Attention-guided robotic laparoscope control

See What Matters Closed-Loop Laparoscopic Field-of-view Control

A learned surgical-attention target is stabilized online, mapped into laparoscope motion under the RCM constraint, and executed on a physical robotic platform for autonomous FoV adjustment.

Watch video View pipeline

Concept

From assistant observation to perception-action control

Traditional camera assistance depends on open-loop human coordination. The proposed framework closes the loop by estimating task-relevant surgical attention, translating the focus target into control commands, and updating the laparoscope viewpoint through robotic motion.

Open-loop human coordination and closed-loop robotic FoV control comparison — Attention turns visual relevance into a closed-loop control reference for robotic laparoscope movement.

System pipeline

Focus region tracking, vision-guided FoV control, and robotic execution

Intraoperative video is processed by a focus-region tracking module, stabilized with prediction and re-ranking,
mapped from image coordinates into laparoscope pose targets, and executed by a robotic arm through inverse kinematics.

System pipeline for vision input, focus region tracking, vision guided FoV control, and robotic arm execution — The full perception-control loop links surgical attention detection to RCM-aware robotic camera motion.

Video demo

Physical closed-loop FoV adjustment

The demo visualizes the robotic laparoscope control process under learned surgical attention guidance, showing how the field of view is updated toward the target operative area.