3D Avatar Emotion Expression via Deformable Mask Rings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current social networking and mobile communication technologies lack the ability to effectively convey and understand emotions and moods in real-time, as virtual identities do not accurately represent a person's feelings or sentiments.
Innovation Solution
A method for creating three-dimensional avatars that transform two-dimensional images of human or animal faces into animated, deformable 3D models, using a deformable mask with concentric rings and mesh structures to simulate facial expressions and emotions, allowing for the representation of feelings and sentiments through mobile devices and cloud systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If two-dimensional images are used for virtual identities, then communication convenience is improved, but emotional expression capability deteriorates
Solution Approach 1:
The patent transforms two-dimensional flat images into three-dimensional avatars with depth, volume, and spatial characteristics. This dimensional transition enables the avatar to convey emotions through facial muscle movements, head orientations, and body postures that are impossible in 2D, thereby restoring emotional expression capability while maintaining digital communication convenience
Solution Approach 2:
The patent creates dynamically deformable 3D avatars that can change shape and posture in real-time based on emotional states. The avatar's facial features, head position, and body orientation are not static but can dynamically adjust to represent different emotions, providing rich emotional expression while maintaining ease of digital communication
2Device complexity
If simple virtual identities are used, then device complexity is reduced, but emotional representation accuracy deteriorates
Solution Approach 1:
The patent segments the avatar into multiple independent components including head, torso, limbs, and facial features. Each segment can be independently controlled and animated, allowing complex emotional expressions to be achieved through coordinated movements of individual segments rather than requiring a completely complex monolithic model
Solution Approach 2:
The patent applies different levels of detail and deformability to different parts of the avatar. The head and facial features have high local quality with detailed muscle structures for subtle emotional expressions, while the body has sufficient detail for posture expression but less complexity than the face. This localized differentiation achieves accurate emotional representation without uniformly increasing overall complexity
3Loss of information
If three-dimensional avatars with emotional expression are created, then emotional communication is improved, but processing complexity increases
Solution Approach 1:
The patent pre-defines a set of standard emotional states and their corresponding avatar configurations. When a user experiences an emotion, the system retrieves and applies the pre-configured avatar state rather than computing it in real-time, significantly reducing processing complexity while maintaining rich emotional communication capability
Solution Approach 2:
The patent controls avatar emotional expression by adjusting a limited set of key parameters such as head orientation angles, facial muscle deformation values, and body posture coordinates. By constraining the expression space to these essential parameters rather than controlling every vertex, the system achieves effective emotional communication with reduced processing complexity
Data Source
AI summary
A two-dimensional image is transformed into at least one portion of a human or animal body into a three-dimensional model. An image is acquired that includes the at least one portion of the human or animal body. An identification is made of the at least one portion within the image. Searches are made for features indicative of the at least one portion of the human or animal body within the at least one portion. One or more identifications are made of a set of landmarks corresponding to the features. An alignment is a deformable mask including the set of landmarks. The deformable mask includes a number of meshes corresponding to the at least one portion of the human or animal body. The 3D model is animated by dividing it into concentric rings, quasi rings and applying different degrees of rotation to each ring.