Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MGCA-Net: Multi-Graph Contextual Attention Network for Two-View Correspondence Learning

Shuyuan Lin, Mengtin Lo, Haosheng Chen, Yanjie Liang, Qiangqiang Wu

TL;DR

MGCA-Net tackles robust two-view correspondence under high outlier conditions by introducing Contextual Geometric Attention (CGA) and Cross-Stage Multi-Graph Consensus (CSMGC). CGA fuses global context with local geometry through Context Position Attention and a multi-branch feed-forward path, while CSMGC enforces geometric consensus across stages via cross-stage sparse graphs and annular convolution. The paper reports substantial gains in outlier rejection and camera pose estimation on YFCC100M and SUN3D compared with both traditional and state-of-the-art learning-based methods, validating improved robustness and representation. Together, these components enable progressive refinement of correspondences and geometry-aware feature interaction, with potential for unsupervised extensions in future work.

Abstract

Two-view correspondence learning is a key task in computer vision, which aims to establish reliable matching relationships for applications such as camera pose estimation and 3D reconstruction. However, existing methods have limitations in local geometric modeling and cross-stage information optimization, which make it difficult to accurately capture the geometric constraints of matched pairs and thus reduce the robustness of the model. To address these challenges, we propose a Multi-Graph Contextual Attention Network (MGCA-Net), which consists of a Contextual Geometric Attention (CGA) module and a Cross-Stage Multi-Graph Consensus (CSMGC) module. Specifically, CGA dynamically integrates spatial position and feature information via an adaptive attention mechanism and enhances the capability to capture both local and global geometric relationships. Meanwhile, CSMGC establishes geometric consensus via a cross-stage sparse graph network, ensuring the consistency of geometric information across different stages. Experimental results on two representative YFCC100M and SUN3D datasets show that MGCA-Net significantly outperforms existing SOTA methods in the outlier rejection and camera pose estimation tasks. Source code is available at http://www.linshuyuan.com.

MGCA-Net: Multi-Graph Contextual Attention Network for Two-View Correspondence Learning

TL;DR

MGCA-Net tackles robust two-view correspondence under high outlier conditions by introducing Contextual Geometric Attention (CGA) and Cross-Stage Multi-Graph Consensus (CSMGC). CGA fuses global context with local geometry through Context Position Attention and a multi-branch feed-forward path, while CSMGC enforces geometric consensus across stages via cross-stage sparse graphs and annular convolution. The paper reports substantial gains in outlier rejection and camera pose estimation on YFCC100M and SUN3D compared with both traditional and state-of-the-art learning-based methods, validating improved robustness and representation. Together, these components enable progressive refinement of correspondences and geometry-aware feature interaction, with potential for unsupervised extensions in future work.

Abstract

Two-view correspondence learning is a key task in computer vision, which aims to establish reliable matching relationships for applications such as camera pose estimation and 3D reconstruction. However, existing methods have limitations in local geometric modeling and cross-stage information optimization, which make it difficult to accurately capture the geometric constraints of matched pairs and thus reduce the robustness of the model. To address these challenges, we propose a Multi-Graph Contextual Attention Network (MGCA-Net), which consists of a Contextual Geometric Attention (CGA) module and a Cross-Stage Multi-Graph Consensus (CSMGC) module. Specifically, CGA dynamically integrates spatial position and feature information via an adaptive attention mechanism and enhances the capability to capture both local and global geometric relationships. Meanwhile, CSMGC establishes geometric consensus via a cross-stage sparse graph network, ensuring the consistency of geometric information across different stages. Experimental results on two representative YFCC100M and SUN3D datasets show that MGCA-Net significantly outperforms existing SOTA methods in the outlier rejection and camera pose estimation tasks. Source code is available at http://www.linshuyuan.com.
Paper Structure (18 sections, 12 equations, 4 figures, 3 tables)

This paper contains 18 sections, 12 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Traditional Methods
  4. Learning-Based Methods
  5. Attention-Based Methods
  6. Methodology
  7. Problem Formulation
  8. Contextual Geometric Attention Module
  9. 3.2.1 Context Position Attention
  10. Cross-Stage Multi-Graph Consensus Module
  11. Loss Function
  12. Experiments
  13. Implementation Details
  14. Datasets and Evaluation Metrics
  15. Outlier Rejection
  16. ...and 3 more sections

Figures (4)

  • Figure 1: Overall architecture of the proposed MGCA-Net.
  • Figure 2: Pipeline of CGA.
  • Figure 3: Overall architecture of the proposed CSMGC.
  • Figure 4: Qualitative results of outlier removal. The first and second rows show outdoor scenes from YFCC100M, while the third and fourth rows depict indoor scenes from SUN3D. False matches are marked in red and correct matches are marked in green.