Motion Model Refinement via Contact Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating character motion from video sequences often produce unsatisfactory results with visually and physically implausible motions, such as feet floating or penetrating the ground, and unnatural body leans.
Innovation Solution
A motion model refinement system that determines contact joint points with a ground plane using a trained contact estimation model and performs physics-based optimization to generate a refined motion model, enforcing constraints like valid rigid mechanics and foot contact to ensure motion plausibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion estimation methods are used to extract character motions from video sequences, then motion data can be obtained for analysis and animation, but the extracted motions contain errors such as feet floating or penetrating the ground and implausible body lean
Solution Approach 1:
The system uses contact estimation models to continuously monitor and provide feedback on foot contact status with the ground plane. This feedback is integrated into the motion estimation process to correct implausible motions in real-time, ensuring feet remain on the ground while maintaining extraction efficiency
Solution Approach 2:
The patent replaces traditional mechanical motion extraction algorithms with a physics-based optimization approach. This substitution uses physical constraints (ground contact, rigid mechanics) to guide the motion estimation, eliminating floating feet and penetrating ground errors while maintaining computational efficiency
2Reliability
If contact constraints and physics-based optimization are applied to refine motion models, then motion plausibility is improved, but computational complexity increases
Solution Approach 1:
The optimization system is segmented into modular components: contact estimation model, ground plane determination module, and physics-based optimization module. Each component handles specific constraints independently, making the overall complex system more manageable and efficient to implement
Solution Approach 2:
The system dynamically adjusts optimization parameters based on the motion context and contact status. By changing parameters such as constraint weights and optimization thresholds adaptively, the system maintains high plausibility while reducing unnecessary computational complexity in simpler motion scenarios
Data Source
AI summary
In some embodiments, a model training system obtains a set of animation models. For each of the animation models, the model training system renders the animation model to generate a sequence of video frames containing a character using a set of rendering parameters and extracts joint points of the character from each frame of the sequence of video frames. The model training system further determines, for each frame of the sequence of video frames, whether a subset of the joint points are in contact with a ground plane in a three-dimensional space and generates contact labels for the subset of the joint points. The model training system trains a contact estimation model using training data containing the joint points extracted from the sequences of video frames and the generated contact labels. The contact estimation model can be used to refine a motion model for a character.