Data-driven Modeling of Skin and Muscle Deformation

How to read this

Category

Method: data-driven skin and muscle deformation model

Contributions

• Synthesizes realistic skin and muscle deformation by separating static pose-dependent shape from dynamic inertial effects
• Builds models from high-density marker data using PCA for shape and spring-damper equations for jiggle and muscle dynamics
• Drives skin deformation from skeletal motion captured with far fewer markers

Context

Extends the authors' earlier work on capturing and animating skin deformation in human motion, adding an explicit dynamic (inertial) component on top of pose-dependent deformation.Builds on: Capturing and Animating Skin Deformation in Human Motion

Correctness

Validated as a data-driven model learned from dense marker captures of real motion; readers should keep in mind it is limited to the captured subjects and motion range, and the linear PCA plus spring-damper decomposition is an approximation of true soft-tissue behavior.

Clarity

Accessible; a first pass conveys the static-versus-dynamic split, a second pass clarifies the PCA and spring-damper formulation.

How to read it

Focus on the separation of pose-dependent from inertial deformation and how the sparse-marker drive works; a second pass on the dynamics model pays off if you want to reproduce jiggle.