Feedback Matters: Augmenting Autonomous Dissection with Visual and Topological Feedback

TL;DR

This work addresses the challenge of robust autonomous tissue dissection under dynamic tissue properties and occlusions by introducing a feedback-enabled framework that reason about topological changes from endoscopic imagery after each cut. It combines an exposure maximization controller, tissue connectivity estimation, and recovery planning to localize progress and adapt actions online, integrating both planning-based and learning-based dissection methods. Key contributions include the definition of visibility-driven feedback signals, an optimization-based exposure controller, and a recovery planner that can propose corrective dissection targets, all validated through extensive real-tissue and phantom experiments showing improved autonomy, reduced errors, and enhanced robustness. The results suggest that actively controlled visibility and topological feedback can significantly advance autonomous dissection performance in realistic surgical settings, with potential impact on reliability and safety in minimally invasive procedures.

Abstract

Autonomous surgical systems must adapt to highly dynamic environments where tissue properties and visual cues evolve rapidly. Central to such adaptability is feedback: the ability to sense, interpret, and respond to changes during execution. While feedback mechanisms have been explored in surgical robotics, ranging from tool and tissue tracking to error detection, existing methods remain limited in handling the topological and perceptual challenges of tissue dissection. In this work, we propose a feedback-enabled framework for autonomous tissue dissection that explicitly reasons about topological changes from endoscopic images after each dissection action. This structured feedback guides subsequent actions, enabling the system to localize dissection progress and adapt policies online. To improve the reliability of such feedback, we introduce visibility metrics that quantify tissue exposure and formulate optimal controller designs that actively manipulate tissue to maximize visibility. Finally, we integrate these feedback mechanisms with both planning-based and learning-based dissection methods, and demonstrate experimentally that they significantly enhance autonomy, reduce errors, and improve robustness in complex surgical scenarios.

Figures (6)

• Figure 1: Illustration of the importance of exposure maximization for tissue dissection outcome verification. When tissue is dissected, an estimation method is uncertain about tissue connectivity condition due to lack of stretching. A proper controller should not only stretch the dissection site, but also consider its visibility under the camera observation model.
• Figure 2: Our Autonomous Dissection Pipeline. Our framework operates in a feedback loop. The process begins with a human-provided dissection goal $\mathcal{G}_{\text{human}}$, which the autonomous agent executes. Next, an exposure maximization controller manipulates the tissue to improve visibility for the subsequent error estimation. Finally, a recovery planner estimates the remaining tissue connectivity. If the dissection is incomplete, it generates a corrective goal, and the loop repeats until the task is successfully completed or a maximum number of attempts is reached.
• Figure 3: Qualitative results of the recovery feedback planner across various tissues and conditions. After executing the last dissection goal, each sub-figure shows the resulting tissue connectivity and the planner’s next proposed target.
• Figure 4: Comparison of our exposure maximization control and naive normal for improving visibility of the dissected region. For the same amount of retraction, our method achieves significantly more exposure of the dissected area.
• Figure 5: Comparison of dissection performance over three attempts: (a) remaining attachments, (b) effective cut ratio, and (c) excessive cut ratio.