Avatar Eye Expression Driving from Eyelid and Gaze Coefficients
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately simulate facial expressions, particularly eye expressions, when user identity and posture information is unavailable, such as when wearing virtual reality glasses.
Innovation Solution
Determine expression coefficients, including a first eye coefficient representing eyelid spacing and a second eye coefficient representing line-of-sight information, to simulate eye expressions without requiring identity or posture information, using key point identification and line-of-sight tracking algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional facial expression simulation methods are used, then identity and posture information can be obtained, but the method cannot work when wearing virtual reality glasses and requires complex 3D reconstruction
Solution Approach 1:
The patent extracts only the essential eye region information (eyelid spacing and line-of-sight) from the full face, eliminating the need for complex 3D reconstruction of the entire face. This extraction approach maintains expression simulation accuracy while significantly reducing system complexity and removing the requirement for identity and posture data.
Solution Approach 2:
The patent uses 2D image copying and analysis instead of 3D reconstruction. By analyzing key points in 2D images to determine eyelid spacing and line-of-sight direction, the system achieves expression simulation without requiring complex 3D models or additional sensors.
2Measurement precision
If comprehensive facial data is collected for expression simulation, then accuracy can be improved, but the processing time and computational resources increase
Solution Approach 1:
The patent extracts only the critical eye region features (eyelid spacing and line-of-sight) from the entire face, eliminating the need to process comprehensive facial data. This selective extraction maintains expression simulation accuracy while significantly reducing computational load and processing time.
Solution Approach 2:
The patent segments the facial expression problem into specific eye region parameters (eyelid spacing, line-of-sight direction) rather than processing the entire face. This segmentation allows for faster, more efficient computation while maintaining the accuracy needed for realistic expression simulation.
3Reliability
If identity and posture information are required for eye expression simulation, then accuracy can be maintained, but the system cannot function when wearing virtual reality glasses
Solution Approach 1:
The patent extracts eye expression information directly from the eye region itself, eliminating the dependency on identity and posture data. This extraction method enables the system to function with virtual reality glasses, as it only requires images captured by the glasses' cameras without needing external sensors or pre-stored user data.
Solution Approach 2:
The patent enables the eye expression simulation system to be self-sufficient by extracting all necessary information (eyelid spacing, line-of-sight) directly from the eye region images captured by virtual reality glasses. The system does not require external identity verification or posture sensing, making it fully compatible with VR device usage.
Data Source
AI summary
The present disclosure provides an expression driving method, apparatus, device and storage medium, wherein the method includes: acquiring an image to be processed of a target object, the image to be processed including an eye region of the target object; determining an expression coefficient of at least one dimension of the target object based on the image to be processed, the expression coefficient of at least one dimension including a first eye coefficient and a second eye coefficient, wherein the first eye coefficient is configured for representing an eyelid spacing of the target object, and the second eye coefficient is configured for representing line-of-sight information of the target object; and driving an avatar based on the determined expression coefficient of at least one dimension.