An attention guided local global feature fusion network for robust point cloud registration
Abstract
Point cloud registration in complex environments remains challenging because measurement noise, partial overlap, and structural variations can reduce correspondence reliability and registration robustness, while computational efficiency is also critical for practical deployment. These difficulties are closely related to insufficient adaptive modeling of local geometric structures, inadequate suppression of weak global graph connections, and limited exploitation of inter-cloud similarity and discrepancy information. To address these issues, we propose AGLGNet, an attention guided local global feature fusion network for robust point cloud registration. First, the Adaptive Local Neighborhood Graph Calibration Module constructs geometry-guided neighborhood representations and adaptively recalibrates neighboring features, thereby enhancing local structural discrimination and reducing sensitivity to noise. Second, the Dynamic Weight Global Graph Learning Module adjusts graph connection weights according to inter-point correlations, suppressing weak relationships and strengthening long-range contextual modeling. Third, the Hybrid Attention Feature Fusion Module jointly models inter-cloud similarity and discrepancy information through cross-attention and further refines the fused representations using self-attention. Extensive experiments on 3DMatch, ModelNet40, KITTI, the Stanford 3D Scanning Repository, and real-world workpiece data demonstrate that the proposed method achieves low registration errors and strong robustness under complete, partial, and noisy point cloud conditions. It also maintains an inference time on the order of 10 ⁻² s, providing a favorable balance among registration accuracy, robustness, and computational efficiency.
Related
- Parallel Local–Global Feature Extraction and Cross-Attention Gated Fusion for Network Intrusion Detection
- VesselRegNet: Multi-Scale Feature Fusion Network for Multi-Modal Joint Retinal Vessel Segmentation and Registration
- Hybrid feature pyramid attention and capsule network (FpaNet+CapsNet) for enhanced spatial and contextual feature extraction for improved brain tumor segmentation
- Feature-aware attention convolutional neural network for efficient pulmonary fibrosis classification from chest CT images with FPGA deployment
- IFMA-Net: Interactive Feature Modulated Attention Network for Accurate Identification of Cotton Pests and Diseases from Field Images
BibTeX
@article{ren2026attention,
title = {{An attention guided local global feature fusion network for robust point cloud registration}},
author = {Linggen Ren and Dong Zhang and Zhenhang Ying and Teng Wu and Qinghe Xu},
journal = {Scientific Reports},
year = {2026},
month = oct,
doi = {10.1038/s41598-026-68790-2},
url = {https://www.nature.com/articles/s41598-026-68790-2},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - An attention guided local global feature fusion network for robust point cloud registration
AU - Linggen Ren
AU - Dong Zhang
AU - Zhenhang Ying
AU - Teng Wu
AU - Qinghe Xu
PY - 2026
DA - 2026/10/08
JO - Scientific Reports
DO - 10.1038/s41598-026-68790-2
UR - https://www.nature.com/articles/s41598-026-68790-2
AB - Point cloud registration in complex environments remains challenging because measurement noise, partial overlap, and structural variations can reduce correspondence reliability and registration robustness, while computational efficiency is also critical for practical deployment. These difficulties are closely related to insufficient adaptive modeling of local geometric structures, inadequate suppression of weak global graph connections, and limited exploitation of inter-cloud similarity and discrepancy information. To address these issues, we propose AGLGNet, an attention guided local global feature fusion network for robust point cloud registration. First, the Adaptive Local Neighborhood Graph Calibration Module constructs geometry-guided neighborhood representations and adaptively recalibrates neighboring features, thereby enhancing local structural discrimination and reducing sensitivity to noise. Second, the Dynamic Weight Global Graph Learning Module adjusts graph connection weights according to inter-point correlations, suppressing weak relationships and strengthening long-range contextual modeling. Third, the Hybrid Attention Feature Fusion Module jointly models inter-cloud similarity and discrepancy information through cross-attention and further refines the fused representations using self-attention. Extensive experiments on 3DMatch, ModelNet40, KITTI, the Stanford 3D Scanning Repository, and real-world workpiece data demonstrate that the proposed method achieves low registration errors and strong robustness under complete, partial, and noisy point cloud conditions. It also maintains an inference time on the order of 10 ⁻² s, providing a favorable balance among registration accuracy, robustness, and computational efficiency.
SN - 2045-2322
ER -