EquiFusion: Skeleton-Agnostic 3D Human Motion Prediction via Equivariant Latent Diffusion
The first model for Stochastic Human Motion Prediction to support training and inference on any skeleton parametization. EquiFusion unlocks training on multiple dataset and cross-dataset evaluation
Nonisotropic Gaussian Diffusion for Realistic 3D Human Motion Prediction
SkeletonDiffusion is a novel nonisotropic diffusion approach for 3D Human Motion Prediction, and the first computer vision method to show that nonisotropic diffusion leads to unequivocally better performance without computational drawbacks for a structured task. We generate diverse and realistic motions achieving state-of-the-art performance on the Human3.6M and AMASS datasets.
ZDySS – Zero-Shot Dynamic Scene Stylization using Gaussian Splatting
We introduce ZDySS, a zero-shot stylization framework for dynamic scenes, allowing our model to generalize to previously unseen style images at inference. Our method demonstrates superior performance and coherence over state-of-the-art baselines in tests on real-world dynamic scenes, making it a robust solution for practical applications.
Gaussian Splatting in Style
We are the first to employ Gaussian Splatting to solve the task of scene stylization, extending the work of neural style transfer to three spatial dimensions.