Neural Animation Layering for Blending Unaligned Motion Sources
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing animation generation systems face challenges in creating realistic and efficient character movements in electronic games, particularly in mixing, blending, and editing unaligned motion sources, leading to unrealistic poses and complex manual corrections.
Innovation Solution
A dynamic animation generation system using a deep learning framework with a modular architecture that includes a motion generator and control modules to synthesize novel movements from unstructured motion capture data, applying layering techniques like additive, override, and blend layering to generate realistic character animations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hand-tuned techniques are used to adjust skeleton positions for realistic movement, then movement realism is improved, but development complexity and time consumption increase substantially
Solution Approach 1:
The patent replaces the manual mechanical adjustment process with an automated neural network system. The motion generator network automatically adjusts skeleton positions and generates realistic movements by learning from motion capture data, eliminating the need for developers to manually tune each skeleton position while maintaining movement realism.
Solution Approach 2:
The system enables self-service by allowing the neural network to autonomously generate and optimize character movements without continuous human intervention. The network learns from training data and automatically produces realistic animations, reducing developer workload while maintaining high movement quality.
2Adaptability or versatility
If multiple motion capture data sources are combined to create diverse movements, then movement variety is improved, but alignment difficulties and unrealistic poses increase
Solution Approach 1:
The patent introduces a gating network as an intermediary that selectively combines motion capture data from multiple sources. The gating network determines which data sources to blend and in what proportions, ensuring smooth transitions and realistic poses while maintaining movement variety. This intermediary layer resolves conflicts between different motion sources and prevents unrealistic poses.
3Reliability
If more detailed character models are created to increase realism, then visual quality is improved, but processing burden on developers increases
Solution Approach 1:
The patent replaces manual model creation and adjustment processes with automated neural network generation. The system automatically creates detailed character models and animates them using learned motion patterns, reducing the time and effort required for developers to create high-quality visual content while maintaining detailed model realism.
Data Source
AI summary
In some embodiments, the dynamic animation generation system can provide a deep learning framework to produce a large variety of martial arts movements in a controllable manner from unstructured motion capture data. The system can imitate animation layering using neural networks with the aim to overcome challenges when mixing, blending and editing movements from unaligned motion sources. The system can synthesize movements from given reference motions and simple user controls, and generate unseen sequences of locomotion, but also reconstruct signature motions of different fighters. For achieving this task, the dynamic animation generation system can adopt a modular framework that is composed of the motion generator, that maps the trajectories of a number of key joints and root trajectory to the full body motion, and a set of different control modules that map the user inputs to such trajectories.