An interactive advertisement delivery system and method based on full-factor cognitive verification and gamified copies

Abstract

The application discloses an interactive advertisement putting system and method based on full-element cognitive verification and game copy, and constructs a technical scheme with an advertisement full-element information structured anchor point as a data basis, a differentiated question generation strategy and a pre-showing-memory-driven puzzle challenge as a cognitive verification means, a budget quantization copy difficulty parameter and a copy leader skill and advertisement anchor point timestamp forced binding as a game-driven mechanism, and a cognitive score cross-module coupling and cross-platform commission tracking as an incentive closed loop. The application improves the cognition of users on advertisement information to full-element deep understanding and impression strengthening, converts puzzle interaction into a cognitive processing tool depending on advertisement memory, and converts the advertisement budget into a quantifiable and challenging game copy difficulty value, so that the advertisement content itself becomes a core driving element of the game mechanism.

Description

Technical Field

[0001] This invention relates to the field of internet advertising technology, and in particular to an interactive advertising delivery system and method based on full-element cognitive verification and gamified copy. Background Technology

[0002] Digital advertising technology has evolved from traditional display ads and rewarded video ads to simple interactive ads. Traditional display ads (such as banner ads and pop-up ads) only present static or dynamic information, which users passively receive, resulting in generally low ad recall and conversion rates. Rewarded video ads incentivize users to complete viewing by offering virtual rewards, but users' understanding of product information often remains superficial, making it difficult to form in-depth cognition. Interactive advertising technologies that have emerged in recent years have attempted to introduce basic interactive forms such as clicks, swipes, and Q&A, but overall, there is still a lack of technical solutions that can systematically improve users' in-depth understanding of advertising information and build a long-term participation economic model.

[0003] A Chinese invention patent with publication number CN110232606A discloses an advertising method that uses fun and rewards to increase user stickiness. Users participate in the advertisement through interactive methods such as watching, listening, moving, and reading; after correct interaction, they receive cash red envelopes or coupon fragments as rewards; fragments can be combined to form complete coupons or traded in the mall. However, this patent provides a simple quiz interaction mechanism but does not propose a technical solution on how to systematically extract all elements of information from the advertising material and automatically generate verification questions. In addition, in this patent and other existing gamified advertising technologies, game elements and advertising content are independent at the information level: the advertiser's budget exists only as a financial parameter and does not establish an algorithmic quantitative conversion relationship with the difficulty parameters of gamified interaction (such as virtual boss health points, expected value of completion); there is no technically binding between the operation events of the game mechanism (such as attack, defense, skill release) and the presentation time of key information in the advertising material, and users can complete game operations and obtain rewards without paying attention to the advertising content; the team cooperation mechanism is not deeply integrated with the advertising interaction, and there is a lack of a combat engine and contribution calculation system for multi-user real-time collaborative challenge of advertising content.

[0004] Furthermore, in existing technologies, each functional module is isolated, and the cognitive verification results are only used for immediate reward distribution judgments. They are not used as key cross-system decision variables to establish coupling relationships with other subsystems such as gamified dungeon combat contributions, cross-platform consumption rebate ratios, and user character attribute growth at the data flow level. Cross-domain integration of user advertising interaction behavior data and external e-commerce platform consumption data has not been achieved, making it impossible to build a two-way closed loop of "internal interaction performance → external consumption incentives → feedback to internal development." A technical system for user advertising creative crowdfunding, adoption and publication, and automatic revenue sharing has not been established. Users are only passive recipients of advertisements or basic interactive participants, and cannot participate in the creation and dissemination of advertising content.

[0005] In existing advertising technology systems, the distribution of advertising tasks is mainly based on user profiles and time dimensions. It has not achieved dynamic matching and sorting of advertising tasks based on geographical location and geofencing, nor has it established a technical channel connecting online advertising interaction behavior with offline store rights and benefits verification. Therefore, it is impossible to track and verify the complete conversion link from advertising interaction to offline consumption. Summary of the Invention

[0006] To overcome the aforementioned problems in the existing technology, this invention proposes an interactive advertising delivery system and method based on full-element cognitive verification and gamified copy.

[0007] The technical solution adopted by this invention to solve its technical problem is: an interactive advertising delivery system based on full-element cognitive verification and gamified dungeons, comprising: The ad creative full-element anchor point association module is used to synchronously obtain a structured anchor point dataset covering all elements of the ad when the advertiser uploads the ad creative; The full-element cognitive verification question bank automatic generation module is used to read various types of anchor data from the anchor association database and automatically match differentiated question generation strategies according to the anchor type. The dungeon skill binding and combat engine module is used to enforce the binding of advertising information presentation and game combat events in the time dimension; The Memory Puzzle Challenge module is used to transform the core ad images uploaded by advertisers into puzzle challenges based on instant memory. The advertiser points pool and difficulty-attempt control module are used to achieve refined management and automated control of advertiser budgets; The user character development and equipment crafting module creates a virtual character for each user, drives character growth through advertising interaction, and provides basic attribute support for team dungeon challenges. The Advertiser Copy Creation and Difficulty Quantification module quantifies and correlates the advertiser's budget with the gamified copy difficulty parameters, enabling a gamified representation of advertising costs. The location-based task distribution and O2O verification module is used to publish advertising tasks, obtain geographical location information for location matching, and perform verification.

[0008] The aforementioned interactive advertising delivery system based on full-element cognitive verification and gamified dungeons includes an anchor point data in the full-element anchor point association module of the advertising material, which includes anchor point type, anchor point content, and timestamp position of the anchor point in the material. The anchor point type includes product performance parameter anchor points, ingredient formula anchor points, celebrity endorsement anchor points, and attribute enumeration anchor points.

[0009] The aforementioned interactive advertising system based on full-element cognitive verification and gamified instances includes a full-element cognitive verification question bank automatic generation module that matches different question generation strategies according to different anchor types. For ingredient and formula anchors, a progressive decomposition strategy is adopted, allowing users to repeatedly and deeply memorize product formulas from different angles. For spokesperson anchors, an association confirmation strategy is adopted, generating questions that confirm the association between the spokesperson and the product. For attribute enumeration anchors, a quantity induction strategy is adopted, counting the number of elements in the list and generating quantity induction questions.

[0010] The aforementioned interactive advertising system based on full-element cognitive verification and gamified dungeons also includes a team-based social and anti-cheating verification module. This module obtains a friend list or an existing team member list through a third-party platform, forms teams in dungeon battles, performs multi-dimensional behavioral correlation analysis, conducts a comprehensive risk score, and classifies and handles cases according to the comprehensive risk score. The cross-platform consumption tracking and differentiated commission module is used for consumption and differentiated commissions on third-party platforms.

[0011] The aforementioned interactive advertising system based on full-element cognitive verification and gamified instances includes an advertiser points pool and a difficulty-attempt control module comprising a points recharge unit, a difficulty and reward configuration unit, a user level matching unit, and a points real-time deduction and automatic removal unit. The points recharge unit is used for pre-recharge via the advertiser's recharge interface. The difficulty and reward configuration unit is used for advertisers to set configuration parameters for each advertisement. The user level matching unit displays puzzles of corresponding difficulty to different users based on their overall user level. The points real-time deduction and automatic removal unit deducts points and updates the advertisement status based on display status and remaining points.

[0012] The aforementioned interactive advertising delivery system based on full-element cognitive verification and gamified instances includes a location-based task distribution and O2O verification module comprising a location data collection and matching unit, a merchant self-service task creation unit, and an O2O rights verification unit. The location data collection and matching unit is used to obtain the user's terminal location and match it with a corresponding task. The merchant self-service task creation unit is used for merchants to submit task creation requests. The O2O rights verification unit is used to generate an electronic verification voucher after the user completes the task.

[0013] The aforementioned interactive advertising delivery system based on full-element cognitive verification and gamified dungeons also includes a user creative crowdfunding and advertising crowdfunding incentive module. The user creative crowdfunding and advertising crowdfunding incentive module is used to collect users' advertising creative content, associate the creative content with the full-element anchor data of the corresponding product, and issue rewards to creators based on the interaction data.

[0014] The aforementioned interactive advertising system based on full-element cognitive verification and gamified instances includes a user creative crowdfunding and advertising crowd-creation incentive module comprising a creative submission unit, a two-level review unit, a creative release and revenue association unit, and a periodic best ad selection unit. The creative submission unit submits advertising creative content including a creative type identifier, associated product ID, and creative text. The two-level review unit reviews the submitted creative content for both merchant and platform approval. The creative release and revenue association unit publishes approved creatives as independent ads, generating cognitive verification questions based on the full-element anchor data of the associated product, integrating the advertiser's points pool budget allocation, and distributing points rewards to creators based on user interaction data. The periodic best ad selection unit periodically calculates a comprehensive score for each creative content by weighting its multi-dimensional performance data, generates a creative content ranking, and distributes points rewards to top-ranked creators.

[0015] The interactive advertising delivery method based on full-element cognitive verification and gamified instances, and based on the aforementioned interactive advertising delivery system, includes the following steps: Merchants publish advertising tasks through a location-based task distribution and O2O verification module, allowing users to discover nearby advertising tasks. Users complete in-depth quizzes using the fully-element cognitive verification question bank automatic generation module, or participate in memory puzzle challenges using the memory puzzle challenge module, earning corresponding points. The advertiser points pool and difficulty-attempt control module deduct points from the advertiser points pool in real time and check whether automatic removal is necessary. Simultaneously, cognitive scores and puzzle performance are updated to the user character development and equipment synthesis modules, improving corresponding attribute values. Users, carrying character attributes and historical cognitive scores, participate in advertiser dungeon challenges through the dungeon skill binding and combat engine modules. The system invites friends to participate through team-based social interaction and anti-cheating verification modules. The dungeon combat engine comprehensively calculates the damage of each member, and BOSS skills are forcibly bound to advertising anchor timestamps. Throughout the dungeon battle, the team-based social interaction and anti-cheating verification modules conduct multi-dimensional behavioral correlation analysis on members imported from third parties within the same team, and perform comprehensive risk scoring based on the analysis results. Based on the comprehensive risk score, tiered actions are taken. After clearing the dungeon, electronic verification vouchers are issued through the location-based task distribution and O2O verification modules. These electronic vouchers can be used for verification at designated stores, or for obtaining differentiated rebates by making purchases on third-party shopping platforms through the cross-platform consumption tracking and differentiated rebate modules.

[0016] The aforementioned interactive advertising method based on full-element cognitive verification and gamified instances also includes the following steps: Creators improve advertising creatives through user creative crowdfunding and advertising crowd-creation incentive modules. After being reviewed and adopted, the creatives are published and automatically connected to the full-element cognitive verification question bank automatic generation module and points pool system. User interaction data of the advertising creatives is continuously tracked, and points rewards are issued to creators based on user interaction data. In each period, the total score of each advertising creative is calculated based on multi-dimensional performance data, a creative advertising ranking is generated, and points rewards are issued to the top-ranked creators.

[0017] The beneficial effects of this invention are: (1) It realizes the systematic cognitive verification and deep memory reinforcement of all elements of advertising information. This invention establishes a structured anchor data system covering multiple types of information such as product performance parameters, ingredient formula, brand slogan, spokesperson information and attribute enumeration, and automatically matches differentiated question generation strategies such as progressive decomposition, association confirmation and quantity summarization for different types of anchors. This requires users to repeatedly process the core information of the advertisement from different semantic perspectives during the question-answering process, which significantly improves the user's understanding depth and memory persistence of all elements of the product.

[0018] (2) A new jigsaw puzzle advertising model based on pre-display memory was established. This invention forces the setting of a pre-display memory window of the complete image before the jigsaw puzzle is fragmented, so that subsequent jigsaw puzzle operations must rely on the user's active recall of the content of the advertising image. This transforms the jigsaw puzzle behavior from a simple image logic game into an effective means of cognitive processing of advertising information, and solves the technical problem of the disconnect between existing jigsaw puzzle advertising and advertising information.

[0019] (3) The advertiser's budget and the difficulty parameters of the gamified dungeon are quantitatively converted and forcibly bound. The present invention automatically converts the advertiser's preset total budget into the virtual boss's health, attack power and expected value for clearing the dungeon through an algorithm, and forcibly binds the release time window of the dungeon boss's skills to the timestamp position of all elements of the anchor data in the advertising material. This technical means ensures that users must concentrate on the core information points of the advertisement in order to effectively cope with the game challenge, fundamentally changing the passive subordinate status of advertising information in the interaction.

[0020] (4) A cross-system incentive closed loop with deep coupling of multi-module data flow was constructed. This invention uses the full-element cognitive score as the key decision variable in the cross-system, and uses it as the core weight parameter of personal damage contribution in dungeon battles, the differentiated calculation factor of cross-platform consumption rebate ratio, and the basis for unlocking consumption rights. This makes the cognitive depth of advertising information directly determine the user's income level in each subsystem, forming a complete technical closed loop from advertising cognition to gamified performance and then to online and offline consumption returns.

[0021] (5) A technical channel for online task distribution and offline rights verification based on geolocation has been realized. This invention enables users to discover and participate in advertising tasks published by nearby merchants through geofencing matching and O2O electronic voucher verification unit. After completing the task, a unique verification code is generated for verification at the offline store, making the complete conversion link from advertising interaction to physical consumption traceable and verifiable.

[0022] (6) It provides an automatic revenue distribution mechanism for user creative crowdfunding and advertising crowdfunding. This invention allows users to submit advertising ideas and publish them as formal advertisements after two-level review. It automatically connects to the full-element anchor point cognitive verification and points pool budget system, automatically calculates and distributes revenue based on the actual user interaction data generated by the creative advertisement, and supports the tracking and revenue sharing of the contribution ratio of multiple collaborative creations. It establishes a complete technical system for users to participate in advertising creation, dissemination and profit.

[0023] The achievement of the above-mentioned technical effects transforms the advertising delivery system from a one-way communication model with "exposure" as the core indicator into a multi-dimensional interactive technology platform with "quantitative verification of cognitive depth" as the engine, gamified dungeons as the driver, and online-offline integration as the closed loop. It is significantly superior to existing technical solutions in terms of user engagement depth, advertising information reach effectiveness, and platform economic sustainability. Detailed Implementation

[0024] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be described in detail below with reference to specific embodiments.

[0025] The interactive advertising system based on full-element cognitive verification and gamified dungeons of the present invention adopts a cloud-native architecture with front-end and back-end separation and supports deployment on multiple platforms (WeChat Mini Program, Douyin Mini Program, standalone APP, H5 page).

[0026] The system server is deployed on a cloud computing platform, and the specific configuration is as follows: Application server: A cloud server instance with a 4-core CPU, 8GB memory, and 100GB SSD cloud storage, deployed with the Spring Boot microservice framework. A single instance supports no less than 2000 concurrent connections. In the production environment, it is managed through a Kubernetes cluster, enabling automatic elastic scaling based on real-time QPS, with a minimum of 3 replicas and a maximum of 20 replicas. Database server: Utilizing a MySQL 8.0 master-slave replication architecture, configured with one master and two slave nodes. The master node handles write operations, while the two slave nodes asynchronously synchronize via binlog, handling read operations and data analysis queries respectively. Each node is configured with an 8-core CPU, 32GB of memory, and 500GB of SSD storage, with innodb_buffer_pool_size set to 70% of the physical memory. Cache server: Deployed as a Redis 7.0 Cluster cluster with 6 nodes (3 masters and 3 slaves), each node configured with 4 CPU cores and 16GB of memory. It is used to cache frequently accessed data such as user fragment inventory, ad recommendation lists, and activity leaderboards. The cache key design follows the naming convention of "business domain:entity type:entity ID". Object storage: Cloud object storage service is used to store advertising videos, advertising images, user-uploaded creative images and original images of masked dating, with CDN acceleration and AES-256 server-side encryption for original image storage; Message service: Deploy a RabbitMQ message queue cluster to asynchronously process non-real-time tasks such as points distribution, order callbacks, and leaderboard updates, ensuring that the response time of core interfaces is not affected by non-critical paths.

[0027] The client is developed based on the uniapp framework, and a single codebase compiles to output WeChat Mini Programs, Douyin Mini Programs, standalone apps (iOS / Android), and H5 pages. It supports iOS 12.0 and above, and Android 8.0 and above, with automatic screen resolution adaptation. The front-end and back-end communicate via HTTPS, with data transmission in JSON format. API version numbers are managed through URL path prefixes (e.g., / api / v1 / ).

[0028] The entire site uses SSL / TLS 1.3 encrypted transmission and deploys a web application firewall to defend against common web attacks such as SQL injection and cross-site scripting (XSS). Sensitive user data (such as mobile phone numbers, shipping addresses, and payment information) is encrypted and stored using the AES-256 algorithm at the database layer, with the keys centrally managed and rotated periodically by a key management service. Risk control data, such as device fingerprint collection and behavioral baseline analysis, is stored independently in a secure domain database, physically isolated from other business data.

[0029] This embodiment of the interactive advertising delivery system includes a module for linking all elements of advertising materials to anchor points, a module for automatically generating a question bank for full-element cognitive verification, a module for memory puzzle challenges, a module for an advertiser points pool and a difficulty-attempt control system, a module for user character development and equipment synthesis, a module for creating and quantifying the difficulty of advertiser dungeons, a module for binding dungeon skills and a combat engine, a module for team social interaction and anti-cheating verification, a module for location-based task distribution and O2O redemption, a module for cross-platform consumption tracking and differentiated commission rebates, and a module for user creative crowdfunding and advertising crowd-creation incentives.

[0030] I. Anchor Point Association Module for All Elements of Advertising Creatives This module is the system's basic data entry point and is configured to synchronously acquire a structured anchor dataset covering all elements of the advertisement when the advertiser uploads the advertisement material.

[0031] Advertisers upload ad creatives (videos ≤ 30 seconds, images ≥ 800×800 pixels) through the merchant interface, and the system provides an anchor data entry interface. The anchor data structure includes three core fields: anchor type, anchor content, and the timestamp position of the anchor in the creative.

[0032] Anchor point types include, but are not limited to, the following five categories: Product performance parameter anchors: record the core technical indicators of the product, such as the car's 0-100 km / h acceleration time, wheelbase, and driving range. An example of the data structure is {type: "performance", content: "0-100 km / h acceleration 3.8 seconds", timestamp: "00:18"}. Ingredient formula type anchor: Records the core ingredients or formula information of the product, such as the vitamin content and electrolyte content of a beverage. The data structure example is {type: "ingredient", content: "Triple vitamins + triple electrolytes", timestamp: "00:08"}; Brand slogan type anchor: Records the core promotional slogan in the advertisement. The data structure example is {type: "slogan", content: "Only one degree of electricity per night", timestamp: "00:15"}; Spokesperson type anchor: Records the brand spokesperson information appearing in the advertisement. The data structure example is {type: "spokesperson", content: "Wu Lei", timestamp: "00:05"}; Attribute enumeration type anchor: Records multiple attribute values of the product shown in the advertisement, such as flavor types and color options. The data structure example is {type: "enumeration", content: ["Lime flavor", "Peach flavor", "Grape flavor", "Pineapple flavor"], timestamp: "00:12"}.

[0033] After all anchor data is submitted, it is stored in the anchor data association database table anchor_data. This table contains the following fields: Unique identifier of anchor data, Associated advertisement identifier, Anchor type (distinguished by enumeration values into five types: performance parameter type, ingredient formula type, brand slogan type, spokesperson type, attribute enumeration type), Anchor content (stored in JSON format, storing different structured JSON data according to different anchor types), Time position where the anchor appears in the material (recorded in seconds), Creation time. This table establishes a foreign key association with the advertiser material table through the associated advertisement identifier field, and creates an index for the associated advertisement identifier field to improve query efficiency.

[0034] For the JSON data structure of the anchor content field, it is defined separately according to the anchor type: Performance parameter type: Stores three sub-fields: parameter name, parameter value, and unit; Ingredient formula type: Stores a text description sub-field; Brand slogan type: Stores a text sub-field; Spokesperson type: Stores a name sub-field; Attribute enumeration type: Stores an item array sub-field.

[0035] II. Automatic generation module for the full-element cognitive verification question bank This module reads various types of anchor data in the anchor data association database and automatically matches different question generation strategies according to the anchor type.

[0036] A progressive decomposition strategy is used for ingredient / formula-related anchors: After reading the content of the ingredient / formula-related anchors, the system executes a semantic decomposition algorithm. Taking the anchor data of the Pulse beverage advertisement as an example, the ingredient / formula-related anchor record is the text "Triple Vitamins + Triple Electrolytes". The specific execution steps of the generation algorithm are as follows: Step 1: The system uses a natural language processing word segmentation component to identify the core component words "vitamin" and "electrolyte" from the anchor text, and also identifies the quantity modifier "triple" corresponding to each component. The system stores the two sets of "component-quantity" correspondences in a temporary data structure.

[0037] Step 2: The system generates the first-level question – a single-ingredient confirmation question. The system obtains the product name associated with the advertisement and selects the first component information from the temporary data structure in Step 1. The system calls the question template “One {Product Name}, ( ) {Ingredient Name}”, substituting the product name and ingredient name into the template to generate the final question. The correct answer is set to the quantity value corresponding to the ingredient. The rule for generating distractors is: based on the quantity value, within the range of integers from the base value minus 2 to the base value plus 2, two integers not equal to the quantity value are randomly selected as distractors. For example, if the base value is 3 and the range is 1 to 5, then distractors “2” and “4” may be generated.

[0038] Step 3: The system generates the second-level question – a multi-component combination confirmation question. The system detects that the anchor text contains information about two components, and calls the composite question template “One-piece {Product Name}, ( ) weight {Component 1} and ( ) weight {Component 2}”, substituting the product name and the two component names to generate the final question. The correct answer is set as a combination of two quantity values, presented in the format of “First Quantity, Second Quantity”. This question requires the user to accurately memorize the quantity information of two components simultaneously, increasing the depth of cognitive processing.

[0039] Step 4: The system generates the third-level question – a reverse association question between brand and ingredients. The system uses the product name as a variable, calls the question template “One Bite ( ), {Quantity}{Ingredient Name}”, and substitutes the quantity and ingredient name to generate the final question. The correct answer is set to the product name. Distractor options are randomly selected from other brand names registered in the same category (beverages) within the system.

[0040] Through the above three-tiered progressive breakdown, users need to repeatedly process the core information of the advertisement from different angles to achieve a deep memory of the ingredients, formula, and selling points.

[0041] The system employs a confirmation strategy based on the anchor point of the spokesperson: After reading the anchor point content, it generates a question confirming the association between the person and the product. The question template is "Who is the spokesperson for this advertisement ( )", and the correct answer is the name of the person in the anchor point content. Distractor options are randomly selected from the system's preset spokesperson database, choosing other figures with similar popularity to the correct answer to prevent users from eliminating distractors based solely on common sense.

[0042] For the quantitative induction strategy targeting attribute enumeration anchors: After reading the content of the attribute enumeration anchors, the system counts the number of elements in the enumeration list and generates a quantitative induction question. The question template is "How many flavors are mentioned in this advertisement?", and the correct answer is set as the element count result of the enumeration list. Distractor options are randomly generated within ±2 of the correct answer.

[0043] All automatically generated questions are stored in batches in a MySQL question bank table. This table contains the following fields: unique question identifier, associated anchor identifier, associated ad identifier, question stem text, correct answer, distractors (multiple distractors stored in JSON array format), question type (identified by strings as progressive decomposition, associated confirmation, and quantity summarization), difficulty level (increasing from 1 to 3 levels), and creation time. The question bank table has indexes on the associated ad identifier and associated anchor identifier fields to support fast queries by ad dimension and anchor dimension.

[0044] When a user finishes watching an advertisement and triggers the quiz, the frontend sends a request to the backend question loading interface, carrying the advertisement identifier as a parameter. The backend queries the question database table for all question records associated with that advertisement, randomly shuffles the question list, and returns it to the frontend for rendering and display. The order of the options for each question on the frontend is also randomly arranged to avoid the option position affecting the answer result.

[0045] III. Memory Puzzle Challenge Module This module transforms the core ad images uploaded by advertisers into a jigsaw puzzle challenge based on instant memory, including a pre-show timing unit, a dynamic fragmentation engine, a jigsaw puzzle interaction unit, and a jigsaw puzzle verification and reward unit.

[0046] (1) Preview Timer Unit: When a user triggers the memory puzzle challenge, the front-end sends a request to the back-end puzzle start interface, carrying the advertising identifier as a parameter. The back-end returns the access address of the core advertising image uploaded by the advertiser and the challenge configuration parameters, including the preset display duration and difficulty level. The front-end renders a full-screen image display interface, with the image centered and the countdown numbers displayed at the top of the page.

[0047] The countdown function is implemented through a separate Web Worker background thread in the browser, ensuring that the countdown is not interrupted when switching pages or locking the phone screen. The Web Worker thread receives the total countdown duration parameter sent by the main thread and immediately starts a timed loop. Every 100 milliseconds, the Web Worker thread decrements the remaining time by 100 milliseconds and sends the current remaining time back to the main thread to update the interface display. When the remaining time reaches zero, the Web Worker thread clears the timer and sends a timer completion signal to the main thread. Upon receiving this signal, the main thread immediately hides the original image display interface and calls the fragmentation interface to enter the jigsaw puzzle challenge stage.

[0048] (2) Dynamic fragmentation engine: Fragmentation is completed in the backend to reduce the computational burden on the frontend device. After receiving the puzzle start request, the backend determines the grid density according to the difficulty level parameter: the easy level uses 3 rows × 3 columns for a total of 9 pieces, the medium level uses 4 rows × 4 columns for a total of 16 pieces, and the hard level uses 5 rows × 5 columns for a total of 25 pieces.

[0049] The backend uses a Java image processing library to perform pixel-level cropping of the original image. The specific processing logic is as follows: First, the original image file is loaded into a memory buffer. The pixel width (original image width divided by the number of grid columns) and pixel height (original image height divided by the number of grid rows) of each fragment are calculated. Then, a nested loop iterates through each row and column, using an image cropping method to extract the corresponding coordinate region from the original image. The starting point of the coordinates is the row number multiplied by the fragment height and the column number multiplied by the fragment width. Each cropped sub-image is considered a fragment, numbered sequentially from 0 to the total number of fragments minus 1.

[0050] After all fragments are extracted, the system executes the Fisher-Yates shuffle algorithm to randomly rearrange the numbered array. The algorithm's logic is as follows: Create an array of sequential numbers from 0 to the total number of fragments minus 1. Then, starting from the last element of the array, traverse backwards. For each position, use a secure random number generator to generate a random integer no greater than the current position index as the swap position, swapping the element at the current position with the element at that random position. After this traversal, a completely randomized fragment order is obtained. The system returns the shuffled fragment order and the object storage access address of each fragment to the front end.

[0051] (3) Puzzle Interaction Unit: After receiving the fragment data, the front end renders the puzzle grid. The grid layout is implemented using CSS Grid technology, with each grid unit being a receiving area and the fragment images being draggable elements.

[0052] On the desktop, interaction is implemented using the HTML5 native drag-and-drop API. Fragment elements are enabled with draggable properties and a drag start event is bound to record the unique identifier of the dragged fragment. The target grid cell is bound to a drag hover event to prevent the browser's default behavior (allowing placement) and a place event to receive the identifier of the dragged fragment. When a fragment is placed into a grid cell that matches its identifier, the fragment snaps to that grid cell and plays a flashing green animation to indicate success; when a fragment is placed into a mismatched grid cell, it automatically bounces back to its original position and plays a flashing red animation to indicate an error.

[0053] On mobile devices, touch events are used for interaction. A touch start event is bound to the fragment element to record the coordinates of the touch origin and the identifier of the touched fragment. A touch move event is bound to update the fragment's position coordinates in real time as it follows the finger's movement. A touch end event is bound to detect whether the finger's coordinates when released are within the coordinate range of a certain grid cell: if within the range and the fragment's identifier matches the grid cell's identifier, the fragment snaps back to its original position; if they do not match, the fragment smoothly bounces back to its original position with an animation; if the release coordinates are outside the range of any grid cell, the fragment also bounces back to its original position.

[0054] The jigsaw puzzle timing also uses a separate Web Worker background thread, which is not affected by the main thread's UI rendering, ensuring timing accuracy.

[0055] (4) Puzzle Verification and Reward Unit: After the user completes all the puzzle pieces and puts them in the correct positions, the front end collects the following parameters: time spent on the puzzle (in seconds), difficulty level, and preset total time. The front end submits the above parameters to the back end reward calculation interface.

[0056] The backend executes the following calculation logic: First, it obtains the corresponding difficulty coefficient based on the difficulty level (easy level coefficient is 1.0, medium level coefficient is 1.5, and hard level coefficient is 2.0). The basic reward is calculated by multiplying the difficulty coefficient by 10 points. The time bonus is calculated as follows: subtract the ratio of used time to total time from 1, take the larger value between this difference and 0, multiply it by the basic reward, and finally multiply it by a time bonus ratio factor of 0.5. The final reward is the sum of the basic reward and the time bonus, rounded to the nearest integer. In other words, the faster the user completes the puzzle, the higher the time bonus reward.

[0057] After the reward calculation is complete, the system executes the following sequence of atomic operations within the same database transaction: First, it deducts the corresponding number of points from the advertiser's points pool account and updates the account balance; then, it adds the corresponding points to the user's points account; next, it inserts a consumption record in the points transaction log table containing fields such as user ID, ad ID, reward points, transaction time, and transaction type; then, it increments the total number of impressions in the ad configuration table by 1; finally, it checks whether the advertiser's points pool balance is lower than the points required for the minimum difficulty reward of the ad, and whether the total number of impressions has reached or exceeded the set maximum number of impressions. When any of these conditions are met, the system updates the ad's status to "removed" and sends a notification to the advertiser via the messaging service that the budget has been exhausted or the number of impressions has been used up.

[0058] IV. Advertiser Points Pool and Difficulty-Attempt Control System This module enables refined management and automated control of advertisers' budgets, including a points recharge unit, a difficulty and reward configuration unit, a user level matching unit, and a real-time points deduction and automatic removal unit.

[0059] (1) Points recharge unit: Advertisers pre-charge through the recharge interface of Alipay and WeChat Pay SDK, and the funds are deposited into the advertiser's independent points pool account in the form of points.

[0060] The advertiser points pool account table contains the following fields: unique account identifier, associated advertiser identifier (unique constraint), total points balance, frozen points, available points, and last update time (automatically refreshed to the current time as records are updated).

[0061] The ad configuration table contains the following fields: configuration unique identifier, associated ad identifier (unique constraint), single reward points for easy difficulty, single reward points for medium difficulty, single reward points for hard difficulty, maximum total number of impressions, maximum number of impressions per user, total number of impressions already displayed (default value is 0), and ad status (distinguished by enumeration values ​​into three states: active, paused, and exhausted).

[0062] Difficulty and Reward Configuration Unit: Advertisers set configuration parameters for each ad in the backend, including: the single-ad reward amount corresponding to different difficulty levels (e.g., 10 points for easy, 20 points for medium, and 30 points for hard), the maximum total number of ad impressions, and the maximum number of impressions for the same user. Configuration data is stored in the ad_config table.

[0063] (2) User Level Matching Unit: The system displays puzzles of corresponding difficulty to different users based on their overall user level. The formula for calculating the user's overall user level is: userLevel = 0.4 * (mean cognitive score / 100) + 0.3 * (total character attribute value / attribute cap) + 0.3 * (historical dungeon completion rate) High-level users are matched with difficult puzzles, medium-level users are matched with medium-difficulty puzzles, and low-level users are matched with easy puzzles.

[0064] First, obtain the user's total character attribute value from the character data table, divide it by the system's preset upper limit of attribute value for normalization, and obtain an attribute coefficient between 0 and 1.

[0065] Then, the average historical cognitive score of the user is obtained from the cognitive score statistics table, and divided by the full score to normalize it to a cognitive coefficient between 0 and 1.

[0066] Next, obtain the total completion rate of all dungeons participated in by the user from the dungeon completion record table, which is the number of completions divided by the number of participations, and then normalize it into a dungeon coefficient.

[0067] Finally, calculate the overall score: multiply the cognitive coefficient by a weight of 0.4, the attribute coefficient by a weight of 0.3, and the copy coefficient by a weight of 0.3, and add the three together to get the overall score.

[0068] The rating criteria are as follows: users with a comprehensive rating score of not less than 0.7 are assigned to the difficult level (Level 3); users with a comprehensive rating score of not less than 0.4 but less than 0.7 are assigned to the medium level (Level 2); and users with a comprehensive rating score less than 0.4 are assigned to the easy level (Level 1).

[0069] (3) Real-time deduction of points and automatic removal of the unit: After each user completes the puzzle and receives the reward, the system performs the following operations in the transaction: First, query the corresponding advertising configuration record according to the advertising identifier to obtain parameters such as the difficulty reward configuration, maximum number of times to be displayed and the current number of times to be displayed for the advertising; The second step is to query the advertiser's identifier by associating the advertisement identifier with the advertiser's identifier, and then query the advertiser's points pool account records to obtain the current available balance; The third step is to deduct the points required for this reward from the available balance and update the total points balance accordingly. Fourth step, increment the total number of times the ad has been displayed in the ad configuration table by 1; Fifth, insert a new consumption record into the points consumption log table, including information such as operation time, associated advertising identifier, and number of points consumed; Step 6: Execute the delisting condition judgment: If the available balance after deduction is lower than the lowest value among the rewards of each difficulty of the advertisement, or the total number of times it has been displayed has reached or exceeded the set maximum total number of times it has been displayed, then the status of the advertisement will be updated to "exhausted". At the same time, the advertisement status update interface will be called to suspend the display of the advertisement to the outside world, and a reminder message of budget exhaustion or number of times used up will be pushed to the advertiser through the notification service.

[0070] All the above database operations are completed within the same transaction. If any step fails, the entire process is rolled back to ensure data consistency.

[0071] V. User Character Development and Equipment Crafting Module This module creates a virtual character for each user, drives character growth through advertising interactions, and provides basic attribute support for team dungeon challenges.

[0072] When a user first enters the platform, the system automatically creates a character record and stores it in the user_characters table, which includes fields such as user_id, intelligence, memory, reaction, and endurance. Users can earn intelligence experience points by completing a full-element cognitive quiz; memory experience points by completing a memory puzzle challenge; and reaction experience points by completing a timed quiz.

[0073] The equipment crafting module guides users to acquire fragment materials by watching designated advertisements. The data structure for equipment crafting tasks is recorded in the `equipment_recipe` table, which includes: equipment ID, equipment name, required fragment type and quantity, base crafting probability, and guaranteed success rate. After collecting all the fragments, the user performs the crafting operation. The crafting algorithm includes a probability factor and a guaranteed success mechanism—if the user's consecutive crafting failures reach the guaranteed success threshold, the next crafting attempt will be forcibly considered successful.

[0074] VI. Advertiser Copy Creation and Difficulty Quantification Module This module quantifies and correlates advertiser budgets with gamified dungeon difficulty parameters, enabling a gamified representation of advertising costs.

[0075] Advertisers set parameters through the instance creation interface on the merchant side: instance name, associated ad creative ID, task time limit (e.g., every Friday 20:00-22:00), team size range (3-5 people), and preset total budget. The system automatically calculates the instance difficulty parameters based on the total budget. BOSS HP H = Total budget * HP coefficient (default 1.5) BOSS ATK A = Base ATK * (1 + Material duration factor) Expected total output required to clear the stage E = H (i.e., all party members need to deal damage equal to BOSS HP) VII. Dungeon Skill Binding and Combat Engine Module This module is one of the core technical features of this system, enabling the forced binding of advertising information presentation with game battle events in the time dimension.

[0076] (1) Instance skill binding: Advertisers can configure multiple skills for virtual bosses in the instance skill configuration interface. The release time window of each skill is bound to the entered anchor point data through the "Associate Anchor Point" drop-down selection box. The system obtains the timestamp of the anchor point from the time position field recorded by the selected anchor point, and uses the timestamp as the center and extends several seconds before and after it (3 seconds before and after by default) as the skill release time window.

[0077] Skill data is stored in the Virtual Boss Skills table, which contains the following fields: unique skill identifier, associated copy identifier, skill name, skill damage value, associated anchor point identifier, start and end seconds of the release window, and end seconds of the release window. This table is linked to the copy quest table and the anchor point data table via foreign keys for the copy identifier and anchor point identifier, respectively.

[0078] Taking a car advertisement as an example: the time position of the wheelbase parameter anchor point in the advertisement video is at 18 seconds. A skill called "Spatial Compression" is bound to this anchor point, and its release window starts at 15 seconds and ends at 21 seconds. During the 15th to 21st second of the dungeon battle, the virtual boss releases this skill. A shield blocking interface pops up, requiring team members to click the shield button and select the correct wheelbase value within a limited time (e.g., within 2 seconds) to successfully block the skill's damage. If the correct selection is not made within the time limit, the entire team will suffer the damage caused by the skill.

[0079] (2) Dungeon Combat Engine: After multiple users enter a dungeon in a team, the system starts a real-time synchronous combat session based on WebSocket. Each member's valid action (such as answering a question correctly or successfully blocking a BOSS skill) generates a damage calculation: Member damage value D = Base damage * Character attack attribute coefficient * Equipment bonus coefficient * Cognitive score coefficient; The cognitive score coefficient is calculated as 0.5 + the user's historical average cognitive score / 200. This means users with a full cognitive score receive a 1.0x bonus, while users with low cognitive scores receive only a 0.5x base bonus. The system accumulates the real-time damage values ​​of all members. When the total accumulated damage is greater than or equal to the BOSS's health (i.e., the expected total output required to clear the dungeon), the dungeon is considered cleared.

[0080] The dungeon battles utilize the WebSocket protocol for real-time bidirectional communication. The server uses the WebSocket module of the Spring framework to establish a message channel. Each dungeon team subscribes to an independent WebSocket topic address upon entering battle, with the address format being: dungeon battle topic prefix plus dungeon identifier plus team identifier.

[0081] During combat, team members' real-time actions are transmitted between the client and server in JSON format. The core fields of the message include: message type (such as member action), user identifier, action type (such as correct answer), base damage value caused by this action, action timestamp, and other information.

[0082] The server-side combat engine maintains a thread-safe memory-mapped table, using the team identifier as the key and the team's cumulative damage (using atomic long integers to ensure concurrency safety) as the value. Upon receiving a valid action message from a member, the following processing logic is executed: First, retrieve the user's average cognitive score from the user cognitive score statistics and calculate the cognitive score bonus coefficient. The calculation formula is: 0.5 plus the result of dividing the average cognitive score by 200, so that the value of the coefficient is between 0.5 and 1.0. Users with a full cognitive score can get a 1.0x damage bonus, while users with a lower cognitive score can only get a 0.5x base bonus.

[0083] Then, the user's total character attribute value is retrieved from the user character data table, and the character attribute bonus coefficient is calculated according to a preset formula. Next, the bonus rate of the user's current equipment is retrieved from the user equipment data table, and the total equipment bonus coefficient is calculated. The base damage is multiplied sequentially by the cognitive score coefficient, character attribute coefficient, and equipment bonus coefficient to calculate the actual damage value of this operation. This actual damage value is added to the team's cumulative damage, and the updated cumulative damage is broadcast to all online members of the team via a WebSocket message channel. The message includes the currently achieved damage, the total expected value required to clear the level, and the current progress percentage.

[0084] After each damage increment, it is determined whether the accumulated damage has reached or exceeded the expected total output value required to clear the dungeon. If it has, the dungeon clear settlement process is triggered, including: stopping the receipt of damage messages for the team, broadcasting a clear notification to all team members, and calling the reward distribution module to generate and issue consumption rights tokens to each team member.

[0085] VIII. Team-based social interaction and anti-cheating verification module This module includes an external social relationship import unit, a cross-domain team matching unit, a behavior consistency verification unit, and a dynamic risk interception unit.

[0086] (1) External Social Relationship Import Unit: Users authorize the platform to access their friend list on third-party social or gaming platforms through the OAuth 2.0 authorization process. Taking WeChat as an example: After the user clicks the "Import Game Friends" button, the system guides the user to the WeChat authorization page. After the user confirms the authorization, WeChat returns a temporary authorization code. The backend uses this authorization code to call the credential acquisition interface of the WeChat Open Platform to obtain an access token, and then uses the access token to call the WeChat relationship link interface to obtain the user's friend list data.

[0087] The acquired friend data is compared and matched with the platform's user system: if a friend has already registered on the platform, they are marked as an "inheritable teammate"; if a friend has not yet registered, their information is temporarily stored and an invitation record with an invitation link is generated.

[0088] Matched friend data is stored in the Imported Friend Relationship Table, which includes the following fields: unique identifier of the relationship record, user identifier of this platform, unique identifier of the friend on the third-party platform, source platform type identifier, and import time. This table has a joint index on the user identifier and platform type combination, supporting quick queries of imported friend lists from specific platforms by user dimension.

[0089] In the dungeon team-up interface, the system queries currently online, unteamed imported friends and highlights them with the "Inheritable Teammates" tag. Users can send team-up invitations to these friends with a single click. Invitation messages are delivered via third-party platform application messaging channels or the platform's in-app push notifications.

[0090] (2) Behavior consistency verification unit: During the dungeon battle, the system performs multi-dimensional behavior correlation analysis on members of the same team imported through a third party, and comprehensively assesses whether the team has the risk of script manipulation or mass cheating.

[0091] Dimension 1 – Operation Timing Correlation Detection. The system extracts the timestamp sequence of each member's operations throughout the entire combat session from the operation log and calculates the time interval between adjacent operations. It then calculates the standard deviation and average of the operation intervals between all members. If the average interval consistently falls below 100 milliseconds (this threshold represents the lower limit of normal human reaction time), it indicates that multiple people's operations are almost synchronously triggered rather than naturally asynchronously responded, and the system increases the risk score for this dimension by 40 points.

[0092] Dimension Two – Device and Network Environment Similarity Detection. The system obtains the fingerprint hash values ​​of the login devices of each team member through the device fingerprint query interface, and also obtains the location of each member's network IP address and the identifier of the connected Wi-Fi network. The system counts the number of valid devices after deduplication: if the number of deduplicated devices is less than 0.3 times the total number of members (i.e., most members share very few devices), it indicates suspicious characteristics of batch manipulation, and the system increases the risk score for this dimension by 30 points.

[0093] Dimension Three – Geographic Location-Behavioral Logic Consistency Detection. The system obtains the GPS location information of each team member and analyzes the geographic location relationships between members and between each member and the offline redemption stores pointed to by the advertisements associated with the instance. If multiple members from geographically distant areas frequently team up to complete the same instance, and their GPS information has no physical connection with the store location for a long period of time, the system increases the risk score for this dimension by 30 points.

[0094] (3) Dynamic Risk Interception Unit: Based on the comprehensive risk score generated by the above three-dimensional behavioral consistency verification, the system automatically executes graded handling measures: High-risk handling – When the comprehensive risk score reaches or exceeds 70 points, the system will immediately freeze the team-making function of all suspicious members in the team, restrict them from creating or joining new teams, and require members to complete facial recognition verification or SMS verification code secondary verification within a specified time. Team-making permissions can only be restored after the verification is passed.

[0095] Medium-risk handling – When the overall risk score is between 40 and 70, the system applies a 0.5x reduction coefficient to the actual damage dealt by the team in the dungeon battle, which means that the damage output of all members is halved, greatly increasing the difficulty of clearing the dungeon; at the same time, the dungeon clearing rewards for the team are delayed for 2 hours, during which time they enter the manual review queue.

[0096] Low-risk handling – When the overall risk score is between 20 and 40, the system will delay the distribution of dungeon completion rewards to the team for 30 minutes. During this period, the system will conduct an automated behavior review, and the rewards will be automatically distributed after the review is passed.

[0097] IX. Location-Based Task Distribution and O2O Verification Module This module includes a geolocation collection and matching unit, a merchant self-service task creation unit, and an O2O rights and benefits verification unit.

[0098] (1) Geographic Location Acquisition and Matching Unit: After user authorization, the system obtains the real-time latitude and longitude coordinates of the user's terminal through GPS or network positioning. The system queries the geofence defined by the effective_area field in the advertising task table, filters out active tasks within the geofence where the user's current location is located, and displays them in order of distance from nearest to farthest. The data structure of each advertising task contains an effective geographic range field set by the advertiser, which is defined by the center coordinate point and the radius (in meters).

[0099] After the user grants location permissions, the front end calls the location acquisition interface of the WeChat mini program or the browser's built-in location interface to obtain the latitude and longitude coordinates of the current device, and reports the coordinate data to the backend location update interface.

[0100] The backend stores user coordinates in Redis cache as key-value pairs. The cache key adopts the naming format of "location prefix + user identifier" and the expiration time is set to 15 minutes. After the expiration time, it is automatically cleared to protect user privacy.

[0101] The system then performs a matching query for nearby tasks, using MySQL's spatial distance calculation function for filtering. The query logic is as follows: It filters all tasks with a status of "active" from the advertising task table, calculates the spherical distance between the center coordinates of the effective geographic range configured for each task and the user's current coordinates, and returns only records whose distance value does not exceed the effective geographic radius of the task. The query results are sorted in ascending order of distance, and the first 20 results are returned to the front end.

[0102] When the front end displays the list of nearby tasks, each task card shows the task name, associated merchant name, distance (in meters or kilometers), task type tag, and a summary of the rewards that can be obtained.

[0103] (2) Merchant self-service task creation unit: Nearby merchants submit task creation requests through the merchant terminal. The data structure includes: task type selection (cognitive quiz / memory puzzle / fragment synthesis), task reward settings (free item / full reduction coupon / discount coupon), redemption store location, effective geographical range of the task, and total task budget. The system automatically calculates the task issuance limit based on the total budget.

[0104] O2O Rights Verification Unit: After a user completes a task, the system generates an electronic verification voucher containing a unique verification code (generated with a UUID) and validity period information. Merchants scan the verification code to call a verification interface. The system verifies the validity of the verification code, whether it has been used, and whether it is within its validity period. Once verification is successful, it is marked as "verified," and the verification time, GPS coordinates of the verification store, and operator information are recorded.

[0105] After a user completes a task (such as answering cognitive questions or completing a memory puzzle), the system calls the verification code generation service to generate an electronic verification voucher for the user that can be used at the merchant's offline stores.

[0106] The logic for generating the verification code is as follows: A globally unique string is generated using a universally unique identifier generator. After removing hyphens, the first 12 characters are extracted and converted to uppercase letters to improve readability. Then, a verification record is created, and the generated verification code, user identifier, associated task identifier, verification status (initially "unused"), validity period (default current time plus 30 days), and associated QR code image address (converted from the verification code to a QR code image via a QR code generation service) are written to the verification record table and persistently stored.

[0107] After the merchant operator scans the QR code presented by the user, the merchant sends a request to the backend verification interface, carrying the verification code and the merchant's store identifier as parameters. The backend executes the following verification process: First, it queries the verification record table by the verification code to confirm the record exists; then it checks whether the verification status is "unused"; finally, it checks whether the current time is before the valid expiration time. After all verifications pass, the system updates the verification status of the verification record to "verified," records the verification operation time, the verification store identifier, and the store's GPS coordinates, and returns a verification success response.

[0108] 10. Cross-platform consumption tracking and differentiated commission module This module is implemented through a jump tracking unit, an order callback listening unit, and a differentiated commission calculation unit.

[0109] When a user clicks on a product entry point on a third-party shopping platform within the platform, the system generates a redirect link with a unique tracking identifier (trace_id), redirecting the user to the corresponding product page on the third-party e-commerce platform. The system registers a callback listener through the third-party platform's affiliate API interface. Once the user completes order placement, payment, and order receipt, the affiliate interface asynchronously pushes an order status update.

[0110] After the order is confirmed and signed for, the commission calculation unit calculates a differentiated commission bonus based on the user's historical recognition score and copy performance: Final commission points = Base commission points * (1 + Cognitive score bonus + Dungeon performance bonus) Where: Cognitive score bonus = 0.3 * (Average user cognitive score / 100) Dungeon performance bonus = 0.2 * (User dungeon completion rate) XI. User Creative Crowdfunding and Advertising Crowdfunding Incentive Module This module includes a creative submission unit, a two-level review unit, a creative release and revenue linkage unit, and a periodic best ad selection unit.

[0111] (1) Creative Submission Unit: Users submit advertising creative content through the front-end rich text editor or image upload component. The data structure includes: creative type identifier (ad copy / visual design / promotion plan), associated product ID, creative text and attachments. Users can choose to mark the creative as "public participation". The system records the complete creative evolution chain. When other users iterate and modify the creative, the system records the contribution ratio of each contributor in the form of a version chain.

[0112] Users select the creative type (slogan, visual design, promotion plan) through the creative submission interface, fill in the creative text, and upload reference images as attachments. If a user selects the "Allow public participation" option, the system marks the creative as a collaborative creation mode, allowing other users to iteratively modify it.

[0113] Version management data for collaborative creation is stored in the Creation Versions collection of the MongoDB document database. Each document contains two core fields: a Creation ID and a Version Chain array. The Version Chain array records the contributor ID, a snapshot of the contributed content, and the submission timestamp for each version in chronological order. The system also maintains a contribution percentage mapping table for each creation, recording the percentage of contribution each contributor makes to the current final version.

[0114] When a user submits modifications to a publicly submitted creative work, the system compares the modified content with the original content to calculate the percentage of the changes made relative to the original. Based on this percentage and the quality score of the modifications, the system updates the contribution percentage for each contributor: the original author's contribution percentage decreases with each modification by others using a certain decay factor, but always remains above a certain minimum percentage (e.g., 20%) to protect the core rights of the original creator.

[0115] (2) Two-level review unit: The creative submitted by the user is first pushed to the merchant's review interface of the corresponding advertiser. The review dimensions include product fit, creative novelty, and feasibility. After passing the merchant's review, it enters the platform's review queue, where a third-party AI content security API is integrated to screen for illegal content. At the same time, the platform's operations personnel conduct the final review.

[0116] Merchant Review Stage: The system pushes the creative to be reviewed to the corresponding advertiser's merchant-side review interface. The review interface displays the full content of the creative and provides reviewers with a three-dimensional scoring system—novelty (1 to 5 points), product fit (1 to 5 points), and feasibility (1 to 5 points), which merchants score item by item. Creatives that achieve a total score of 12 or higher across the three categories pass the merchant review. Creatives that pass the merchant review automatically enter the platform's review queue; creatives that fail the review are returned to the user's end with suggested modifications.

[0117] Platform review stage: The platform integrates a third-party AI content security service interface to automatically detect violations in creative text and images (including sensitive content such as pornography, political content, and terrorism). Creatives that pass the AI ​​detection undergo a final manual review by platform operators, and are marked as "adopted" after confirmation that they do not violate the platform's content guidelines.

[0118] (3) Creative Release and Revenue Linkage Unit: After approval, the creative will be released as an independent advertisement on the platform. The system will automatically generate cognitive verification questions based on the full-element anchor data of the corresponding product and connect it to the advertiser's points pool budget allocation. The system will continuously track the user interaction data of the creative advertisement and automatically calculate and distribute points rewards to the creator based on preset weights such as views, average cognitive score, and conversion rate.

[0119] Once approved, creative ads are published on the platform as standalone ads. The system automatically associates them with all the anchor data of the corresponding product and generates cognitive verification questions. At the same time, it accesses the points pool budget set by the advertiser or allocated by the system by default.

[0120] The system automatically calculates the revenue of each creative advertisement and distributes rewards via a scheduled task (executed every Sunday at 2:00 AM). The calculation logic is as follows: The system queries a list of all active creative ads. For each active creative ad, it extracts statistics across four dimensions: total cumulative user views / interactions, average user perception score, conversion rate (number of users who completed a specified conversion action divided by total number of interacting users), and number of user-initiated votes.

[0121] The weighted formula for the overall score is as follows: the natural logarithm of the total number of interactions multiplied by a weight of 0.3, plus the average cognitive score multiplied by a weight of 0.25, plus the conversion rate multiplied by a factor of 100 and a weight of 0.25, plus the natural logarithm of the number of votes multiplied by a weight of 0.2. Taking the natural logarithm of the total number of interactions and votes is to smooth out the impact of extremely large values ​​on the score, making the score more stable and comparable.

[0122] The base reward amount is calculated by multiplying the overall score by the number of points per point (default is 100 points per point) and then rounding up.

[0123] If the idea is created by a single person, the full base reward will be awarded to that creator. If the idea is created collaboratively by multiple people, the system will split the base reward according to the recorded contribution ratio: the points earned by each contributor will be the base reward multiplied by the contribution ratio of that contributor, and any decimals after splitting will be rounded down.

[0124] All reward points are distributed to each contributor's points account one by one through the points distribution service interface, and the distribution details are recorded in the points transaction log table.

[0125] (4) Periodic Best Ad Selection Unit: The system calculates the overall score for each creative ad on a weekly / monthly basis using a weighted scoring algorithm based on multi-dimensional performance data (total user interactions, average cognitive score, conversion rate, and number of user votes). The formula for the overall score is: Score = 0.3 * ln(Total Interactions + 1) + 0.25 * Average Cognitive Score + 0.25 * Conversion Rate + 0.2 * ln(Number of Votes + 1) The system automatically generates a creative ad ranking list, awards extra points to top-ranked creators, and prioritizes displaying winning ads in the "Creative Ranking" section on the platform's homepage.

[0126] Based on the above system, the specific interaction method is as follows: Step 1: Discover nearby advertising tasks When a user opens the app, a location authorization prompt appears. After the user agrees, the system obtains the current GPS coordinates (longitude 120.39, latitude 36.07) through the device's positioning module and uploads the coordinates to the backend location update interface. The backend writes the coordinates to the Redis cache and simultaneously queries for active advertising tasks within a 3-kilometer radius, sorting them by distance from nearest to farthest and returning the results.

[0127] The user's homepage displays the three closest tasks in the "Nearby Tasks" list. The first task is "XX Milk Tea Shop - New Product Tasting Task," tagged as "Knowledge Quiz," with a distance of "150m" and a reward of "Complete the quiz to receive a free signature drink." The user clicks to enter the task details page.

[0128] Step 2: Complete the comprehensive cognitive quiz. The task details page automatically plays a 15-second advertisement video for a new milk tea product. The video displays information such as the product name, tea base type (white peach oolong), optional sweetness levels (sugar-free, half-sugar, and full-sugar), and brand ambassador (a well-known celebrity).

[0129] After the video finishes playing, the system loads three questions from the full-element anchor question bank associated with the advertisement, arranges them randomly, and displays them. The questions are as follows: The first question is a progressive breakdown question about ingredient formulation. The system has automatically generated the question based on the anchor data: "What type of tea base does this new product use?" The correct option is "White Peach Oolong," and the distractors include "Four Seasons Spring" and "Jasmine Green Tea."

[0130] The second question is a spokesperson association confirmation question. The question stem is: "Who is the spokesperson appearing in the advertisement?" The correct option is the name of a certain artist, and the distractors are two other historical spokespeople of the brand.

[0131] The third question is a quantitative induction question of attribute enumeration. The question stem is: "How many sweetness levels are mentioned in the advertisement?" The correct option is "3 types", and the distractor options are "2 types" and "4 types".

[0132] The user answered each question carefully and correctly, and the system determined that the cognitive score was 100%.

[0133] Step 3: Challenge your memory puzzle After completing the cognitive quiz, the task page displays a prompt to "Continue with the memory puzzle challenge to earn extra points." The user clicks to enter the puzzle challenge.

[0134] The system matches and displays a puzzle challenge with a difficulty level of "medium (4×4)" based on the advertiser's difficulty configuration and the user's current overall level (the user has a high historical cognitive score, the role attributes are at a medium to high level, and the overall level is determined to be medium).

[0135] Before the puzzle begins, the system displays a high-resolution promotional image of the milk tea shop's signature drink on the screen for 5 seconds. A countdown progress bar appears at the top of the page with the prompt "Please remember this image." The countdown is precisely timed via a separate background thread and is unaffected by page switching.

[0136] After the 5-second countdown ends, the system immediately cuts the image into 16 4×4 fragments on the backend. After the Fisher-Yates shuffling algorithm completely scrambles the fragment order, the fragment list is returned to the frontend for rendering and display.

[0137] Users, relying on a brief memory of the advertisement image, drag and drop individual pieces to their desired positions. The system monitors the matching degree between the pieces and the target positions in real time. When a piece is dragged into the correct position within the tolerance range, it automatically snaps into place and displays a green confirmation animation. After 45 seconds, the user accurately places all 16 pieces to complete the puzzle. The system records the actual time taken and calculates the reward: the basic reward is 15 points (medium-level coefficient 1.5×10), and an additional 5 points are awarded for completing the puzzle ahead of the preset 90-second time limit, for a total reward of 20 points.

[0138] Step 4: Obtain electronic verification vouchers and verify them offline. After the user completes the cognitive quiz and jigsaw puzzle challenge, the system automatically generates an electronic verification voucher. The system calls the verification code generation service to generate a unique 12-digit verification code and a corresponding QR code image. The verification record status is "unused," and the validity period expires 30 days later.

[0139] The user walks to the XX Milk Tea shop 150 meters away and shows the cashier their electronic verification voucher QR code. The cashier scans the QR code using the merchant's app, and the merchant sends the resulting verification code along with the store's logo to the backend verification interface. The backend performs three checks: verification of the verification code's existence, usage status, and validity period. Once all checks pass, the verification status is updated to "verified," and the timestamp and store coordinates are recorded. The user successfully receives a free signature drink.

[0140] Step 5: Participate in the dungeon challenge The milk tea brand simultaneously launched a team-based challenge called "New Product Promotion Week" on the platform. The challenge is open every Friday from 8:00 PM to 10:00 PM, and teams need to have 3 to 5 members. Users can view the challenge in the challenge plaza and click the "Team Challenge" button.

[0141] The user had previously authorized the import of their in-game friend relationship chain from "Honor of Kings". The system found that friend B and friend C were currently online but not in a team in the imported friend list, and highlighted these two friends with the "Inheritable Teammate" tag on the team interface. The user sent a team invitation to them with one click, and the invitation information was pushed through the WeChat application's messaging channel. After clicking the invitation link, friends B and C entered the platform and joined the same team, and the three-person team was formed.

[0142] After the three players entered the instance, the instance's combat engine activated, synchronizing the combat session in real time. An advertisement video played concurrently with the battle. When the video reached the section introducing the tea's origin (corresponding to the time frame of the ingredient formula anchor), the virtual boss unleashed the "Tea Ceremony Quiz" skill. Simultaneously, a timed quiz window popped up on all three players' screens. They had 3 seconds to correctly answer "Which region does this tea originate from?" to block the skill's full-team damage. Because the three players had already deeply memorized the tea information from the advertisement during step 2, they all successfully blocked the attack within the time limit.

[0143] The combat engine continuously receives all valid actions from the three members, calculating each member's cognitive score coefficient, character attribute coefficient, and equipment bonus coefficient in turn. It then calculates the actual damage value of each action and adds it to the team's total damage. After approximately 12 minutes of cooperative combat, the team's cumulative damage reaches the virtual boss's health threshold, and the system determines that the dungeon is cleared. The system invokes the reward distribution service to generate a "Second Drink Half-Price Consumption Voucher" for each of the three members, valid for 7 days. The voucher details include applicable store information and a unique redemption code.

[0144] Step 6: User Creative Crowdfunding Revenue User A came up with a more appealing creative idea based on the advertising slogan of a milk tea brand, based on their daily life experience. The user opened the app, went to the "Creative Square" module, clicked the "Submit Idea" button, selected "Slogan" as the creative type, filled in the new advertising slogan, selected the milk tea brand as the associated product, checked the "Allow public participation" option, and then submitted the idea.

[0145] The creative idea was automatically pushed to the merchant's review interface of the milk tea brand. The brand scored the new advertising slogan in three dimensions: "novelty", "product fit" and "feasibility", with a total score of 14 points (≥12 points), and passed the merchant review.

[0146] After passing the merchant's review, the idea automatically enters the platform's review queue. After passing the violation detection by the AI ​​content security interface, the platform's operations staff will conduct a final manual review to confirm that it complies with the content guidelines and mark it as "adopted".

[0147] The creative ad was subsequently published as a standalone ad on the platform. The system automatically linked it to the milk tea brand's existing full-element anchor data, generated corresponding cognitive verification questions, and allocated a basic points pool budget.

[0148] Within a week, the creative ad garnered 5,000 user interactions (views and quizzes), with an average user awareness score of 92 and a conversion rate of 8%. The system performed revenue calculations during a scheduled task at midnight Sunday: substituting the data into the comprehensive scoring formula, the total interaction volume (approximately 8.52 after taking the natural logarithm) multiplied by 0.3 equals 2.556; the average awareness score of 92 multiplied by 0.25 equals 23; the conversion rate of 8 multiplied by 0.25 equals 2; and the number of votes was initially zero. The total comprehensive score was approximately 27.556. The base reward was 27.556 multiplied by 100 points per point, rounded up to 2756 points. Since this creative was created by a single person without any collaborating contributors, all 2756 points were credited to User A's points account.

[0149] Meanwhile, the creative ad's overall score ranked among the top three in this week's "Creative Ranking," and the system automatically displayed it on the platform's homepage "Creative Ranking" section, giving it more frequent exposure and recommendations.

[0150] As can be seen from the above complete operation process, this system deeply couples modules such as full-element cognitive verification, memory puzzle, dungeon challenge, LBS redemption, cross-domain social team formation and user creative incentives at the data flow level, realizing a complete technical closed loop from deep reach of advertising information to realization of user value.

[0151] The above embodiments are merely exemplary embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art can make various modifications or equivalent substitutions to the present invention within its scope and spirit, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of the present invention.

Claims

1. An interactive advertising delivery system based on full-element cognitive verification and gamified instances, characterized in that, include: The ad creative full-element anchor point association module is used to synchronously obtain a structured anchor point dataset covering all elements of the ad when the advertiser uploads the ad creative; The full-element cognitive verification question bank automatic generation module is used to read various types of anchor data from the anchor association database and automatically match differentiated question generation strategies according to the anchor type. The dungeon skill binding and combat engine module is used to enforce the binding of advertising information presentation and game combat events in the time dimension; The Memory Puzzle Challenge module is used to transform the core ad images uploaded by advertisers into puzzle challenges based on instant memory. The advertiser points pool and difficulty-attempt control module are used to achieve refined management and automated control of advertiser budgets; The user character development and equipment crafting module creates a virtual character for each user, drives character growth through advertising interaction, and provides basic attribute support for team dungeon challenges. The Advertiser Copy Creation and Difficulty Quantification module quantifies and correlates the advertiser's budget with the gamified copy difficulty parameters, enabling a gamified representation of advertising costs. The location-based task distribution and O2O verification module is used to publish advertising tasks, obtain geographical location information for location matching, and perform verification.

2. The interactive advertising delivery system based on full-element cognitive verification and gamified dungeons according to claim 1, characterized in that, The anchor data in the full-element anchor association module of the advertising material includes anchor type, anchor content, and timestamp position of the anchor in the material. The anchor type includes product performance parameter anchors, ingredient formula anchors, celebrity endorsement anchors, and attribute enumeration anchors.

3. The interactive advertising delivery system based on full-element cognitive verification and gamified dungeons according to claim 2, characterized in that, The automatic question generation module of the all-element cognitive verification question bank matches different question generation strategies according to different anchor point types. For ingredient formula anchor points, a progressive decomposition strategy is adopted, allowing users to repeatedly memorize the product formula from different angles. For spokesperson anchor points, an association confirmation strategy is adopted, generating questions that confirm the association between the person and the product. For attribute enumeration anchor points, a quantity induction strategy is adopted, counting the number of elements in the list and generating quantity induction questions.

4. The interactive advertising delivery system based on full-element cognitive verification and gamified dungeons according to claim 1, characterized in that, It also includes a team-based social and anti-cheating verification module. This module obtains a friend list or an existing team member list through a third-party platform, teams up in dungeon battles, conducts multi-dimensional behavioral correlation analysis, performs a comprehensive risk score, and classifies and handles cases according to the comprehensive risk score. The cross-platform consumption tracking and differentiated commission module is used for consumption and differentiated commissions on third-party platforms.

5. An interactive advertising delivery system based on full-element cognitive verification and gamified dungeons as described in claim 1, characterized in that, The advertiser points pool and difficulty-attempt control module includes a points recharge unit, a difficulty and reward configuration unit, a user level matching unit, and a points real-time deduction and automatic removal unit. The points recharge unit is used for advertisers to pre-charge through the recharge interface. The difficulty and reward configuration unit is used for advertisers to set configuration parameters for each advertisement. The user level matching unit displays puzzles of corresponding difficulty to different users based on their overall level. The points real-time deduction and automatic removal unit deducts points and updates the advertisement status based on the display status and remaining points.

6. The interactive advertising delivery system based on full-element cognitive verification and gamified dungeons according to claim 1, characterized in that, The location-based task distribution and O2O verification module includes a location collection and matching unit, a merchant self-service task creation unit, and an O2O rights verification unit. The location collection and matching unit is used to obtain the location of the user terminal and match the corresponding task. The merchant self-service task creation unit is used for merchants to submit task creation requests. The O2O rights verification unit is used to generate electronic verification vouchers after the user completes the task.

7. An interactive advertising delivery system based on full-element cognitive verification and gamified dungeons as described in claim 1, characterized in that, It also includes a user creative crowdfunding and advertising crowdfunding incentive module, which is used to collect users' advertising creative content, associate the creative content with the full-element anchor data of the corresponding product, and issue rewards to creators based on interaction data.

8. An interactive advertising delivery system based on full-element cognitive verification and gamified dungeons as described in claim 7, characterized in that, The user-generated content crowdfunding and advertising crowd-creation incentive module includes a creative submission unit, a two-tier review unit, a creative release and revenue linkage unit, and a periodic best ad selection unit. The creative submission unit is used to submit advertising creative content, including a creative type identifier, associated product ID, and creative text. The two-tier review unit is used to review the submitted creative content by both the merchant and the platform. The creative release and revenue linkage unit is used to release approved creatives as independent ads, generating cognitive verification questions based on the full-element anchor data of the corresponding product, and integrating the advertiser's points pool budget allocation. Points rewards are also distributed to creators based on user interaction data. The periodic best ad selection unit periodically calculates a comprehensive score for each creative content by weighting its multi-dimensional performance data, generates a creative content ranking, and distributes points rewards to top-ranked creators.

9. An interactive advertising delivery method based on full-element cognitive verification and gamified instances, characterized in that, The interactive advertising delivery system according to any one of claims 1-6 includes the following steps: Merchants publish advertising tasks through a location-based task distribution and O2O verification module, allowing users to discover nearby advertising tasks. Users complete in-depth quizzes using the fully-element cognitive verification question bank automatic generation module, or participate in memory puzzle challenges using the memory puzzle challenge module, earning corresponding points. The advertiser points pool and difficulty-attempt control module deduct points from the advertiser points pool in real time and check whether automatic removal is necessary. Simultaneously, cognitive scores and puzzle performance are updated to the user character development and equipment synthesis modules, improving corresponding attribute values. Users, carrying character attributes and historical cognitive scores, participate in advertiser dungeon challenges through the dungeon skill binding and combat engine modules. The system invites friends to participate through team-based social interaction and anti-cheating verification modules. The dungeon combat engine comprehensively calculates the damage of each member, and BOSS skills are forcibly bound to advertising anchor timestamps. Throughout the dungeon battle, the team-based social interaction and anti-cheating verification modules conduct multi-dimensional behavioral correlation analysis on members imported from third parties within the same team, and perform comprehensive risk scoring based on the analysis results. Based on the comprehensive risk score, tiered actions are taken. After clearing the dungeon, electronic verification vouchers are issued through the location-based task distribution and O2O verification modules. These electronic vouchers can be used for verification at designated stores, or for obtaining differentiated rebates by making purchases on third-party shopping platforms through the cross-platform consumption tracking and differentiated rebate modules.

10. The interactive advertising delivery method based on full-element cognitive verification and gamified dungeons according to claim 9, characterized in that, The interactive advertising delivery system based on any one of claims 7-8 further includes the following steps: creators improve advertising creatives through user creative crowdfunding and advertising crowdfunding incentive modules, and after being reviewed and adopted, the creatives are published and automatically connected to the full-element cognitive verification question bank automatic generation module and the points pool system. The system continuously tracks user interaction data of the advertising creatives and issues points rewards to creators based on user interaction data. In each period, the system calculates the total score of each advertising creative based on multi-dimensional performance data, generates a creative advertising ranking list, and issues points rewards to the top-ranked creators.