Egoexor

EgoExOR is a dataset for comprehensive understanding of surgical activities, featuring both egocentric and exocentric video perspectives. The dataset was created by authors including Ege Özsoy and Arda Mamur, with official code for a paper submitted to the NeurIPS 2025 Datasets & Benchmarks Track. It was last updated on the Hugging Face platform on September 29, 2025.

Use Cases

• Train computer vision models for surgical activity recognition based on multiview video data.
• Develop and benchmark models for egocentric action understanding in clinical environments.
• Research human-robot collaboration in surgery based on detailed visual data of operating room activities.

Strengths

• Dataset is associated with a paper submitted to the NeurIPS 2025 conference, suggesting academic rigor.
• Features both egocentric and exocentric video perspectives, which likely provides complementary visual information.
• Last updated on 2025-09-29, indicating recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and total size are unknown, which may limit suitability assessment.
• License information is unavailable, which could restrict usage.

Provenance

Source

Hugging Face dataset uploaded by author 'ardamamur'.

Collection Method

Likely collected from operating room environments using egocentric (wearable) and exocentric (fixed) cameras.

Time Range

null

Freshness

Last updated 2025-09-29 15:00:32; freshness should be verified.

Geography

null