Facial Layer Retargeting Without Spatial Correspondence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing animation systems face challenges in efficiently and compellingly transferring facial expressions between characters with different facial structures, as they require defining spatial correspondences, which can be cumbersome and limit the variety of characters that can be created with realistic emotional expressions.
Innovation Solution
A computer-implemented method that uses facial layers, including emotion, speech, and eye-blink layers, to retarget facial expressions by generating corresponding layers for a target character without defining spatial correspondences, employing weighted masks and simplex coefficients to modulate the influence of these layers over time, allowing for parameter-parallel retargeting between the source and target characters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If spatial correspondences are defined between source and target characters, then facial expressions can be transferred with accurate geometric alignment, but the process becomes cumbersome and limits the variety of characters that can be created
Solution Approach 1:
The facial expression transfer process is segmented into independent parametric components (emotion parameters, speech parameters, eye-blink parameters) rather than requiring complete spatial correspondence. Each parameter type can be transferred independently to different character types, simplifying the overall process while maintaining accuracy for each facial element.
Solution Approach 2:
The patent transforms the facial expression transfer problem from spatial coordinate mapping to parameter space transformation. By representing facial expressions as sets of parameters (emotion, speech, eye-blink) rather than spatial correspondences, the system enables flexible transfer across characters with arbitrary facial structures without requiring complex geometric alignments.
2Manufacturing precision
If spatial correspondences are defined between source and target characters, then facial expressions can be transferred with accurate geometric alignment, but the variety of characters that can be created is limited
Solution Approach 1:
The parametric facial expression representation serves as a universal interface that can be applied to any character type regardless of facial structure. The same set of parameters (emotion, speech, eye-blink) can control facial expressions across humans, animals, fantasy creatures, and mechanical characters, enabling the system to handle diverse character varieties while maintaining transfer accuracy.
3Manufacturing precision
If traditional facial expression transfer methods are used, then spatial alignment can be achieved, but the transfer process is inefficient and computationally intensive
Solution Approach 1:
The patent replaces the mechanical spatial correspondence system with a parametric representation system. Instead of computing complex spatial mappings between facial geometries, the system directly transfers parametric values (emotion, speech, eye-blink parameters) that inherently encode the desired facial expression, dramatically improving computational efficiency while maintaining alignment accuracy.
Data Source
AI summary
Techniques are disclosed for retargeting facial expressions. Input is received that represents a facial expression of a first character. Facial layers are generated based on the received input. The facial layers include one or more parameters extracted from the received input. A facial expression for a second character and corresponding to the facial expression of the first character is generated, based on the facial layers and without defining any spatial correspondence between the first character and the second character.