GazeScope: A Framework of Gaze Attention-Based Automatic Field-of-View Adjustment for Laparoscopic Robots
Jing Zhang, Baichuan Wang, Zhijie Pan, Mengtang Li
Abstract
The procedure of laparoscopic minimally invasive surgery (MIS) heavily relies on the effective and efficient adjustment of the laparoscopic field-of-view (FoV). However, most existing robot-assisted laparoscopic FoV adjustment methods either require additional surgeon interactions or neglect surgeon's intentions. This paper therefore proposes GazeScope, a novel framework for automatic laparoscopic FoV adjustment, which considers gaze attention, the positions of surgical tools in the image, and the eye-hand consistency. A gaze attention-based FoV unlocking strategy is proposed to identify and eliminate unnecessary FoV adjustments, thereby improving the stability of the surgical view. Compared to traditional image-based visual servoing (IBVS) methods, GazeScope offers FoV adjustment that better aligns with manual adjustments by professionals. Meanwhile, GazeScope reduces unnecessary FoV adjustments by at least 66% and adjusting time by 50% in representative test cases, demonstrating its ability to provide a more stable operational view.
BibTeX
@inproceedings{ral2025_gazescopeaframew,
title = {GazeScope: A Framework of Gaze Attention-Based Automatic Field-of-View Adjustment for Laparoscopic Robots},
author = {Jing Zhang and Baichuan Wang and Zhijie Pan and Mengtang Li},
booktitle = {RA-L 2025},
year = {2025}
}