Deformable State Estimation for Autonomous Surgical Tissue Retraction Under Partial Observability

2026-07-15

AI Summary

Accurately estimating the complete state of deformable tissues under sparse and noisy observations remains a significant challenge for autonomous surgical retraction. This work proposes a learning-based state estimator that reconstructs full tissue meshes within a low-dimensional PCA latent space using a multilayer perceptron, augmented with geometric-aware regularization to enforce smooth and physically plausible deformations. To the best of our knowledge, this is the first application of a geometry-regularized learning framework to the task of deformable surgical tissue retraction under partial observability. The method achieves 98.1% of the ideal state estimation performance in multi-step planning scenarios, demonstrating both high fidelity and computationally efficient inference.

Abstract

Surgical tissue retraction requires effective manipulation planning under partial and noisy perception. We study state estimation for deformable tissue retraction, where only sparse observations of the tissue surface are available at decision time. We propose a learned state estimator that reconstructs the full deformable mesh state from 40 noisy vertex observations. The estimator combines a multilayer perceptron with a low-dimensional PCA latent representation and is trained using geometry-aware regularization that encourages smooth and physically plausible deformations. We evaluate the approach in a 2D deformable sheet simulation using single-step and multi-step retraction planning. Results show that the learned estimator achieves 98.1% of oracle performance in multi-step retraction while supporting efficient inference. These results demonstrate that learned, geometry-regularized state estimation can support effective deformable manipulation under realistic perception constraints.

Problem

• deformable state estimation
• surgical tissue retraction
• partial observability
• sparse observations
• autonomous manipulation

Innovation

• deformable state estimation
• geometry-aware regularization
• partial observability
• learned mesh reconstruction
• autonomous surgical retraction