Stereo Video Dataset for Occluded and Unoccluded Surgical Robot Tracking

4.1 GB of video data captures surgical robot motion under both clear and operator-occluded conditions. The dataset includes 18 sequences with robot and camera motion, plus five videos where the robot is occluded by the author. Created by Yanghe Hao and shared under a CC-BY-4.0 license, it was last updated in May 2026.

Use Cases

• Training computer vision models for marker-free robot pose estimation based on stereo video sequences.
• Benchmarking tracking algorithm robustness against operator-induced occlusion mentioned in the description.
• Developing and testing differentiable rendering pipelines for surgical simulation using the provided motion data.

Strengths

• Dataset size is explicitly 4.1 GB, indicating substantial video content.
• Provides a controlled comparison with 18 unoccluded and 5 occluded video sequences.
• Clear consent and authorship are documented for the operator and author appearing in videos.

Limitations

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

Provenance

Source

Yanghe Hao

Collection Method

Likely captured using stereo camera systems during robot operation.

Freshness

Last updated 2026-05-12 15:44:43; freshness should be verified.