Cloth Digitalization

Cloth digitalization focuses on accurately representing and manipulating cloth in digital environments, primarily for robotics and computer graphics applications. Current research emphasizes learning-based approaches, employing diffusion models, Bayesian methods, and mesh-based networks to reconstruct cloth from various input modalities (e.g., 2D images, 3D scans), often incorporating physics-based priors to improve accuracy and realism. This field is significant for advancing robotics capabilities in tasks involving cloth manipulation and for creating more realistic and interactive virtual environments, particularly in areas like virtual fashion and character animation.