🤖 AI Summary
To address point cloud tracking failure and erroneous data association caused by large soft-tissue deformations in endoscopic surgery, this paper proposes a physics-guided deep non-rigid point cloud registration framework. Methodologically, we pioneer the integration of Position-Based Dynamics (PBD) physics simulation into the training pipeline to generate high-fidelity deformation correspondences—overcoming the fundamental challenge of inaccessible ground-truth registrations in real surgical scenarios. We further design an end-to-end deep point cloud matching network that jointly enforces geometric consistency and physical plausibility, enabling robust data association and accurate non-rigid registration. Evaluated on multiple challenging surgical datasets exhibiting severe tissue deformation, our method significantly outperforms existing state-of-the-art approaches: achieving a 23.6% improvement in tracking accuracy and a 31.4% reduction in reconstruction consistency error. This establishes a reliable perceptual foundation for real-time intraoperative navigation under dynamic tissue deformation.
📝 Abstract
Manipulation of tissue with surgical tools often results in large deformations that current methods in tracking and reconstructing algorithms have not effectively addressed. A major source of tracking errors during large deformations stems from wrong data association between observed sensor measurements with previously tracked scene. To mitigate this issue, we present a surgical perception framework, SuPerPM, that leverages learning-based non-rigid point cloud matching for data association, thus accommodating larger deformations. The learning models typically require training data with ground truth point cloud correspondences, which is challenging or even impractical to collect in surgical environments. Thus, for tuning the learning model, we gather endoscopic data of soft tissue being manipulated by a surgical robot and then establish correspondences between point clouds at different time points to serve as ground truth. This was achieved by employing a position-based dynamics (PBD) simulation to ensure that the correspondences adhered to physical constraints. The proposed framework is demonstrated on several challenging surgical datasets that are characterized by large deformations, achieving superior performance over state-of-the-art surgical scene tracking algorithms.
