Skeleton-free Pose Transfer for Stylized 3D Characters
This addresses a novel scenario for 3D animation and game design by enabling pose transfer for skeleton-free characters with diverse shapes and topologies, though it is incremental in advancing pose transfer methods.
The paper tackles the problem of automatically transferring poses between stylized 3D characters without skeletal rigging, achieving effective results on unseen characters and inanimate objects through a novel pose transfer network.
We present the first method that automatically transfers poses between stylized 3D characters without skeletal rigging. In contrast to previous attempts to learn pose transformations on fixed or topology-equivalent skeleton templates, our method focuses on a novel scenario to handle skeleton-free characters with diverse shapes, topologies, and mesh connectivities. The key idea of our method is to represent the characters in a unified articulation model so that the pose can be transferred through the correspondent parts. To achieve this, we propose a novel pose transfer network that predicts the character skinning weights and deformation transformations jointly to articulate the target character to match the desired pose. Our method is trained in a semi-supervised manner absorbing all existing character data with paired/unpaired poses and stylized shapes. It generalizes well to unseen stylized characters and inanimate objects. We conduct extensive experiments and demonstrate the effectiveness of our method on this novel task.