# SemICP: Semantic Non-Rigid Point Cloud Registration with Elastic Energy Regularization

arXiv: https://arxiv.org/abs/2503.00972 · HTML: https://mlsanity.com/paper/2503.00972/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 30 references that are in the map.
- First posted on arXiv 2 Mar 2025.

## Builds on (most cited first, 11 of 11)

- [A large annotated medical image dataset for the development and evaluation of segmentation algorithms](https://mlsanity.com/paper/1902.09063/) (2019)
- [Deep Closest Point: Learning Representations for Point Cloud Registration](https://mlsanity.com/paper/1905.03304/) (2019)
- [Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registration](https://mlsanity.com/paper/1605.03344/) (2016)
- [Learn2Reg: comprehensive multi-task medical image registration challenge, dataset and evaluation in the era of deep learning](https://mlsanity.com/paper/2112.04489/) (2021)
- [Fast and Robust Iterative Closest Point](https://mlsanity.com/paper/2007.07627/) (2020)
- [Quasi-Newton Solver for Robust Non-Rigid Registration](https://mlsanity.com/paper/2004.04322/) (2020)
- [Fast and Robust Non-Rigid Registration Using Accelerated Majorization-Minimization](https://mlsanity.com/paper/2206.03410/) (2022)
- [Boundary Constraint-free Biomechanical Model-Based Surface Matching for Intraoperative Liver Deformation Correction](https://mlsanity.com/paper/2403.09964/) (2024)
- [Deep Learning-based Point Cloud Registration for Augmented Reality-guided Surgery](https://mlsanity.com/paper/2405.03314/) (2024)
- [LIBR+: Improving Intraoperative Liver Registration by Learning the Residual of Biomechanics-Based Deformable Registration](https://mlsanity.com/paper/2403.06901/) (2024)
- [Towards Transcervical Ultrasound Image Guidance for Transoral Robotic Surgery](https://mlsanity.com/paper/2211.16544/) (2022)

## Built on by (most cited first, 1 of 1)

- [Registration of 3D Point Sets Using Exponential-based Similarity Matrix](https://mlsanity.com/paper/2505.04540/) (2025)

## Abstract (arXiv)

Purpose: Accurate point cloud registration is essential in computer-aided interventions (CAI) to align multi-modal medical images for intraoperative guidance. Classical methods, such as Iterative Closest Point (ICP), remain attractive for their explainability and minimal training requirements, but typically ignore anatomical semantics and biomechanical properties during regularization. Methods: We present Semantic ICP (SemICP), a novel non-rigid point cloud registration framework that combines semantically informed point matching with deformation regularization. Semantic labels are used to improve correspondence matching by constraining correspondences to be anatomically consistent. A novel control-point deformation representation with linear-elastic energy regularization is introduced to encourage biomechanically plausible deformations. SemICP was evaluated on four datasets on US-CT, MR-CT, MR-MR and MR-US registration against established baselines. It was also tested with labels from AI-based segmentation in a fully automatic segmentation-registration pipeline. Results: Across all datasets, SemICP achieves lower Hausdorff distance, mean surface distance, and target registration error than competing methods. The fully automatic registration pipeline was shown to be effective for US-MR registration and to improve the alignment of expert-annotated structures. Conclusion: SemICP improves deformable point cloud registration accuracy and robustness by combining semantic correspondence constraints and linear energy regularization. Combined with AI-based segmentation, SemICP provides an effective pipeline for multi-modal registration in CAI.