Soft Tissue Solver for Realistic Facial Animation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Realistic animation of facial expressions in computer animation is challenging due to the limited movement of the skeleton, requiring labor-intensive techniques to simulate soft tissue behavior, which hinders the ability of animated characters to convey emotions realistically.
Innovation Solution
A soft tissue solver is created by associating deformation objects with a surface mesh, allowing for automatic generation based on salient points and deformation operators, enabling quantitative aesthetics and real-time interactive animation, similar to skeleton animation using inverse kinematics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional skeleton and skin animation techniques are used, then character animation can be achieved, but realistic facial expression animation is difficult due to limited skeleton movement
Solution Approach 1:
The face is divided into multiple deformation regions (e.g., forehead, cheeks, jaw, lips) with salient points identified in each region. This segmentation allows independent control of different facial areas, enabling realistic expressions while managing complexity through regional decomposition.
Solution Approach 2:
A soft tissue solver acts as an intermediary between the skeleton and the surface mesh. It translates limited skeleton movements into realistic soft tissue deformations by computing deformation operators that affect deformation regions, bridging the gap between rigid bone structure and flexible facial appearance.
2Manufacturing precision
If manual posing and morphing techniques are used to animate facial expressions, then realistic animation can be achieved, but the process becomes labor intensive requiring many months of work
Solution Approach 1:
Salient points and deformation regions are pre-defined and automatically identified on the face model before animation. Control objects are pre-positioned at these salient points, and deformation operators are pre-configured to affect specific deformation regions. This preliminary setup enables rapid animation without manual posing for each expression.
Solution Approach 2:
The soft tissue solver automatically computes deformation operators based on control object movements without requiring manual intervention. Given salient points and a format file, the system self-configures the deformation network and autonomously generates realistic facial expressions from skeleton movements, eliminating the need for months of manual animation work.
3Manufacturing precision
If detailed manual control of surface mesh points is used, then realistic soft tissue behavior can be simulated, but the system becomes difficult to operate and control
Solution Approach 1:
Control objects serve as intermediaries that animateurs can easily manipulate, which then automatically drive the complex soft tissue solver. The animator only needs to move control objects at salient points rather than directly controlling thousands of mesh vertices, simplifying operation while maintaining deformation accuracy through the automated solver.
Solution Approach 2:
Instead of directly manipulating the complex surface mesh, the system uses simplified control objects that copy and represent the essential control points (salient points). These control objects serve as proxies that are easier to animate while the soft tissue solver copies their movement patterns and translates them into realistic mesh deformations.
Data Source
AI summary
A face, as well as any other soft tissue of a character, can be animated much in the same way that a skeleton is animated by creating a soft tissue solver attached to the surface mesh. In particular, deformation objects are associated with regions of the surface mesh. The deformation objects deform the mesh according to deformation operators in response to a change in a control object. This soft tissue solver can be generated automatically given a set of salient points specified on an input mesh and a format file for the class of objects of which the input mesh is an example. The format file specifies what the salient points are, and the relative placement of the deformation objects and control objects as functions of the salient points. Specific deformation operators can be defined and associated, through the format file, with the deformation objects and control objects. The format file for a class of objects, such as for human and humanoid heads, and related operators can be determined based on experimentation and observation of the object being modeled. Through such a soft tissue solver, the behavior of the deformation of the surface mesh is quantified and controlled by direct manipulation of the points on the mesh. Thus, the surface of the object can be animated in a manner similar to how animation of a skeleton is defined using inverse kinematics. Because the surface of the object is directly related to its appearance or aesthetics, this animation capability is referred to herein as “quantitative aesthetics.”