MambaNetLK: Enhancing Colonoscopy Point Cloud Registration with Mamba

2025-10-31

AI Summary

In colonoscopy, repetitive tissue textures and preoperative-intraoperative domain shift degrade feature discriminability and stability in 3D point cloud registration, undermining lesion localization and navigation reliability. To address this, we propose MambaNetLK—a novel framework that introduces the Mamba state-space model for correspondence-free point cloud registration, serving as a cross-modal feature extractor. Integrated with PointNetLK and Lucas-Kanade iterative optimization, it enables long-range dependency modeling and high generalization capability at linear time complexity. Evaluated on our newly constructed clinical dataset C3VD-Raycasting-10k, MambaNetLK reduces rotational error by 56.04% and translational error by 26.19% over the best-performing baseline. Moreover, it demonstrates strong generalization and robustness to initial pose uncertainty on ModelNet40. This work is the first to leverage Mamba’s selective state-space modeling for efficient, accurate, and domain-robust point cloud registration in surgical navigation.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Machine Learning: Learning with Manifolds
• Natural Language Processing: Safety and Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Accurate 3D point cloud registration underpins reliable image-guided colonoscopy, directly affecting lesion localization, margin assessment, and navigation safety. However, biological tissue exhibits repetitive textures and locally homogeneous geometry that cause feature degeneracy, while substantial domain shifts between pre-operative anatomy and intra-operative observations further degrade alignment stability. To address these clinically critical challenges, we introduce a novel 3D registration method tailored for endoscopic navigation and a high-quality, clinically grounded dataset to support rigorous and reproducible benchmarking. We introduce C3VD-Raycasting-10k, a large-scale benchmark dataset with 10,014 geometrically aligned point cloud pairs derived from clinical CT data. We propose MambaNetLK, a novel correspondence-free registration framework, which enhances the PointNetLK architecture by integrating a Mamba State Space Model (SSM) as a cross-modal feature extractor. As a result, the proposed framework efficiently captures long-range dependencies with linear-time complexity. The alignment is achieved iteratively using the Lucas-Kanade algorithm. On the clinical dataset, C3VD-Raycasting-10k, MambaNetLK achieves the best performance compared with the state-of-the-art methods, reducing median rotation error by 56.04% and RMSE translation error by 26.19% over the second-best method. The model also demonstrates strong generalization on ModelNet40 and superior robustness to initial pose perturbations. MambaNetLK provides a robust foundation for 3D registration in surgical navigation. The combination of a globally expressive SSM-based feature extractor and a large-scale clinical dataset enables more accurate and reliable guidance systems in minimally invasive procedures like colonoscopy.

Problem

• Addressing feature degeneracy in colonoscopy point cloud registration
• Overcoming domain shifts between pre-operative and intra-operative data
• Enhancing alignment stability for surgical navigation accuracy

Innovation

• Uses Mamba State Space Model for feature extraction
• Integrates PointNetLK with Lucas-Kanade algorithm
• Leverages large-scale clinical dataset C3VD-Raycasting-10k