Facial Animation Rig Generation via Muscle Model Strain Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing animation systems face challenges in creating realistic and varied facial expressions for computer-generated characters, as they require tedious specification of numerous facial muscles and their movements, which is impractical to determine directly from a live actor, especially considering the variability in muscle attachment and strain ranges among individuals.
Innovation Solution
A computer-implemented method that uses high-resolution facial scans, a muscle model, and AI to generate a strain vector representing plausible facial expressions, allowing animators to create expressions by modifying strain values within constrained boundaries, thereby simplifying the animation process and enabling the simulation of unlimited facial expressions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If an animator specifies in detail the surface of the live actor's body, then the visual accuracy of the computer-animated character is improved, but the complexity and difficulty of the animation process increases significantly
Solution Approach 1:
The patent uses facial scanning technology to create a digital 3D model (copy) of the actor's facial surface geometry. This digital model captures the detailed surface topology without requiring manual specification, thereby maintaining visual accuracy while reducing animator workload. The scanned mesh serves as a reusable digital asset that can be deformed programmatically.
Solution Approach 2:
The patent replaces the manual mechanical process of specifying each facial muscle and surface point with an automated computational system. Machine learning models and algorithms automatically map facial expressions to mesh deformations, substituting the manual mechanical specification process with automated computational deformation techniques.
2Ease of operation
If an animator manually specifies numerous facial muscles and their movements, then the control over facial expressions is improved, but the time and effort required increases significantly
Solution Approach 1:
The system captures actual facial muscle movement patterns from the actor through scanning and motion capture, creating a digital copy of the muscle animation data. This copied muscle movement information is then applied to the digital twin, allowing realistic facial expressions to be generated without manually specifying each muscle movement, thus reducing time while maintaining control quality.
Solution Approach 2:
The patent substitutes manual muscle specification with automated machine learning models that predict muscle deformations from facial expressions. The system uses trained neural networks to automatically compute muscle movements based on expression targets, replacing the time-consuming manual muscle control process with efficient computational prediction.
3Adaptability or versatility
If the system captures scans of different facial expressions and blends them, then the ability to create varied expressions is improved, but the process becomes tedious and complex
Solution Approach 1:
The system creates a digital twin that copies the actor's complete facial geometry, texture, and muscle structure. This comprehensive digital copy enables versatile expression generation through programmatic deformation rather than manual blending of scanned expressions, maintaining expression variety while simplifying the creation process.
Solution Approach 2:
The patent replaces the manual process of capturing and blending facial expression scans with automated procedural generation using the digital twin. The system uses the copied muscle and surface data to programmatically generate novel expressions through controlled deformation, substituting the tedious blending process with efficient computational animation.
4Manufacturing precision
If the system accounts for variability in muscle attachment and strain ranges among individuals, then the realism of facial expressions is improved, but the complexity of determining accurate muscle models increases
Solution Approach 1:
The patent performs a comprehensive facial scan that copies the actor's unique muscle attachment points, insertion points, and strain ranges directly from their actual face. This empirical copying of individual anatomical variations creates an accurate personalized muscle model without requiring manual determination, thereby maintaining realism while reducing the complexity of model creation.
Data Source
AI summary
An animation system wherein scanned facial expressions are processed to form muscle models that can be used to generate expressions based on specification of a strain vector and a control vector of the muscle model.