MACAW: Reliable And Efficient Surgical Debridement Using Monocular Adaptive Compact Attention Windows
Abstract
Augmenting the dexterity of human surgeons has the potential to free them from tedious subtasks. We consider debridement (removal of diseased or dead tissue fragments), which is challenging due to imprecision in spatial perception and cable actuation. We develop an augmented dexterity system for surgical debridement that uses visual servoing to align the cable-driven gripper with the target position in the image plane, and then introduces a novel approach to depth control, MACAW: Monocular Adaptive Compact Attention Windows. Across 100 physical trials using the da Vinci Research Kit (dVRK) robot, camera-frame servoing reduced average gripper position offset from 37 to fewer than 5 pixels within 4 optimization steps, taking an average of only 0.39s. MACAW significantly outperforms procedural and learned VLA baselines, achieving a 93% success rate at 11 seconds per fragment, yielding a throughput of 304 fragments per hour. Extending MACAW to a bimanual debridement setup maintains a 92% success rate at an average of 7 seconds per fragment, increasing the throughput to 473 fragments per hour.
Method Overview
Camera-frame servoing
Relative Position Recognition using the upper MACAW
Interactive Depth Perception using the lower MACAW
Key Contributions
- An augmented dexterity system for surgical debridement using only one fixed monocular RGB camera for perception,
- A novel camera-frame servoing technique that enables dynamic gripper alignment with goal positions in the image plane,
- MACAW: Monocular Adaptive Compact Attention Windows, a novel depth perception method,
- Unimanual (one-arm) physical experiments with deformable fragments that suggest up to 93% success rate with throughput of 304 fragments per hour, significantly outperforming procedural baselines and learned policies (including ACT, Diffusion Policy, π₀, and π₀.₅),
- Bimanual debridement with MACAW maintains a 92% success rate and increases throughput to 473 fragments per hour.
Experimental Results
Comparison Study (×10 Speed)
Procedural Baseline 1
ACT
Diffusion Policy
π₀
π₀.₅
MACAW (Ours)
Failure Cases
Failure Case 1: Fragment Rolls Over
Failure Case 2: Release Failure
Extension Testing
Extension Test 1
Extension Test 2
Bimanual Debridement with MACAW
BibTeX
@article{MACAW2026,
title={MACAW: Reliable And Efficient Surgical Debridement Using Monocular Adaptive Compact Attention Windows},
journal={Under Review},
year={2026}
}