CoMatcher: Multi-View Collaborative Feature Matching

2025-04-02

AI Summary

To address feature matching ambiguities in complex scenes caused by severe occlusion and extreme viewpoint variations, this paper proposes a multi-view collaborative matching paradigm. Moving beyond conventional pairwise matching, our approach establishes a group-level matching framework that explicitly models cross-view projection consistency and global geometric constraints while integrating complementary contextual information. We introduce a deep multi-view feature matching network coupled with a 3D scene representation learning module to enable end-to-end trajectory optimization. Evaluated under challenging conditions—including heavy occlusion and large viewpoint changes—our method significantly improves matching accuracy and trajectory completeness: matching error is reduced by 28.6%, and trajectory completion rate increases by 34.2% compared to state-of-the-art two-view methods. The proposed framework demonstrates superior robustness and generalization, setting a new benchmark for multi-view geometric reasoning in cluttered and dynamic environments.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

This paper proposes a multi-view collaborative matching strategy for reliable track construction in complex scenarios. We observe that the pairwise matching paradigms applied to image set matching often result in ambiguous estimation when the selected independent pairs exhibit significant occlusions or extreme viewpoint changes. This challenge primarily stems from the inherent uncertainty in interpreting intricate 3D structures based on limited two-view observations, as the 3D-to-2D projection leads to significant information loss. To address this, we introduce CoMatcher, a deep multi-view matcher to (i) leverage complementary context cues from different views to form a holistic 3D scene understanding and (ii) utilize cross-view projection consistency to infer a reliable global solution. Building on CoMatcher, we develop a groupwise framework that fully exploits cross-view relationships for large-scale matching tasks. Extensive experiments on various complex scenarios demonstrate the superiority of our method over the mainstream two-view matching paradigm.

Problem

• Address ambiguous estimation in image set matching
• Overcome 3D structure uncertainty from limited 2D views
• Enhance matching reliability in complex multi-view scenarios

Innovation

• Multi-view collaborative matching strategy
• Deep multi-view matcher CoMatcher
• Cross-view projection consistency utilization