Server-Based AI Cheat Detection in Electronic Games
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Competitive electronic games face integrity issues due to users employing modified game applications or exploits, which conventional moderation struggles to detect, especially in sophisticated cheating scenarios.
Innovation Solution
A system utilizing machine vision and machine learning algorithms to monitor video information from client devices, identifying virtual objects and changes within the game environment, and comparing these against a ruleset to detect and report illegal or anomalous events, thereby automating the detection of cheats and ensuring fair play.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human moderators manually review gameplay to detect cheating, then detection accuracy for sophisticated cheats improves, but productivity and coverage of monitoring decrease
Solution Approach 1:
The system enables automated self-detection of cheating by having the game application automatically capture gameplay footage and transmit it to servers for AI analysis, eliminating the need for manual moderator intervention for routine detection while maintaining high accuracy
Solution Approach 2:
The patent replaces the mechanical system of human moderators with an automated computer vision system using machine learning models that analyze video frames to detect cheating behaviors, achieving both high detection accuracy and scalable productivity
2Productivity
If automated systems are used to monitor gameplay, then productivity and monitoring coverage improve, but detection precision for sophisticated cheats decreases
Solution Approach 1:
The system uses an intermediary layer of AI/ML models that act as a bridge between automated video analysis and cheating detection, enabling the system to process large volumes of gameplay data while maintaining high detection accuracy through sophisticated pattern recognition
Solution Approach 2:
The patent employs multiple AI/ML models with different detection parameters and thresholds to analyze various aspects of gameplay, allowing the system to maintain high detection accuracy across diverse cheating methods while processing large volumes of data
3Speed
If computing resources are dedicated to running the game application, then game performance improves, but resources available for cheat detection decrease
Solution Approach 1:
The system segments the cheat detection function into a separate server-based AI analysis component, allowing the client device to dedicate all computing resources to running the game application while the server handles the computationally intensive task of video frame analysis and cheating detection
Solution Approach 2:
The patent introduces an intermediary server system that receives video data from the game application and performs cheat detection analysis, allowing the game client to maintain optimal performance while the server provides reliable cheat detection capabilities
Data Source
AI summary
A method, at a server computer, for verifying the results of an electronic game includes obtaining a unique identifier from a client device running a game application. The server computer then receives, from the client device, video information in a real-time video stream. The video information is evaluated at the server computer using a machine learning model to identify a virtual object in a first frame of the video information and identify the virtual object in a second frame of the video information. The method also includes determining a change in the virtual object between the first frame and the second frame and comparing the change to one or more rules of a ruleset. The method includes determining a legality of the change in the virtual object based at least partially on the ruleset. If an illegal event is identified, the change is presented to a moderator for review.