Guess The Unseen: Dynamic 3D Scene Reconstruction from Partial 2D Glimpses

Input: Monocular Video Novel Pose Synthesis Novel View Synthesis Output: Animatable Person 3D-GS Occluded partial observations For Teaser (light) Figure 1. We present a method to reconstruct dynamic scenes from a monocular video capturing partial 2D observations. As a key advantage, our method can estimate the unseen body parts by leveraging a pre-trained diffusion model [40] via SDS method [37]. The reconstructed scenes can be rendered to any viewpoint and each human body can be transformed into any body posture controlled by SMPL [27] parameters.