We develop a 3D equivariant diffusion model that jointly learns the generative process of both fragment poses and the 3D structure of the linker.

Sign languages are the primary form of communication for millions of deaf individuals.

However, developing models that can process weight-space objects is challenging due to their high dimensional nature.

We propose ManiPose, a manifold-constrained multi-hypothesis model for human-pose 2D-to-3D lifting.