SOMA: From Surface Observations to Muscle Anatomy
Researchers have developed SOMA, a novel model that infers muscle deformations from surface observations using RGB cameras. This method aims to overcome the limitations of existing parametric body models, which typically only capture the skin's 3D surface and lack insight into underlying biomechanical structures. SOMA provides anatomically grounded animations without the computational expense of traditional simulations, offering a scalable and cost-effective solution for applications in medicine, sports, and entertainment. The project also introduces SKIM, a subject-specific soft-tissue deformation dataset, and makes both data and code publicly available. AI
IMPACT Enables more realistic and biomechanically accurate virtual human models for various applications.
