is switched to execute the above real-time bidding logic.

[0047] Specifically, the ad delivery system listens for and responds to ad fill requests sent by media clients in real time. When a request arrives, the ad delivery system first parses the request, extracting the ad placement identifier and terminal environment parameters (such as device model, ad delivery system version, network type, screen resolution, etc.). Based on the ad placement identifier, the ad delivery system retrieves the previously established three-way mapping relationship, obtains the traffic rules bound to the ad placement, and then obtains the base price strategy, bidding mode parameters (real-time bidding mode, private programmatic buying mode, or priority trade mode) and targeting condition parameters referenced by the traffic rules.

[0048] If the retrieved traffic rule indicates a real-time bidding mode, a bidding invitation is immediately broadcast to all connected demand-side platforms (DSPs). The invitation includes a traffic rule identifier and context information about the current ad placement. Each DSP responds with a bidding request, which must include the traffic rule identifier and its bid value (in cents). The ad delivery system first filters all received bidding requests: removing those without the correct traffic rule identifier and those with bids below the minimum bid threshold set in the base price strategy. For the filtered bidding requests, the ad delivery system sorts them by bid value from highest to lowest. Then, based on the second-highest bid rule commonly used in programmatic advertising—that is, the actual transaction price is the highest bidder's bid, but the winner is the second-highest bidder—the system selects the second-highest bidder as the winning advertiser. This mechanism encourages DSPs to bid according to their true intentions, avoiding payment premiums caused by artificially inflated highest bids. After determining the winning advertiser, the advertising system returns a winning bid notification, which includes the transaction price (i.e., the second highest price) and a winning indicator.

[0049] Specifically, if the bidding mode indicated by the traffic rules is a private programmatic buying mode, the advertising delivery system will not initiate a public auction. In this case, the advertising delivery system directly retrieves the transaction agreement identifier associated with the ad slot identifier. This transaction agreement identifier, which was pre-established and stored in the rule configuration database in step S110, uniquely maps to a specific demand-side platform. The advertising delivery system directly selects this demand-side platform as the winning advertiser, without needing to inquire about prices or participate in bidding; the transaction price is executed according to a pre-agreed fixed price. This mode is suitable for high-quality resource transactions that guarantee both price and volume, typically with one media outlet corresponding to one advertiser, and the media outlet promising a volume rebate.

[0050] Specifically, if the bidding mode indicated by the traffic rules is the priority transaction mode, a price request is first sent to a preset priority demand-side platform, which typically has a priority purchase agreement with the media. The advertising system waits for a preset timeout period (e.g., 200ms for the default page type, 120ms for other types). If a valid bid is received from the priority demand-side platform within this preset timeout period, and the bid value is not lower than the minimum bid threshold in the base price strategy, the priority demand-side platform is directly selected as the winning advertiser, and the transaction price is that bid. If no valid bid is received within the timeout period, or the received bid is lower than the minimum bid threshold, the advertising system automatically switches to real-time bidding mode, re-executes the bidding process according to the aforementioned real-time bidding logic, sends bidding invitations to all connected demand-side platforms, and determines the final winning advertiser based on the second-highest price rule.

[0051] Specifically, after the winning advertiser is determined, the ad delivery system retrieves the standard ad creative that has been previously approved and associated with the corresponding traffic rules or transaction agreement from the ad management module based on the winning advertiser's identifier (such as the Demand-Side Platform ID or a specific Demand-Side Platform ID bound in the transaction agreement identifier). Since the standard ad creative has already undergone format verification, size adaptation, or intelligent generation in step S120, and has been manually reviewed and confirmed, it can be directly used for subsequent synthesis and delivery without any repetitive work of format conversion or size adjustment. This design ensures low latency and high efficiency throughout the entire process from winning the bid to creative retrieval, meeting the millisecond-level response requirements of programmatic advertising.

[0052] S140: Obtain real-time feature information of the audio-visual content stream currently being played by the media client, parse the predefined semantic tags of the macro placeholders in the creative template, fill the corresponding macro placeholders with the feature values ​​in the real-time feature information that match the predefined semantic tags, and dynamically synthesize them with the standard advertising materials to generate the final advertising content.

[0053] In some embodiments, obtaining real-time feature information of the audiovisual content stream currently being played by the media client includes: Listen for playback status events from media clients or receive playback metadata periodically reported by media clients; The playback metadata or the current playback frame is parsed: if it is structured metadata, the show name, episode number, character tags and scene description text are directly extracted; if it is an unstructured video stream, the image recognition interface is called to extract the scene object features of the current frame, the speech recognition interface is called to extract the audio text features, and the natural language processing interface is called to extract key entity words from the scene description text or audio text. The extracted drama titles, episode numbers, character tags, scene object features, and key entity words are mapped to the corresponding fields in the predefined standard data model. If no valid information is extracted from a certain field, it is left empty, generating a structured feature information dataset with a uniform format. In this context, the field names in the standard data model correspond one-to-one with the predefined semantic tags of the macro placeholders in the creative template.

[0054] Specifically, after identifying the winning advertiser and retrieving their standard advertising materials, the ad delivery system immediately initiates real-time feature information collection of the audiovisual content stream currently playing on the media client. The ad delivery system obtains this information in two ways: firstly, by monitoring playback status events on the media client (such as playback start, chapter switching, pause, etc.), or by receiving playback metadata proactively reported by the media client at fixed intervals (e.g., every 500 milliseconds); secondly, for scenarios that do not support structured metadata reporting, the ad delivery system directly performs frame extraction on the currently playing video frames.

[0055] Specifically, after acquiring the raw data, the ad delivery system performs intelligent parsing. If the received data is structured metadata (e.g., playback information from the smart TV's ad delivery system interface), it directly extracts the drama title, episode number, character tags, and scene description text. If the received data is an unstructured video stream, the ad delivery system calls multiple analysis interfaces in parallel: it calls an image recognition interface (based on a convolutional neural network-based object detection model) to analyze the currently extracted video frames and output object features in the scene (such as "mask," "tissue," "water cup," "lipstick," etc.); it calls a speech recognition interface to convert the current audio segment into text; and it calls a natural language processing interface to extract key entity words (such as "cough," "makeup," "skincare," "drinking water") from the scene description text or audio transcription text. All extracted information—including drama title, episode number, character tags, scene object features, and key entity words—is uniformly mapped to the corresponding fields in a predefined standard data model. In this standard data model, field names correspond one-to-one with predefined semantic tags for macro placeholders in the creative template. For example, the field "video_name" corresponds to the macro placeholder %vidname, the field "episode_number" corresponds to %vidnumber, and the field "scene_object" corresponds to %scene_obj. If no valid information is extracted from a field, that field is set to null, ultimately generating a structured feature information dataset with a uniform format.

[0056] In some embodiments, mapping and filling the real-time feature information into the corresponding macro placeholders and dynamically synthesizing it with the standard advertising material includes: Iterate through the macro placeholders in the creative template and parse the predefined semantic tags carried by each macro placeholder; Find the field corresponding to the predefined semantic label in the structured feature information dataset, extract the feature value in the field as the basic feature value, replace the corresponding macro placeholder with the basic feature value, and complete the basic initialization of text information or speech synthesis parameters. Extract scene object features from the structured feature information dataset, convert them into semantic representations, and calculate the semantic similarity between the semantic representation and each keyword in the preset keyword library; Based on the comparison results between the semantic similarity and the preset threshold, the current ad synthesis strategy is dynamically determined: If the maximum similarity exceeds the preset threshold, the ad style associated with the keyword is selected, and the standard ad material is overlaid with the information flow text or badge graphic corresponding to the ad style to generate contextual ad content containing contextual elements. If the maximum similarity does not exceed the preset threshold, the voice synthesis engine is called to generate a customized voice stream based on the voice synthesis parameters after basic initialization, and the customized voice stream is mixed with the video track of the standard advertising material or the text information after basic initialization is rendered to the picture layer of the standard advertising material to generate the content of the standard version advertisement.

[0057] Specifically, the advertising placement system then traverses all macro placeholders in the current creative template, and analyzes the predefined semantic tags carried by each macro placeholder (such as %vidname, %vidnumber, %role_name, %scene_desc). Fields corresponding to the tags are searched in the generated structured feature information dataset, and the feature values in these fields are extracted as basic feature values, and then these feature values are used to replace the corresponding macro placeholders one by one. For example, %vidname is replaced with the actual drama name "Blossoms", and %vidnumber is replaced with "Episode 12". This replacement process completes the basic initialization of the text information or voice synthesis parameters. At the same time, the advertising placement system additionally extracts the scene object features (such as "cold" "cough") in the structured feature information dataset, converts the features into semantic representation vectors, and calculates the semantic similarity between the vectors and each keyword in the preset keyword library (such as "cold" "cough" "fever" "headache" "makeup" "lipstick" "skin care" "cream" "drinking" "mineral water", etc.). The similarity calculation can adopt cosine similarity or a semantic matching method based on a pre-trained language model (such as BERT).

[0058] Specifically, based on the comparison between the calculated maximum semantic similarity and a preset threshold (e.g., 0.75), the ad delivery system dynamically determines the current ad synthesis strategy. If the maximum similarity exceeds the preset threshold, it indicates that the current playback scenario is highly relevant to a certain preset keyword, and the ad delivery system selects an ad style associated with that keyword. For example, when keywords such as "cold" or "cough" are detected, a medicine or health-related corner ad style is triggered: the ad delivery system overlays the standard ad creative (usually a brand icon or product image) with the corresponding information flow text (such as "a certain cold medicine, relieves cold symptoms") or corner graphic of the ad style, generating contextualized ad content containing contextual elements, which is finally displayed in the lower left or lower right corner of the video screen as a semi-transparent overlay. If the maximum similarity does not exceed a preset threshold, it indicates that there are no strongly related scene triggering conditions. In this case, the ad delivery system adopts a standard ad synthesis strategy: it calls the text-to-speech (TTS) engine to generate a customized voice stream based on the initialized voice synthesis parameters (including the drama name, episode number, and preset voice type—male or female), such as "A certain shampoo brand invites you to watch episode 12 of 'Blossoms'." Then, it mixes this customized voice stream with the video track of a standard ad creative (usually a 5-second standard video). Alternatively, it renders the initialized text information directly onto the visual layer of the standard ad creative (e.g., as a subtitle overlay) to generate standard ad content. These standard ads are typically played at the beginning or end of the drama, lasting no more than 5 seconds. Combined with episode number prompts, this effectively reduces user aversion to ads. Through this scene-adaptive synthesis, the ad delivery system achieves semantic-level fusion between ad content and playback content, improving both the accuracy of ad targeting and the user's viewing experience.

[0059] S150, the final advertising content is pushed to the target advertising position of the media client for rendering and display.

[0060] In some embodiments, the method further includes: When the final advertisement content is rendered and displayed, the exposure monitoring address bound in the creative template is triggered to record the exposure timestamp, device identifier, and ad placement identifier; Monitor user clicks on the final advertisement content. When a click event is detected, trigger the click monitoring address and record the click timestamp and click coordinates. The recorded exposure and click data are structured and encapsulated to generate an ad performance dataset, which is then sent asynchronously to the data analysis server.

[0061] Specifically, the advertising delivery system pushes the final advertising content dynamically synthesized in step S140—whether it's a corner ad with overlaid contextual layers or a banner ad mixed with a customized audio stream—to the target ad slot on the media client for rendering and display via a real-time transmission protocol. To ensure measurable advertising effectiveness and optimizable delivery strategies, the advertising delivery system automatically performs a series of monitoring and data collection operations during rendering and display.

[0062] Specifically, the moment the final ad content begins rendering and displaying on the client, the ad delivery system immediately triggers the pre-bound exposure monitoring address (usually a URL) in the creative template. The ad delivery system sends an asynchronous request to this address, carrying parameters such as the exposure timestamp (accurate to milliseconds), device identifier (e.g., device ID or IDFA), ad placement identifier, transaction protocol identifier or traffic rule identifier, and the current ad's transaction price. Upon receiving the exposure request, the media client or third-party monitoring server records a valid exposure. Simultaneously, the ad delivery system overlays a transparent click listener layer over the ad display area, continuously monitoring user clicks on the final ad content. When a user click event is detected, the ad delivery system obtains the click coordinates (offset relative to the top-left corner of the ad placement) and the click timestamp, and triggers the click monitoring address bound to the creative template, also sending a data packet containing the click timestamp, click coordinates, device identifier, and ad placement identifier. For ads associated with a landing page address, the ad delivery system also redirects the user's browser or in-app page to that landing page address, completing the full link tracking from exposure to click to conversion.

[0063] In some embodiments, the method further includes: Receive statistical results returned by the data analysis server, including exposure conversion rate and click-through rate under different traffic rules; Based on the statistical results, parameter adjustment instructions are automatically generated to update the targeting condition parameters in the traffic rules or the minimum bid threshold in the base price strategy; or, Based on the statistical results, template optimization instructions are automatically generated to update the default fill values ​​of macro placeholders or the associated scene keyword weights in the creative template.

[0064] Specifically, the ad delivery system structures and encapsulates the collected exposure and click data, generating an ad performance dataset according to a preset data format (such as JSON or Protobuf). This dataset contains complete closed-loop information for a single ad display: from the winning bid price and the traffic rule identifier used, to the terminal environment parameters during the display, and whether the user clicked and the subsequent landing page behavior. The ad delivery system sends this dataset to the data analysis server asynchronously and non-blockingly to avoid network latency or server processing time affecting the main workflow response speed of ad display.

[0065] Specifically, the data analysis server aggregates and performs statistical calculations on the received datasets, generating multi-dimensional statistical results, including but not limited to exposure conversion rate (the ratio of impressions to display times) under different traffic rules, click-through rate (the ratio of clicks to impressions), fill rate and average transaction price under different base price strategies, and user interaction depth under different creative templates (especially different macro placeholder configurations). These statistical results are returned to the advertising delivery system in real time or near real time.

[0066] Specifically, after receiving statistical results from the data analysis server, the ad delivery system automatically triggers the optimization engine. The optimization engine analyzes the statistical results based on preset optimization rules: if the click-through rate (CTR) under a certain traffic rule is significantly lower than the average level of the ad delivery system, the optimization engine automatically generates parameter adjustment instructions to update the targeting parameters in that traffic rule (e.g., narrowing or changing geographic targeting, adjusting the delivery time period) or the minimum bid threshold in the base price strategy (e.g., appropriately lowering the base price to attract more demand-side platforms to participate in bidding, thereby improving the fill rate). If the exposure conversion rate under a certain creative template is excellent, while another template performs poorly, the optimization engine automatically generates template optimization instructions to update the default fill value of macro placeholders in the creative template (e.g., replacing the default brand slogan with the currently best-performing copy) or adjust the weight of related scenario keywords (e.g., increasing the weight coefficient of keywords related to high-conversion scenarios in semantic similarity calculation). These adjustment instructions are confirmed by operations personnel or executed automatically (depending on the automation level configured in the ad delivery system), thus forming a complete intelligent closed loop from delivery, monitoring, analysis to optimization. Through this closed loop, the advertising delivery system can continuously improve itself, enhancing the accuracy and effectiveness of ad delivery while reducing the cost of manual intervention. The entire advertising delivery system's interface supports access from multiple devices, including PCs and mobile devices, allowing operators to monitor and adjust delivery strategies, creative templates, and optimization rules anytime, anywhere.

[0067] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this disclosure is not limited to the described order of actions, because according to this disclosure, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this disclosure.

[0068] The above is an introduction to the method embodiments. The following describes the present disclosure further through device embodiments.

[0069] Figure 2 A structural diagram of an integrated intelligent advertising delivery device provided by an embodiment of this disclosure is shown, such as... Figure 2 As shown, an integrated intelligent advertising delivery device 200 may include: The configuration management module 210 is used to receive basic configuration instructions for ad placement, establish a binding relationship between traffic rules and base price strategies, and bind creative templates to target ad slots; wherein, the creative template includes basic display parameters and macro placeholders with predefined semantic tags, and the macro placeholders are used to identify the data type of audiovisual content associated parameters.

[0070] The material processing module 220 is used to receive the original advertising materials pushed by the advertiser, identify the push format of the original advertising materials, start the corresponding intelligent processing flow according to the identification result, and generate standard advertising materials that conform to the creative template specifications.

[0071] The bidding matching module 230 is used to respond to the ad fill request sent by the media client, parse the ad slot identifier carried in the ad fill request, execute matching logic based on the traffic rules bound to the ad slot identifier, determine the winning advertiser, and retrieve the standard ad creative corresponding to the winning advertiser.

[0072] The dynamic synthesis module 240 is used to obtain real-time feature information of the audio-visual content stream currently being played by the media client, parse the predefined semantic tags of the macro placeholders in the creative template, fill the feature values ​​in the real-time feature information that match the predefined semantic tags into the corresponding macro placeholders, and dynamically synthesize them with the standard advertising materials to generate the final advertising content.

[0073] The push module 250 is used to push the final advertising content to the target advertising position of the media client for rendering and display.

[0074] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0075] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0076] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0077] Figure 3 A schematic block diagram of an electronic device 300 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device 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 the present disclosure described and / or claimed herein.

[0078] Electronic device 300 includes a computing unit 301, which can perform various appropriate actions and processes according to a computer program stored in ROM 302 or a computer program loaded into RAM 303 from storage unit 308. RAM 303 can also store various programs and data required for the operation of electronic device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via bus 304. I / O interface 305 is also connected to bus 304.

[0079] Multiple components in electronic device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of displays, speakers, etc.; storage unit 308, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows electronic device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0080] The computing unit 301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 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 301 performs the various methods and processes described above, such as method 100. For example, in some embodiments, method 100 may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of method 100 described above may be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform method 100 by any other suitable means (e.g., by means of firmware).

[0081] The various implementations of the advertising delivery systems and technologies described above herein can be implemented in digital electronic circuit advertising delivery systems, integrated circuit advertising delivery systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), on-chip advertising delivery systems (SOCs), payload programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include: implementations in one or more computer programs that can be executed and / or interpreted on a programmable advertising delivery system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a stored advertising delivery system, at least one input device, and at least one output device, and transmitting data and instructions to the stored advertising delivery system, the at least one input device, and the at least one output device.

[0082] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0083] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction-executing advertising delivery system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor advertising delivery systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, 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.

[0084] To provide interaction with the user, the advertising delivery system and techniques described herein can be implemented on a computer having: a display device for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including voice input, speech input, or tactile input).

[0085] The advertising delivery systems and technologies described herein can be implemented in computational advertising delivery systems that include backend components (e.g., as a data server), or computational advertising delivery systems that include middleware components (e.g., an application server), or computational advertising delivery systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with the implementations of the advertising delivery systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the advertising delivery system can be interconnected via digital data communication (e.g., a communication network) of any form or medium. Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0086] A computer-based advertising delivery system can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is established by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, a server for a distributed advertising delivery system, or a server incorporating blockchain technology.

[0087] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this disclosure can be achieved, and this is not limited herein.

[0088] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. An integrated intelligent advertising delivery method, applied to an advertising delivery system, characterized in that, include: Receive basic configuration instructions for ad placement, establish a binding relationship between traffic rules and base price strategy, and bind creative templates to target ad slots; wherein, the creative templates include basic display parameters and macro placeholders with predefined semantic tags, the macro placeholders being used to identify the data type of audiovisual content associated parameters; Receive the original advertising materials pushed by the advertiser, identify the push format of the original advertising materials, start the corresponding intelligent processing flow according to the identification result, and generate standard advertising materials that conform to the creative template specifications; In response to an ad fill request sent by a media client, the ad slot identifier carried in the ad fill request is parsed, and matching logic is executed based on the traffic rules bound to the ad slot identifier to determine the winning advertiser and retrieve the standard ad creative corresponding to the winning advertiser. The system obtains real-time feature information of the audio-visual content stream currently being played by the media client, parses the predefined semantic tags of the macro placeholders in the creative template, fills the feature values ​​in the real-time feature information that match the predefined semantic tags into the corresponding macro placeholders, and dynamically synthesizes them with the standard advertising materials to generate the final advertising content. The final advertising content is pushed to the target advertising space in the media client for rendering and display.

2. The method according to claim 1, characterized in that, The process of establishing the binding relationship between traffic rules and the base price strategy, and binding the creative template to the target ad placement, includes: Receive at least one base price strategy configured for a single ad slot, the base price strategy including a minimum bid threshold and an applicable subject range, and generate a unique strategy identifier for each base price strategy; Based on the unique policy identifier, a traffic rule is created, and the policy identifier is written into the data structure of the traffic rule, thereby establishing a one-to-one binding relationship between the traffic rule and the base price policy at the data level; The traffic rules also include bidding mode parameters, targeting condition parameters, and validity period parameters. The bidding mode parameters are used to identify one of the following: real-time bidding mode, private programmatic buying mode, or priority trading mode. Configure a unique creative template for the target ad slot, and obtain the unique template identifier of the creative template and the unique rule identifier of the traffic rule; Establish a three-way mapping relationship between the target ad placement identifier, the unique rule identifier, and the unique template identifier; The macro placeholders in the creative template are configured with corresponding predefined semantic tags, which correspond to the field names in the predefined standard data model.

3. The method according to claim 2, characterized in that, When the bidding mode parameter in the traffic rule is identified as a private programmatic buying mode, the method further includes: Generate a unique transaction agreement identifier; Establish a unique mapping relationship between the transaction agreement identifier, the target ad slot identifier, and the specific demand-side platform identifier, and store this mapping relationship in the rule configuration database for subsequent ad fill request retrieval.

4. The method according to claim 2, characterized in that, Before identifying the push format of the original advertising creative, the method further includes: Identify the media type of the original advertising creative; If the media type is video data, there is no need to identify its push format. The video data is uploaded to the transcoding server, encrypted and hashed, and transcoded using a preset video encoding standard. A multi-bitrate playback address list containing multiple bitrates is generated, and an anti-leeching token based on timestamp signature is added to each playback address. The processed video data is then identified as standard advertising material. If the media type is a static image, then its push format is further identified.

5. The method according to claim 4, characterized in that, The step of initiating the corresponding intelligent processing flow based on the recognition result to generate standard advertising materials that conform to the creative template specifications includes: When the push notification is identified as a single-format ad creative, the pixel size of the original material and multiple adaptation sizes of the target ad space are obtained. If the pixel size of the original material is larger than the target adaptation size, multiple sizes of material are generated according to a preset center cropping strategy or proportional scaling strategy. If the pixel size of the original material is smaller than the target adaptation size, a solid color background layer is generated, the difference between the original material and the target adaptation size is calculated, the original material is centered and superimposed on the solid color background layer to fill the target adaptation size, and adaptive code adapted to different terminal resolutions is generated. The generated multiple sizes of material and adaptive code are uniformly defined as standard ad creatives. When the push notification is identified as a collection of scattered material packages combined with content descriptions, the image elements and text descriptions in the scattered material packages are parsed and input into a pre-trained material generation model, which outputs multiple candidate ad creatives that conform to the creative template specifications. Automated compliance checks and manual reviews are performed on the candidate ad creatives, and the approved candidate creatives are pushed to the advertiser's terminal. Only after receiving confirmation instructions from the advertiser are the corresponding candidate creatives determined as standard ad materials.

6. The method according to claim 2, characterized in that, The step of executing matching logic based on traffic rules bound to the ad slot identifier to determine the winning advertiser includes: The ad fill request is parsed to extract the ad placement identifier and terminal environment parameters, and the traffic rules bound to the ad placement identifier are retrieved. If the bidding mode indicated by the traffic rule is the real-time bidding mode, then bidding invitations are sent to multiple demand-side platforms, and bidding requests carrying traffic rule identifiers and bid values ​​are received from each demand-side platform; bidding requests with bid values ​​lower than the minimum bid threshold in the bound base price strategy are filtered out, and the remaining bidding requests are sorted according to bid values, and the demand-side platform with the second highest bid value is selected as the winning advertiser according to the second highest price transaction rule; If the bidding mode indicated by the traffic rule is a private programmatic buying mode, then retrieve the transaction protocol identifier associated with the ad slot identifier, and directly select the specific demand-side platform mapped to the transaction protocol identifier as the winning advertiser. If the bidding mode indicated by the traffic rule is the priority transaction mode, then a price inquiry request is first sent to the preset priority demand-side platform; if a valid bid is received from the priority demand-side platform and the valid bid is not lower than the minimum bid threshold in the base price strategy, then the priority demand-side platform is directly selected as the winning advertiser; if no valid bid is received or the bid is lower than the minimum bid threshold, then the real-time bidding mode is switched to execute the above real-time bidding logic.

7. The method according to claim 2, characterized in that, The step of obtaining real-time feature information of the audiovisual content stream currently being played by the media client includes: Listen for playback status events from media clients or receive playback metadata periodically reported by media clients; The playback metadata or the current playback frame is parsed: if it is structured metadata, the show name, episode number, character tags and scene description text are directly extracted; if it is an unstructured video stream, the image recognition interface is called to extract the scene object features of the current frame, the speech recognition interface is called to extract the audio text features, and the natural language processing interface is called to extract key entity words from the scene description text or audio text. The extracted drama titles, episode numbers, character tags, scene object features, and key entity words are mapped to the corresponding fields in the predefined standard data model. If no valid information is extracted from a certain field, it is left empty, generating a structured feature information dataset with a uniform format. In this context, the field names in the standard data model correspond one-to-one with the predefined semantic tags of the macro placeholders in the creative template.

8. The method according to claim 7, characterized in that, The step of mapping and filling the real-time feature information into the corresponding macro placeholders and dynamically synthesizing it with the standard advertising material includes: Iterate through the macro placeholders in the creative template and parse the predefined semantic tags carried by each macro placeholder; Find the field corresponding to the predefined semantic label in the structured feature information dataset, extract the feature value in the field as the basic feature value, replace the corresponding macro placeholder with the basic feature value, and complete the basic initialization of text information or speech synthesis parameters. Extract scene object features from the structured feature information dataset, convert them into semantic representations, and calculate the semantic similarity between the semantic representation and each keyword in the preset keyword library; Based on the comparison results between the semantic similarity and the preset threshold, the current ad synthesis strategy is dynamically determined: If the maximum similarity exceeds the preset threshold, the ad style associated with the keyword is selected, and the standard ad material is overlaid with the information flow text or badge graphic corresponding to the ad style to generate contextual ad content containing contextual elements. If the maximum similarity does not exceed the preset threshold, the speech synthesis engine is invoked to generate a customized speech stream based on the speech synthesis parameters after basic initialization. The customized speech stream is then mixed with the video track of the standard advertising material, or the text information after basic initialization is rendered to the screen layer of the standard advertising material to generate standard advertising content.

9. The method according to claim 1, characterized in that, The method further includes: When the final advertisement content is rendered and displayed, the exposure monitoring address bound in the creative template is triggered to record the exposure timestamp, device identifier, and ad placement identifier; Monitor user clicks on the final advertisement content. When a click event is detected, trigger the click monitoring address and record the click timestamp and click coordinates. The recorded exposure and click data are structured and encapsulated to generate an ad performance dataset, which is then sent asynchronously to the data analysis server.

10. The method according to claim 9, characterized in that, The method further includes: Receive statistical results returned by the data analysis server, including exposure conversion rate and click-through rate under different traffic rules; Based on the statistical results, parameter adjustment instructions are automatically generated to update the targeting condition parameters in the traffic rules or the minimum bid threshold in the base price strategy; or, Based on the statistical results, template optimization instructions are automatically generated to update the default fill values ​​of macro placeholders or the associated scene keyword weights in the creative template.

Images