RayCo4D: Ray-Conditioned Continuous Motion Model for Monocular 4D Dynamic Reconstruction

Key points

• Reconstructing dynamic 3D scenes from monocular video requires inferring motion from noisy and intermittently visible observations.1
• Each node maintains an $SE(3)$ pose and body-frame velocity at sparse support times.1
• The inferred shared motion is transferred to Gaussian primitives through dual-quaternion blending, while reference-anchored local $SE(3)$ residuals capture fine-scale non-rigid deformation beyond the shared scaffold.1
• Experiments on diverse monocular dynamic scenes show that RayCo4D achieves competitive novel-view synthesis quality, improves structural and temporal consistency under rapid motion and incomplete tracks, and supports continuous-time rendering.1

Sources (1)

• [1]RayCo4D: Ray-Conditioned Continuous Motion Model for Monocular 4D Dynamic Reconstruction
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 03:17 PM

  We introduce RayCo4D, a ray-conditioned continuous motion model for monocular 4D reconstruction that explicitly separates visual measurements from latent 3D motion states.

  Reconstructing dynamic 3D scenes from monocular video requires inferring motion from noisy and intermittently visible observations.

Extractive summary: sentences quoted from the sources.