DL3DV-10K Sample: 11 Scenes with 4K Video and 3D Reconstruction Data

DL3DV-10K Sample is a subset of a larger dataset for 3D computer vision, created by DL3DV. It contains 11 scenes, each with a raw 4K video file and processed 3D reconstruction results from COLMAP. The dataset was last updated on the Hugging Face platform on 2024-01-05.

Use Cases

• Training 3D scene reconstruction models based on provided COLMAP sparse point clouds and camera parameters.
• Benchmarking novel view synthesis methods using the paired high-resolution 4K video and 3D data.
• Developing video-based SLAM (Simultaneous Localization and Mapping) systems using the processed feature matches and camera trajectories.
• Researching multi-view geometry and feature matching using the provided features.h5 and matches.h5 files.

Strengths

• Includes high-resolution 4K video for each of the 11 scenes.
• Provides processed 3D reconstruction outputs (COLMAP sparse models, camera parameters, feature matches) for immediate use.
• File structure is organized and documented, with an example provided in the description.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

DL3DV

Collection Method

Likely involves capturing video and processing it with the COLMAP structure-from-motion pipeline.

Time Range

null

Freshness

Last updated 2024-01-05 02:21:19; freshness should be verified.

Geography

null