Related Experiment Video
Updated: Jan 9, 2026
06:25
Capturing Representative Hand Use at Home Using Egocentric Video in Individuals with Upper Limb Impairment
Published on: December 23, 2020
2.9K
EgoSurgery-HTS: A Dataset for Egocentric Hand-Tool Segmentation in Open Surgery Videos.
Nathan Darjana1, Ryo Fujii1, Hideo Saito1
1Faculty of Science and Technology Keio University Yokohama Japan.
Healthcare Technology Letters
|December 11, 2025
Summary
Researchers developed EgoSurgery-HTS, a new dataset for segmenting surgical tools and hands in egocentric videos. This benchmark improves understanding of surgical procedures and surgeon actions.
Area of Science:
- Computer Vision
- Medical Imaging
- Surgical Robotics
Background:
- Egocentric open-surgery videos offer detailed insights into surgical procedures and surgeon behavior.
- Pixel-level understanding of hands and surgical tools is vital for interpreting surgeon actions and intentions.
Purpose of the Study:
- Introduce EgoSurgery-HTS, a novel dataset with pixel-wise annotations for segmenting surgical tools, hands, and their interactions.
- Provide a benchmark suite for evaluating segmentation methods in egocentric surgical videos.
Main Methods:
- Developed a labelled dataset for instance segmentation of 14 surgical tool types.
- Included instance segmentation for hands.
- Implemented hand-tool segmentation to identify main operating hands and manipulated tools.
Main Results:
- Conducted extensive evaluations of state-of-the-art segmentation methods on the EgoSurgery-HTS dataset.
- Demonstrated significant improvements in hand and hand-tool segmentation accuracy compared to existing datasets.
Conclusions:
- EgoSurgery-HTS facilitates more accurate modelling of surgical procedures and human behavior.
- The dataset and benchmark suite advance research in computer vision for surgical applications.
