Reconstructing Personalized Anatomical Models for Physics-based Body Animation

摘要

We present a method to create personalized anatomical models readyrnfor physics-based animation, using only a set of 3D surface scans.rnWe start by building a template anatomical model of an averagernmale which supports deformations due to both 1) subject-specificrnvariations: shapes and sizes of bones, muscles, and adipose tissuesrnand 2) skeletal poses. Next, we capture a set of 3D scans of anrnactor in various poses. Our key contribution is formulating andrnsolving a large-scale optimization problem where we compute bothrnsubject-specific and pose-dependent parameters such that our resultingrnanatomical model explains the captured 3D scans as closely asrnpossible. Compared to data-driven body modeling techniques thatrnfocus only on the surface, our approach has the advantage of creatingrnphysics-based models, which provide realistic 3D geometry of thernbones and muscles, and naturally supports effects such as inertia,rngravity, and collisions according to Newtonian dynamics.