Digital Human Facial Expression Transfer Using Reference Model Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The process of binding facial expressions to digital human figures is labor-intensive and time-consuming, typically requiring professional designers and taking weeks to complete, especially for high-quality, ultra-realistic digital humans.
Innovation Solution
A method and apparatus for automatically transferring facial expressions by screening a target reference model from a library, acquiring its expression library, and transferring the last frame of the expression to an object model, utilizing feature curve matching and registration techniques to ensure accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual binding by professional designers is used, then expression binding quality can be ensured, but time consumption increases significantly (1-2 weeks or more)
Solution Approach 1:
The patent creates a reference model library containing pre-established expression binding data from professional designs. These reference models serve as templates that can be copied and applied to multiple object models, eliminating the need for repeated manual binding work while preserving the quality standards established by professional designers.
Solution Approach 2:
The patent performs expression binding work in advance by creating a comprehensive reference model library with pre-bound expressions. This preliminary action allows the system to quickly match and apply appropriate reference models to object models later, significantly reducing the time required for expression binding in actual applications.
2Manufacturing precision
If manual binding by professional designers is used, then expression binding can be completed accurately, but labor costs increase
Solution Approach 1:
The patent establishes a reference model library with pre-computed expression binding data created by professional designers. This library serves as a reusable resource that can be automatically matched and applied to multiple object models, eliminating the need for continuous manual intervention and reducing labor costs while maintaining binding accuracy.
Solution Approach 2:
The patent implements an automatic matching mechanism that enables object models to self-assign appropriate reference models from the library based on similarity metrics. This self-service capability eliminates the need for continuous professional designer involvement, reducing labor costs while maintaining binding quality through algorithmic matching.
3Productivity
If automated expression transfer is implemented, then processing speed increases, but matching accuracy between reference model and object model must be ensured
Solution Approach 1:
The patent replaces manual visual inspection and judgment with automated image recognition and feature extraction algorithms. These computational methods objectively compare geometric features, texture characteristics, and structural properties between reference models and object models, ensuring consistent and accurate matching decisions without human intervention.
Solution Approach 2:
The patent implements a feedback mechanism where the matching system evaluates the similarity between reference models and object models using multiple features (geometric, textural, structural). The system provides feedback on matching quality and can adjust selection criteria to optimize both speed and accuracy, ensuring that the fastest match also meets quality thresholds.
Data Source
AI summary
Method and apparatus for transferring facial expression of digital human, electronic device, and storage medium which relates to the fields of augmented reality technologies, virtual reality technologies, computer vision technologies, deep learning technologies, or the like, and can be applied to scenarios, such as metaverse, a virtual digital human, or the like, An implementation includes: selecting an identification of a target reference model matched with an object model from a preset reference model library; the reference model library including a plurality of reference models; acquiring an expression library of the target reference model based on the identification of the target reference model; and transferring a last frame of an expression in the expression library of the target reference model into the object model to obtain a last frame of an expression of the object model.