Motion Capture Marker System for Dynamic Surface Deformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion capture techniques fail to accurately capture and animate dynamic surface deformations such as jiggling flesh and bulging muscles, as they rely on simplified skeletal motion and lack detailed skin dynamics, leading to unrealistic computer-generated imagery.
Innovation Solution
A method using a large set of markers placed on bony and fleshy body parts to simultaneously capture rigid body motion and deformations, with local models and radial basis functions to estimate occluded marker positions and blend marker motions, allowing for more accurate rendering of skin deformations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional motion capture techniques using 40-60 markers are used, then the system complexity is reduced and ease of operation is improved, but the measurement precision of surface deformations deteriorates
Solution Approach 1:
The patent segments the motion capture system into two distinct marker sets: a sparse set of 40-60 markers for capturing rigid skeletal motion, and a dense set of markers for capturing detailed surface deformations. This segmentation allows each subsystem to specialize in specific aspects of motion, achieving high measurement precision for surface deformations while maintaining manageable system complexity through modular design.
2Measurement precision
If a dense marker set is used to capture detailed surface deformations, then the measurement precision of skin motion is improved, but the difficulty of detecting and measuring increases due to marker occlusion
Solution Approach 1:
The patent implements dynamic hole-filling techniques that adapt to the subject's motion state. The system dynamically identifies occluded markers and uses temporal-spatial interpolation based on neighboring markers and biomechanical models to estimate their positions. This dynamic approach maintains high measurement precision for skin motion while automatically handling the difficulty of marker detection and measurement during occlusion events.
3Manufacturing precision
If basic skinning techniques are used, then the ease of manufacture of the animation system is improved, but the manufacturing precision of skin deformation deteriorates due to the candy wrapper effect
Solution Approach 1:
The patent transforms the skinning problem from a geometric transformation task into a parameter-driven deformation task. By representing skin deformation as a combination of rigid body motion parameters and local deformation parameters, the system achieves high manufacturing precision for skin deformation while maintaining ease of manufacture through parameter-based control rather than complex geometric calculations.
4Manufacturing precision
If skeleton subspace deformation techniques are used, then the ease of operation is improved, but the measurement precision of volume preservation deteriorates
Solution Approach 1:
The patent introduces an intermediary dense marker set that acts as a mediator between the skeletal motion capture system and the final skin deformation rendering. This intermediary layer provides direct measurement data for volume preservation, enabling high manufacturing precision while keeping the overall system easy to operate by maintaining the simplicity of skeleton-based motion capture for the primary motion control.
Data Source
AI summary
A method for a computer system includes receiving global positional data associated with a set of markers from a plurality of markers associated with a surface of an object at one or more time instances, wherein global positional data associated with a first marker from the plurality of markers is absent from a first time instance, using local statistical methods to determine global positional data associated with the first marker at the first time instance in response to the global positional data associated with the set of markers at the one or more time instances, and determining a model of the object in response to the global positional data associated with the set of markers and the global positional data associated with the first marker.