Syn4D: Large-Scale Synthetic Multiview Dataset for Dynamic Scenes

Syn4D is a large-scale, fully-synthetic multiview dataset of dynamic scenes created by Voxel51. It is designed for research in 4D reconstruction, depth estimation, 3D point tracking, novel-view synthesis, and human pose estimation. The dataset was last updated on 2026-05-26.

Use Cases

• Train 4D reconstruction models based on dense, complete geometric annotations.
• Develop depth estimation algorithms using per-pixel depth maps.
• Research 3D point tracking leveraging dense long-range 3D point tracks.
• Advance novel-view synthesis for dynamic scenes using multi-view camera trajectories.
• Improve human pose estimation models using parametric SMPL-X human body annotations.

Strengths

• Provides dense, complete, and accurate geometric annotations as described.
• Covers a diverse collection of dynamic scenes.
• Includes multi-view data with camera trajectories and long-range 3D point tracks.

Limitations

• Dataset size, row count, and file formats are unknown, limiting suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data is fully synthetic, which may limit real-world applicability.

Provenance

Source

Voxel51

Collection Method

Fully-synthetic generation.

Freshness

Last updated 2026-05-26 22:23:07; freshness should be verified.