3D Facial Expression Generation via Semantic Transfer Intensities
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for generating realistic 3D facial expressions using blend shapes lack accuracy and realism due to their reliance on linear and global transfer methods, resulting in unrealistic movements and expressions such as half-open eyes or facial creases.
Innovation Solution
The proposed solution involves detecting the semantic type of a blend shape to assign non-linear transfer intensities to corresponding points between the blend shape and the facial expression source, ensuring that different portions of the face contribute differently to the generated expression, thus preserving semantic integrity and realism.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If linear and global transfer methods are used to generate 3D facial expressions from blend shapes, then the generation process is simple and efficient, but the accuracy and realism of the expressions deteriorate
Solution Approach 1:
The patent divides the facial expression transfer process into local regions (e.g., eyes, mouth, cheeks) and applies different transfer intensities to each region based on its semantic type. This segmentation allows the system to preserve both efficiency through automated processing and accuracy through region-specific adjustments, resolving the contradiction between simple global transfer and precise local expression generation.
2Ease of operation
If conventional blend shape techniques are used, then the process is easy to operate, but unrealistic movements and expressions occur
Solution Approach 1:
The patent applies different transfer intensities to different local regions of the face based on their semantic types. For example, regions involved in smiling receive higher transfer intensity from the blend shape, while regions that should remain stable maintain higher fidelity to the source. This local quality approach ensures operational simplicity while achieving realistic expressions without unrealistic movements.
3Device complexity
If global transfer is applied uniformly to all facial points, then the transfer process is consistent and simple, but semantic integrity of different facial regions is lost
Solution Approach 1:
The patent dynamically adjusts transfer intensities based on the semantic type of each facial region and the specific expression being generated. Rather than using a fixed uniform transfer ratio, the system adaptively determines the appropriate blend between source and target for each region, preserving semantic integrity while maintaining manageable process complexity through automated dynamic adjustment.
Data Source
AI summary
A digital medium environment is described to generate a three dimensional facial expression from a blend shape and a facial expression source. A semantic type is detected that defines a facial expression of the blend shape. Transfer intensities are assigned based on the detected semantic type to the blend shape and the facial expression source, respectively, for individual portions of the three dimensional facial expression, the transfer intensities specifying weights given to the blend shape and the facial expression source, respectively, for the individual portions of the three dimensional facial expression. The three dimensional facial expression is generated from the blend shape and the facial expression source based on the assigned transfer intensities.