EA Allowing Players To Customise In-Game Character Animations
"Working at Electronic Arts" | Source: Glassdoor
Electronic Arts was recently awarded a patent proposing a system that allows players to customise the animations of in-game characters using a deep learning framework.
The system includes an animation-editing interface within the video game, enabling players to modify individual poses or frames of an animation.
Modifications made by the player to one pose are extrapolated across the entire animation using a motion generation machine learning model, resulting in smooth and realistic animations.
The system can handle various types of animations, including locomotion and asynchronous motions, and can update the animation rule set to incorporate player modifications.
The patent also introduces the concept of monitoring gameplay to identify specific video game states triggering the application of the modified animation rule set, allowing players to customise their character’s behaviour in the virtual world.
Earlier today, we came across a recently published patent from Electronic Arts titled, “SYSTEM FOR CUSTOMIZING IN-GAME CHARACTER ANIMATIONS BY PLAYERS,” filed in December 2021 under the name of Electronic Arts Inc. The patent, published earlier this month, describes a system for allowing players of a video game to customise the animations of in-game characters. The system uses a deep learning framework to enable users to modify the poses and movements of characters in the video game.
“System and methods for using a deep learning framework to customize animation of an in-game character of a video game. The system can be preconfigured with animation rule sets corresponding to various animations. Each animation can be comprised of a series of distinct poses that collectively form the particular animation,” reads the abstract for the patent.
“The system can provide an animation-editing interface that enables a user of the video game to make modifications to at least one pose or frame of the animation. The system can realistically extrapolate these modifications across some or all portions of the animation. In addition or alternatively, the system can realistically extrapolate the modifications across other types of animations.”
In traditional video game development, creating realistic character animations requires a lot of time and effort from video game developers. This patent proposes a solution to simplify this process by allowing players to customise the animations themselves. The system uses a deep learning framework to enable users to modify the poses and movements of characters in the video game.
The system works by providing an animation-editing interface within the video game. Players can make modifications to individual poses or frames of an animation. For example, they can change the position, angle, or velocity of a character’s limbs or body parts. To ensure smooth and realistic animations, the system uses a motion generation machine learning model to extrapolate the modifications made by the player across the entire animation.
This means that the changes made to one pose will be realistically applied to all other poses in the animation. The system can also update the animation rule set, which defines the sequence of poses that make up an animation, to incorporate the modifications made by the player. This way, the customised animations will be used in the video game when the character performs specific actions.
The patent also mentions that the system can handle different types of animations, such as locomotion like walking, running, and jumping, and asynchronous motions like punching, kicking, and bending. It also suggests the possibility of resolving any motion artefacts that may occur due to the modifications made by the player.
According to the claims in the patent, the system involves creating altered poses for a second action by modifying the initial pose of a character. This means that players can adjust the poses of characters for one action, such as running, based on user input. However, it doesn’t end there!
Furthermore, the system can utilise the modifications made to the running action poses as a foundation for generating altered poses for another action, like jumping. These modified poses are generated using a machine learning model. Lastly, the animation rule set for the second action (jumping) is updated with these modified poses, replacing the original ones.
Additionally, the patent introduces the concept of monitoring gameplay and identifying specific video game states that trigger the application of the modified animation rule set. This involves providing an animation editor interface to the player, allowing them to select and modify character poses for customisation.
With this patent from Electronic Arts, video game developers gain the ability to offer players an immersive experience that goes beyond visual customisation and extends to the very movements of their in-game characters. Players can now take greater control over locomotion, asynchronous actions, and various other animations that define their character’s behaviour in the virtual world.
Nevertheless, it is crucial to acknowledge that this currently exists as a patent and does not ensure actual implementation. Nonetheless, Electronic Arts’ proposed system stands as a testament to the innovative drive propelling advancements in the video industry. The future holds thrilling prospects for players worldwide, allowing them to genuinely leave their impact on the virtual realms they engage with.
What do you think about this? Do tell us your opinions in the comments below!