Network game anti-cheating method and device
By setting up a reporting mechanism in online games and analyzing game record data, identifying cheat scripts in combination with the target network model, and conducting manual review, the problem of identifying and preventing cheating behaviors in online games is solved, and the purification of the game environment and the healthy development of the industry is achieved.
Patent Information
- Application Number
- CN202411775900.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-05
- Publication Date
- 2025-04-11
AI Technical Summary
There are cheating behaviors in online games, which affect the game experience and the healthy development of the industry. It is difficult for existing technologies to effectively identify and prevent cheating.
By setting up a reporting mechanism and analyzing game record data, abnormal player behavior is identified; using the target network model to identify cheat scripts in the game screen, and using manual review to ensure accuracy.
Effectively identify and prevent cheating, improve the degree of purification of the game environment, reduce the probability of mis-blocking and misblocking, and promote the sustainable and healthy development of the game industry.
Smart Images
Figure CN119792948A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of script detection, and in particular to an anti-cheating method and device for online games. Background Art
[0002] With the development of the Internet, online games have become an increasingly important part of economic activities. However, in the development of the game industry, cheating has become a difficult problem to manage.
[0003] Common cheating methods are usually to use some third-party software or programs to modify game data or game loopholes to seek benefits for players, thereby enhancing the skills or abilities of game characters and surpassing the normal game rules and settings. Such software or programs can greatly enhance the abilities of game characters, such as by modifying game data, using packet and packet capture tools to submit false parameters to the game itself or the game server, thereby changing the abilities of characters in the game, so as to easily obtain the benefits of victory, rewards and pleasure.
[0004] This cheating behavior seriously affects the gaming experience of the majority of players who abide by the rules, and inevitably has an impact on the healthy development of the online gaming industry.
[0005] In order to solve the above problems, the present invention proposes an anti-cheating method and device for online games. Summary of the invention
[0006] In order to make up for the defects of the prior art, the present invention provides a simple and efficient online game anti-cheating method and device.
[0007] The present invention is achieved through the following technical solutions:
[0008] An anti-cheating method for online games comprises the following steps:
[0009] Step S1: Set up a reporting mechanism
[0010] When the number of reports of a game player reaches a custom threshold, the game record data of the corresponding game player is obtained from the game terminal database;
[0011] Step S2: Analyze the game player's game record data
[0012] Compare the average win rate of the game player during the period of being reported with the average win rate before being reported. If the difference ratio reaches the custom threshold, it is considered suspected of cheating;
[0013] Alternatively, the gamer's game record in a single game can be analyzed. If the improvement ratio of the gamer's game record in a single game before and during the reporting period exceeds a custom threshold, the player is considered to be suspected of cheating.
[0014] Marking gamers suspected of cheating as abnormal types;
[0015] Step S3: Scan the game folders of game players marked as abnormal types
[0016] When gamers download games, they are given permission to scan their game folders to detect cheating behavior.
[0017] If the scan finds cheat script files in the game folder, or cheat scripts in the game screen, the corresponding game player is judged to have cheated, the penalty mechanism is activated, and the player is punished accordingly according to the custom penalty rules;
[0018] If no abnormal files are detected in the game folder, the game screen image of the marked game player is obtained;
[0019] Step S4: Identify the game screen image of the marked game player based on the trained target network model
[0020] Customize batch selection of game screen images with cheating behavior from the game terminal database, customize the target objects in the game screen images, and mark them; generate a training sample set using the marked game screen images, and perform feature learning training on the target network model based on the training sample set;
[0021] Based on the trained target network model, identify whether the game screen image of the labeled game player belongs to the script image:
[0022] If the probability of a script image reaches a custom probability threshold, the corresponding game player is judged to have cheated, and the penalty mechanism is activated to impose corresponding penalties according to the custom penalty rules;
[0023] If the probability of being a script image is lower than the custom probability threshold, the image to be tested will be marked as a suspected plug-in image, and the image to be tested will be pushed to the manual review window for further manual review to determine whether the game player has cheated.
[0024] In step S4, the feature points of each game screen are marked in all acquired game screen image data, that is, the real position of the target object and the external representation type of the target object are marked, so that each marked game screen becomes a training sample.
[0025] The external representation types include frames, rays, auxiliary frames, function selection item frames, distances, external icons, and external scenes.
[0026] In step S4, the real position annotation information of the target object and the external representation type annotation information of the target object are displayed on the game screen, and the relationship between the target object and the real position annotation information and the external representation type annotation information, as well as the customized specific information content are stored in a data table.
[0027] An anti-cheating device for online games, comprising a reporting mechanism management module, a game data reading module, a game data analysis module, a marking module, a game folder scanning module, a target network model recognition module, a manual review module and a cheating punishment module;
[0028] The reporting mechanism management module is responsible for collecting reporting data in real time and conducting statistical analysis on the reporting data. When the number of reports of game players reaches a custom threshold, the game data reading module is triggered;
[0029] The game data reading module is responsible for obtaining the game record data of the corresponding game player and the game screen images of the marked game player from the game terminal database, and for customizing batch selection of game screen images with cheating behavior;
[0030] The game data analysis module is responsible for comparing the average win rate of game players during the period of being reported with the average win rate before being reported. If the difference ratio reaches a custom threshold, it is considered suspected of cheating;
[0031] Alternatively, the gamer's game record in a single game can be analyzed. If the improvement ratio of the gamer's game record in a single game before and during the reporting period exceeds a custom threshold, the player is considered to be suspected of cheating.
[0032] The labeling module is responsible for labeling game players suspected of cheating as abnormal types, customizing the target objects in the game screen images with cheating behaviors selected in batches, and labeling them;
[0033] The game folder scanning module is responsible for scanning the game folders of game players marked as abnormal types; when game players download games, they are provided with scanning permissions for their game folders to detect cheating behaviors;
[0034] If the scan finds cheat script files in the game folder, or cheat scripts in the game screen image, the corresponding game player is judged to have cheated, triggering the cheating penalty module;
[0035] If no abnormal files are detected in the game folder, the game data reading module is triggered to obtain the game screen image of the marked game player;
[0036] The target network model recognition module is responsible for generating a training sample set using the annotated game screen images, and performing feature learning training on the target network model based on the training sample set; and identifying the annotated game player's game screen images based on the trained target network model to determine whether they belong to script images:
[0037] If the probability of a script image reaches a custom probability threshold, the corresponding game player is judged to have cheated, and the penalty mechanism is activated to impose corresponding penalties according to the custom penalty rules;
[0038] If the probability of the image being a script is lower than the custom probability threshold, the image to be tested is marked as a suspected plug-in image and pushed to the manual review module;
[0039] The manual review module is responsible for further manual review to determine whether the game player has cheated;
[0040] The cheating penalty module is responsible for initiating the penalty mechanism and imposing corresponding penalties according to the custom penalty rules.
[0041] The labeling module marks the feature points of each game screen in all acquired game screen image data, that is, marks the real position of the target object and the external representation type of the target object, so that each marked game screen becomes a training sample.
[0042] The real position annotation information of the target object and the external representation type annotation information of the target object are displayed on the game screen, and the relationship between the target object and the real position annotation information and the external representation type annotation information, as well as the customized specific information content, are stored through a data table.
[0043] The external representation types include frames, rays, auxiliary frames, function selection item frames, distances, external icons, and external scenes.
[0044] An anti-cheating device for online games, characterized in that it comprises a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the above method steps when executing the computer program.
[0045] A readable storage medium, characterized in that: a computer program is stored on the readable storage medium, and the computer program implements the above method steps when executed by a processor.
[0046] The beneficial effects of the present invention are as follows: the online game anti-cheating method and device learns and trains the target network model so that it has a receptive field of multiple granularities, thereby improving the recognition ability of the model, being able to effectively identify cheating scripts, improving the accuracy of plug-in image recognition, reducing the probability of false blocking and wrong blocking, purifying the game environment, and being conducive to the sustainable and healthy development of the game industry. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0048] Attached Figure 1 The figure is a schematic diagram of the network game anti-cheating method of the present invention. DETAILED DESCRIPTION
[0049] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0050] There are many types of online games, mainly RPG, ACT, AAG, SLG, SRPG, RTS, FTG, STG, FPS, TPS, PZL, SPG, RCG (RAC), CAG, TAB, MSC, WAG, MUD, MOBA. This example mainly checks cheating in MOBA games. MOBA games refer to multiplayer online battle arena (MOBA), also known as action real-time strategy (Action RTS, ARTS). It is a subcategory of video games (Video Game) - real-time strategy games (RTS). In this type of game, equipment is generally required to be purchased in battle. Players are usually divided into two teams, and the two teams compete with each other in scattered game maps. Each player controls the selected character through an RTS-style interface. However, unlike traditional hardcore RTS games, this type of game usually does not have organizational units such as buildings, resources, and training arms that are common in RTS games, and players only control the characters they choose. Cheating in MOBA games involves using scripts to know the release positions of enemy skills in advance, thereby avoiding the enemy's skills, knowing the enemy's movement positions in advance, and improving the hit rate of one's own skills; or attaching the mouse to the enemy's character model.
[0051] Machine learning is a multi-disciplinary subject that involves probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures, thereby continuously improving their own performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence. The core of machine learning is to allow computers to learn patterns from data and make predictions or decisions on new data without explicit programming by humans.
[0052] Machine learning can be divided into types such as supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning.
[0053] Among them, supervised learning is to obtain an optimal model through training of existing training data sets, and then use this model to map all inputs to corresponding outputs, and make simple judgments on the outputs to achieve the purpose of prediction and classification.
[0054] Unsupervised learning involves discovering structures or associations in data through analysis of input data without known outputs or labels.
[0055] Semi-supervised learning combines the characteristics of supervised learning and unsupervised learning, with some data being labeled and some being unlabeled.
[0056] Reinforcement learning emphasizes how to act based on the environment to maximize expected benefits.
[0057] Simply put, machine learning is about allowing computers to learn from data, optimize themselves, and improve performance, so that they can make more accurate predictions and decisions. Machine learning plays a vital role in this instance, effectively analyzing the player's game screen and then determining whether the player is cheating.
[0058] The online game anti-cheating method comprises the following steps:
[0059] Step S1: Set up a reporting mechanism
[0060] When the number of reports of a game player reaches a custom threshold, the game record data of the corresponding game player is obtained from the game terminal database;
[0061] Step S2: Analyze the game player's game record data
[0062] Compare the average win rate of the game player during the period of being reported with the average win rate before being reported. If the difference ratio reaches the custom threshold, it is considered suspected of cheating;
[0063] Alternatively, the gamer's game record in a single game can be analyzed. If the improvement ratio of the gamer's game record in a single game before and during the reporting period exceeds a custom threshold, the player is considered to be suspected of cheating.
[0064] Marking gamers suspected of cheating as abnormal types;
[0065] Step S3: Scan the game folders of game players marked as abnormal types
[0066] When gamers download games, they are given permission to scan their game folders to detect cheating behavior.
[0067] If the scan finds cheat script files in the game folder or cheat scripts in the game screen, the corresponding game player is judged to have cheated, and the penalty mechanism is activated to impose corresponding penalties according to the custom penalty rules (the account will be banned for different lengths according to the number of cheating games and the impact of cheating);
[0068] If no abnormal files are detected in the game folder, the game screen image of the marked game player is obtained;
[0069] Step S4: Identify the game screen image of the marked game player based on the trained target network model
[0070] Customize batch selection of game screen images with cheating behavior from the game terminal database, customize the target objects in the game screen images, and mark them; generate a training sample set using the marked game screen images, and perform feature learning training on the target network model based on the training sample set;
[0071] Based on the trained target network model, identify whether the game screen image of the labeled game player belongs to the script image:
[0072] If the probability of a script image reaches a custom probability threshold, the corresponding game player is judged to have cheated, and the penalty mechanism is activated to impose corresponding penalties according to the custom penalty rules;
[0073] If the probability of being a script image is lower than the custom probability threshold, the image to be tested will be marked as a suspected plug-in image, and the image to be tested will be pushed to the manual review window for further manual review to determine whether the game player has cheated.
[0074] During the manual review process, professionals will check suspected cheating images to determine whether there is indeed cheating behavior. If the suspected cheating image is confirmed to be a cheating image by manual review, the system will impose corresponding penalties on it.
[0075] In step S4, the feature points of each game screen are marked in all acquired game screen image data, that is, the real position of the target object and the external representation type of the target object (that is, the cheating representation type) are marked, so that each marked game screen becomes a training sample.
[0076] The external representation types include frames, rays, auxiliary frames, function selection item frames, distances, external icons, and external scenes.
[0077] For example, in a first-person shooter game, the game logic records the coordinates of all players in battle, and marks the enemy with a box on the screen after reading the specified coordinates. This allows the enemy's position to be predicted and can be developed into an automatic aiming aid. Distance is a form of perspective rendering. In MOBA games, it is manifested as the adsorption of the mouse pointer. In order to improve the player's skill hit rate, general cheating scripts will adsorb the mouse pointer on the enemy's character model, thereby greatly improving the player's skill hit rate, or through a series of red boxes to predict the enemy's skill release position, so as to make advance responses to avoid the enemy's non-directional skills. Non-directional skills refer to optional directional skills that can be released without locking the enemy's character model, because the enemy's character model cannot be locked. Therefore, the hit of this type of skill requires the player to predict the movement of the enemy's character model. The cheating script can help the player adsorb the mouse on the enemy's character model, greatly improving the hit rate of non-directional skills.
[0078] In step S4, the real position annotation information of the target object and the external representation type annotation information of the target object are displayed on the game screen, and the relationship between the target object and the real position annotation information and the external representation type annotation information, as well as the customized specific information content are stored in a data table.
[0079] For example, a training sample can only contain one target object. This means that a training sample is an image that contains only one specific target object. The actual location of this target object on the image may be different in different training samples.
[0080] On the other hand, it is also possible that a training sample contains multiple target objects. These target objects can belong to the same or different plug-in performance categories. For example, a training sample may contain target objects of different plug-in categories such as box plug-in, ray plug-in, and distance plug-in. In this case, the true positions of target objects of different plug-in categories in different training samples may be different.
[0081] For example, if a training sample is an image containing multiple target objects, and these target objects represent positive samples of cheating performance, then in this training sample, multiple target objects of cheating performance will appear at the same time.
[0082] The network game anti-cheating device includes a reporting mechanism management module, a game data reading module, a game data analysis module, a marking module, a game folder scanning module, a target network model recognition module, a manual review module and a cheating punishment module;
[0083] The reporting mechanism management module is responsible for collecting reporting data in real time and conducting statistical analysis on the reporting data. When the number of reports of game players reaches a custom threshold, the game data reading module is triggered;
[0084] The game data reading module is responsible for obtaining the game record data of the corresponding game player and the game screen images of the marked game player from the game terminal database, and for customizing batch selection of game screen images with cheating behavior;
[0085] The game data analysis module is responsible for comparing the average win rate of game players during the period of being reported with the average win rate before being reported. If the difference ratio reaches a custom threshold, it is considered suspected of cheating;
[0086] Alternatively, the gamer's game record in a single game can be analyzed. If the improvement ratio of the gamer's game record in a single game before and during the reporting period exceeds a custom threshold, the player is considered to be suspected of cheating.
[0087] The labeling module is responsible for labeling game players suspected of cheating as abnormal types, customizing the target objects in the game screen images with cheating behaviors selected in batches, and labeling them;
[0088] The game folder scanning module is responsible for scanning the game folders of game players marked as abnormal types; when game players download games, they are provided with scanning permissions for their game folders to detect cheating behaviors;
[0089] If the scan finds cheat script files in the game folder, or cheat scripts in the game screen image, the corresponding game player is judged to have cheated, triggering the cheating penalty module;
[0090] If no abnormal files are detected in the game folder, the game data reading module is triggered to obtain the game screen image of the marked game player;
[0091] The target network model recognition module is responsible for generating a training sample set using the annotated game screen images, and performing feature learning training on the target network model based on the training sample set; and identifying the annotated game player's game screen images based on the trained target network model to determine whether they belong to script images:
[0092] If the probability of a script image reaches a custom probability threshold, the corresponding game player is judged to have cheated, and the penalty mechanism is activated to impose corresponding penalties according to the custom penalty rules;
[0093] If the probability of the image being a script is lower than the custom probability threshold, the image to be tested is marked as a suspected plug-in image and pushed to the manual review module;
[0094] The manual review module is responsible for further manual review to determine whether the game player has cheated;
[0095] The cheating penalty module is responsible for initiating the penalty mechanism and imposing corresponding penalties according to the custom penalty rules (the account will be banned for different lengths according to the number of games cheated and the impact of cheating).
[0096] The labeling module marks the feature points of each game screen in all acquired game screen image data, that is, marks the real position of the target object and the external representation type of the target object (that is, the cheating representation type), so that each marked game screen becomes a training sample.
[0097] The real position annotation information of the target object and the external representation type annotation information of the target object are displayed on the game screen, and the relationship between the target object and the real position annotation information and the external representation type annotation information, as well as the customized specific information content, are stored through a data table.
[0098] The external representation types include frames, rays, auxiliary frames, function selection item frames, distances, external icons, and external scenes.
[0099] The network game anti-cheating device comprises a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the above method steps when executing the computer program.
[0100] The readable storage medium stores a computer program, and when the computer program is executed by a processor, the above method steps are implemented.
[0101] The embodiment described above is only one specific implementation of the present invention. Common changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included in the protection scope of the present invention.
Claims
1. An anti-cheating method for online games, characterized in that: The following steps are involved: Step S1: Set up a reporting mechanism When the number of reports of a game player reaches a custom threshold, the game record data of the corresponding game player is obtained from the game terminal database; Step S2: Analyze the game player's game record data Compare the average win rate of the game player during the period of being reported with the average win rate before being reported. If the difference ratio reaches the custom threshold, it is considered suspected of cheating; Alternatively, the gamer's game record in a single game can be analyzed. If the improvement ratio of the gamer's game record in a single game before and during the reporting period exceeds a custom threshold, the player is considered to be suspected of cheating. Marking gamers suspected of cheating as abnormal types; Step S3: Scan the game folders of game players marked as abnormal types When gamers download games, they are given permission to scan their game folders to detect cheating behavior. If the scan finds cheat script files in the game folder, or cheat scripts in the game screen, the corresponding game player is judged to have cheated, the penalty mechanism is activated, and the player is punished accordingly according to the custom penalty rules; If no abnormal files are detected in the game folder, the game screen image of the marked game player is obtained; Step S4: Identify the game screen image of the marked game player based on the trained target network model Customize batch selection of game screen images with cheating behavior from the game terminal database, customize target objects in the game screen images, and mark them; Generate a training sample set using the annotated game screen images, and perform feature learning training on the target network model based on the training sample set; Based on the trained target network model, identify whether the game screen image of the labeled game player belongs to the script image: If the probability of a script image reaches a custom probability threshold, the corresponding game player is judged to have cheated, and the penalty mechanism is activated to impose corresponding penalties according to the custom penalty rules; If the probability of being a script image is lower than the custom probability threshold, the image to be tested will be marked as a suspected plug-in image, and the image to be tested will be pushed to the manual review window for further manual review to determine whether the game player has cheated.
2. The online game anti-cheating method according to claim 1, characterized in that: In step S4, the feature points of each game screen are marked in all acquired game screen image data, that is, the real position of the target object and the external representation type of the target object are marked, so that each marked game screen becomes a training sample.
3. The online game anti-cheating method according to claim 2, characterized in that: In step S4, the real position annotation information of the target object and the external representation type annotation information of the target object are displayed on the game screen, and the relationship between the target object and the real position annotation information and the external representation type annotation information, as well as the customized specific information content are stored in a data table.
4. The online game anti-cheating method according to claim 2, characterized in that: The external representation types include frames, rays, auxiliary frames, function selection item frames, distances, external icons, and external scenes.
5. An anti-cheating device for online games, characterized in that: It includes reporting mechanism management module, game data reading module, game data analysis module, annotation module, game folder scanning module, target network model identification module, manual review module and cheating punishment module; The reporting mechanism management module is responsible for collecting reporting data in real time and conducting statistical analysis on the reporting data. When the number of reports of game players reaches a custom threshold, the game data reading module is triggered; The game data reading module is responsible for obtaining the game record data of the corresponding game player and the game screen images of the marked game player from the game terminal database, and for customizing batch selection of game screen images with cheating behavior; The game data analysis module is responsible for comparing the average win rate of game players during the period of being reported with the average win rate before being reported. If the difference ratio reaches a custom threshold, it is considered suspected of cheating; Alternatively, the gamer's game record in a single game can be analyzed. If the improvement ratio of the gamer's game record in a single game before and during the reporting period exceeds a custom threshold, the player is considered to be suspected of cheating. The labeling module is responsible for labeling game players suspected of cheating as abnormal types, customizing the target objects in the game screen images with cheating behaviors selected in batches, and labeling them; The game folder scanning module is responsible for scanning the game folders of game players marked as abnormal types; When gamers download games, they are given permission to scan their game folders to detect cheating behavior. If the scan finds cheat script files in the game folder, or cheat scripts in the game screen image, the corresponding game player is judged to have cheated, triggering the cheating penalty module; If no abnormal files are detected in the game folder, the game data reading module is triggered to obtain the game screen image of the marked game player; The target network model recognition module is responsible for generating a training sample set using the annotated game screen images and performing feature learning training on the target network model based on the training sample set; Based on the trained target network model, the game screen images of the labeled game players are identified to determine whether they belong to script images: If the probability of a script image reaches a custom probability threshold, the corresponding game player is judged to have cheated, and the penalty mechanism is activated to impose corresponding penalties according to the custom penalty rules; If the probability of the image being a script is lower than the custom probability threshold, the image to be tested is marked as a suspected plug-in image and pushed to the manual review module; The manual review module is responsible for further manual review to determine whether the game player has cheated; The cheating penalty module is responsible for initiating the penalty mechanism and imposing corresponding penalties according to the custom penalty rules.
6. The anti-cheating device for online games according to claim 5, characterized in that: The labeling module marks the feature points of each game screen in all acquired game screen image data, that is, marks the real position of the target object and the external representation type of the target object, so that each marked game screen becomes a training sample.
7. The anti-cheating device for online games according to claim 6, characterized in that: The real position annotation information of the target object and the external representation type annotation information of the target object are displayed on the game screen, and the relationship between the target object and the real position annotation information and the external representation type annotation information, as well as the customized specific information content, are stored through a data table.
8. The anti-cheating device for online games according to claim 6, characterized in that: The external representation types include frames, rays, auxiliary frames, function selection item frames, distances, external icons, and external scenes.
9. An anti-cheating device for online games, characterized in that: The method comprises a memory and a processor; the memory is used to store a computer program, and the processor is used to implement the method according to any one of claims 1 to 4 when executing the computer program.
10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.