Language-Conditioned Behavioral Cloning for Autonomous Endoscope Guidance
Language-Conditioned Behavioral Cloning for Autonomous Endoscope Guidance
This research paper proposes a language-conditioned behavioral cloning approach to automate endoscope guidance during minimally invasive surgery. The authors define six clinically relevant guidance behaviors and collect 235 subtask demonstrations. Preliminary experiments on phantoms show the policy achieves 100% success on focus, overview, and stop behaviors, 80% on search-tip and follow-tip behaviors, and 50% on horizon leveling. The results indicate feasibility for learning heterogeneous surgeon-requested camera behaviors with a single visuomotor policy, though failure modes suggest future work is needed to incorporate temporal context for more demanding tasks.
The study demonstrates a language-conditioned behavioral cloning policy for autonomous endoscope guidance, achieving high success rates on specific subtasks in phantom experiments while noting limitations in more complex behaviors.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
number of subtask demonstrations collected
235
View original evidence
levant guidance behaviors and collect 235 subtask demonstrations. In preliminary phantom experiments, our policy achieves 100% success on focus, overview, and stop, 80% on search-tip and follow-tip behaviors, and 50% on horizon leveling. These results demonstrate the feasibility of learning heterogeneous surgeon-requested camera behaviors with a single visuomotor policy, while the observed failure modes motivate futu
Open source S3number of guidance behaviors defined
6
View original evidence
arXiv:2610.41978v1 Announce Type: new Abstract: Automated endoscope guidance reduces the need for continuous manual camera control during minimally invasive surgery. However, the desired camera motion depends on the surgeon's immediate intentions and cannot always be inferred from the visual operative scene alone. We propose to automate the task using language-conditioned behavioral cloning; we define 6 clinically re
Open source S2
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
arXiv:2610.41978v1 Announce Type: new Abstract: Automated endoscope guidance reduces the need for continuous manual camera control during minimally invasive surgery. However, the desired camera motion depends on the surgeon's immediate intentions and cannot always be inferred from the visual operative scene alone. We propose to automate the task using language-conditioned behavioral cloning; we define 6 clinically re
Open source S2
levant guidance behaviors and collect 235 subtask demonstrations. In preliminary phantom experiments, our policy achieves 100% success on focus, overview, and stop, 80% on search-tip and follow-tip behaviors, and 50% on horizon leveling. These results demonstrate the feasibility of learning heterogeneous surgeon-requested camera behaviors with a single visuomotor policy, while the observed failure modes motivate futu
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org