Open4DHOI: Open-Source Human-Object Interaction Video and Motion Data

Open4DHOI is an open-source release of annotated video data for human-object interaction research. It contains source videos, object and human masks, reconstructed human motion, and point annotations, generated from records with full annotation progress. The dataset was authored by acane2 and last updated on 2026-05-19.

Use Cases

• Train video segmentation models based on provided object and human masks.
• Develop human motion reconstruction algorithms based on the motion data files.
• Benchmark 3D human-object interaction tasks based on the combined video, mesh, and annotation data.

Strengths

• Data is released as open-source, facilitating reproducibility.
• Includes multiple aligned data modalities: video, 3D meshes, masks, and motion data.

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

huggingface

Collection Method

Generated from upload_records.json for records with annotation_progress == 4.

Freshness

Last updated 2026-05-19 08:13:25; freshness should be verified.