Virtual Character Animation Control Using Joint Torque Transitions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing animation production methods, particularly for animated characters in games, are labor-intensive and time-consuming due to frame-by-frame manual drawing, and existing machine learning techniques require extensive data and time to learn complex animations.
Innovation Solution
An animation implementation method using a trained control policy network to generate animations for virtual characters by adjusting joint torques based on initial and target posture data, leveraging reinforcement learning and physics-based simulation to automate the process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If frame-by-frame processing is used to ensure animation quality, then animation rendering quality is improved, but processing time increases significantly
Solution Approach 1:
The patent segments the animation processing into two distinct modes: frame-by-frame processing for key frames requiring high quality, and incremental processing for intermediate frames that can tolerate lower quality. This segmentation allows the system to apply different processing strategies to different parts of the animation sequence, optimizing both quality and time consumption.
Solution Approach 2:
The patent applies partial action by processing only the necessary portions of animation frames at full resolution. Key frames receive complete frame-by-frame processing, while intermediate frames use incremental processing with reduced detail, achieving acceptable overall quality with significantly reduced processing time.
2Speed
If frame interpolation technology is used to improve animation smoothness, then animation fluidity is improved, but processing complexity increases
Solution Approach 1:
The patent segments interpolation processing into two stages: first generating a coarse intermediate frame using low-resolution processing, then refining specific regions requiring higher detail. This two-stage approach reduces the overall computational complexity compared to full high-resolution interpolation while maintaining animation smoothness.
Solution Approach 2:
The patent applies partial action by performing complete interpolation only on key regions that require smooth transitions, while using simplified interpolation methods for less critical areas. This reduces processing complexity while maintaining the perceived animation fluidity.
3Productivity
If resource allocation is increased to improve processing speed, then processing efficiency is improved, but resource consumption increases
Solution Approach 1:
The patent implements dynamic resource allocation where the processing mode (frame-by-frame or incremental) is automatically selected based on the specific requirements of each frame and the current system state. This dynamic approach optimizes processing efficiency by allocating resources only where and when they are most needed, rather than uniformly across all frames.
Solution Approach 2:
The patent changes processing parameters such as resolution level, processing mode, and detail threshold based on the importance and position of each frame in the animation sequence. This allows efficient processing of less critical frames with reduced parameters while maintaining high quality for key frames, optimizing the balance between processing efficiency and resource consumption.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
An animation implementation method, comprising: obtaining a first target animation clip, the first target animation clip comprising a first key frame, and the first key frame comprising initial posture data of a target virtual character; inputting the initial posture data and a set target task into a trained control policy network to obtain target posture data of the target virtual character; obtaining, according to the initial posture data and the target posture data of the target virtual character, torques for adjusting a first preset number of joints of the target virtual character; adjusting the target virtual character from an initial posture to a target posture by using the torques of the first preset number of joints, so as to obtain a second key frame; and acquiring a second target animation clip, the second target animation clip consisting of at least the first key frame and the second key frame.