Dataset

Syn4D: A Multiview Synthetic 4D Dataset

Contains synthetic multiview 4D scene sequences with RGB video, depth, masks, tracking geometry, and object mesh metadata for dynamic scene tasks.

Hugging Face ↗Jun 21, 2026Paper ↗

  • multimodal
  • multiview-video
  • depth
  • tracking

What it is

Synthetic multiview 4D video dataset with RGB, depth, masks, tracking geometry, and object mesh metadata for dynamic scene reconstruction and tracking. Synthetic multiview dynamic scenes packaged in V1 stride-1/stride-5 splits plus a simulation subset, with RGB videos, depth, masks, tracking geometry, and object mesh metadata.

Key results

  • 3 subsets: v1 stride-1, v1 stride-5, and syn4d_sim
  • Modalities include RGB videos, depth, masks, and geometry
  • License: CC-BY-4.0

Useful for benchmarking multiview 4D reconstruction, geometry tracking, and dynamic scene understanding.

One useful ML email a week.

Worth your time, not all of your time.

We track new ML work across papers, code, and data, then keep only the pieces worth a closer look. Selected, not collected.

Papers.Data.Code

The Weekly ML Signal

Selected, not collected.

One useful ML email a week: the papers, repos and datasets worth your attention.

Worth your time, not all of your time.