EnvPoser: Environment-aware Realistic Human Motion Estimation from Sparse Observations with Uncertainty Modeling

This supplementary material provides additional information to complement our main paper. Sec. A details the network architecture and training procedures of our proposed method. Sec. B summarizes the datasets used in our study. Sec. C presents extended ablation studies, encompassing various environment points sampling strategy and further analysis of component design. Sec. D showcases additional qualitative comparisons between our approach and state-of-the-art methods. Finally, Sec. E summarizes our work, discusses its limitations, and outlines potential directions for future research.