Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Preoperative-to-intraoperative Liver Registration for Laparoscopic Surgery via Latent-Grounded Correspondence Constraints

Ruize Cui, Jialun Pei, Haiqiao Wang, Jun Zhou, Jeremy Yuen-Chun Teoh, Pheng-Ann Heng, Jing Qin

TL;DR

Land-Reg is introduced, a correspondence-driven deformable registration framework that explicitly learns latent-grounded 2D-3D landmark correspondences as an interpretable intermediate representation to bridge cross-modal alignment.

Abstract

In laparoscopic liver surgery, augmented reality technology enhances intraoperative anatomical guidance by overlaying 3D liver models from preoperative CT/MRI onto laparoscopic 2D views. However, existing registration methods lack explicit modeling of reliable 2D-3D geometric correspondences supported by latent evidence, leading to limited interpretability and potentially unstable alignment in clinical scenarios. In this work, we introduce Land-Reg, a correspondence-driven deformable registration framework that explicitly learns latent-grounded 2D-3D landmark correspondences as an interpretable intermediate representation to bridge cross-modal alignment. For rigid registration, Land-Reg embraces a Cross-modal Latent Alignment module to map multi-modal features into a unified latent space. Further, an Uncertainty-enhanced Overlap Landmark Detector with similarity matching is proposed to robustly estimate explicit 2D-3D landmark correspondences. For non-rigid registration, we design a novel shape-constrained supervision strategy that anchors shape deformation to matched landmarks through reprojection consistency and incorporates local-isometric regularization to alleviate inherent 2D-3D depth ambiguity, while a rendered-mask alignment enforces global shape consistency. Experimental results on the P2ILF dataset demonstrate the superiority of our method on both rigid pose estimation and non-rigid deformation. Our code will be available at https://github.com/cuiruize/Land-Reg.

Preoperative-to-intraoperative Liver Registration for Laparoscopic Surgery via Latent-Grounded Correspondence Constraints

TL;DR

Land-Reg is introduced, a correspondence-driven deformable registration framework that explicitly learns latent-grounded 2D-3D landmark correspondences as an interpretable intermediate representation to bridge cross-modal alignment.

Abstract

In laparoscopic liver surgery, augmented reality technology enhances intraoperative anatomical guidance by overlaying 3D liver models from preoperative CT/MRI onto laparoscopic 2D views. However, existing registration methods lack explicit modeling of reliable 2D-3D geometric correspondences supported by latent evidence, leading to limited interpretability and potentially unstable alignment in clinical scenarios. In this work, we introduce Land-Reg, a correspondence-driven deformable registration framework that explicitly learns latent-grounded 2D-3D landmark correspondences as an interpretable intermediate representation to bridge cross-modal alignment. For rigid registration, Land-Reg embraces a Cross-modal Latent Alignment module to map multi-modal features into a unified latent space. Further, an Uncertainty-enhanced Overlap Landmark Detector with similarity matching is proposed to robustly estimate explicit 2D-3D landmark correspondences. For non-rigid registration, we design a novel shape-constrained supervision strategy that anchors shape deformation to matched landmarks through reprojection consistency and incorporates local-isometric regularization to alleviate inherent 2D-3D depth ambiguity, while a rendered-mask alignment enforces global shape consistency. Experimental results on the P2ILF dataset demonstrate the superiority of our method on both rigid pose estimation and non-rigid deformation. Our code will be available at https://github.com/cuiruize/Land-Reg.
Paper Structure (11 sections, 6 equations, 4 figures, 1 table)

This paper contains 11 sections, 6 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methodology
  3. Overview
  4. Rigid Registration via Latent-Grounded Correspondence
  5. Shape Constrained Deformable Registration
  6. Experiments
  7. Dataset and Pseudo Label Generation
  8. Implementation Details
  9. Experimental Comparisons
  10. Ablation Study
  11. Conclusion

Figures (4)

  • Figure 1: Comparison of landmark-based rigid liver registration methods.
  • Figure 2: Architecture of Land-Reg for 2D-3D liver registration.
  • Figure 3: Qualitative results from two randomly selected samples. Top: correspondence learning performance of Land-Reg, including comparisons of correspondence and overlapping landmark detection between GT and our predictions. Bottom: registration results comparison of Land-Reg and competitive methods.
  • Figure 4: Ablation results of crucial components proposed in Land-Reg.