Abstract:This paper studies graph clustering with application to feature matching and proposes an effective method, termed as GC-LAC, that can establish reliable feature correspondences and simultaneously discover all potential visual patterns. In particular, we regard each putative match as a node and encode the geometric relationships into edges where a visual pattern sharing similar motion behaviors corresponds to a strongly connected subgraph. In this setting, it is natural to formulate the feature matching task as a graph clustering problem. To construct a geometric meaningful graph, based on the best practices, we adopt a local affine strategy. By investigating the motion coherence prior, we further propose an efficient and deterministic geometric solver (MCDG) to extract the local geometric information that helps construct the graph. The graph is sparse and general for various image transformations. Subsequently, a novel robust graph clustering algorithm (D2SCAN) is introduced, which defines the notion of density-reachable on the graph by replicator dynamics optimization. Extensive experiments focusing on both the local and the whole of our GC-LAC with various practical vision tasks including relative pose estimation, homography and fundamental matrix estimation, loop-closure detection, and multimodel fitting, demonstrate that our GC-LAC is more competitive than current state-of-the-art methods, in terms of generality, efficiency, and effectiveness. The source code for this work is publicly available at: https://github.com/YifanLu2000/GCLAC.

Feature Matching via Graph Clustering with Local Affine Consensus

Matching Images Based on Consistency Graph and Region Adjacency Graphs.

Hierarchical Clustering Matching for Features with Repetitive Patterns in Visual Odometry

Robust Image Matching Via Local Graph Structure Consensus

Geometrically Based Linear Iterative Clustering for Quantitative Feature Correspondence

MCM: an Efficient Geometric Constraint Method for Robust Local Feature Matching.

Learning Bipartite Graph Matching for Robust Visual Localization.

Remote sensing image feature matching via graph classification with local motion consistency

Deep Graphical Feature Learning for the Feature Matching Problem

Learning to Match Features with Discriminative Sparse Graphneuralnetwork

Layered Graph Matching with Composite Cluster Sampling

Adaptive Assignment for Geometry Aware Local Feature Matching

Improving Feature-based Visual Localization by Geometry-Aided Matching

Robust Feature Matching Via Graph Neighborhood Motion Consensus

Density-Guided Locality Consensus for Fast Feature Matching

Layered Graph Matching by Composite Cluster Sampling with Collaborative and Competitive Interactions

Learning for Feature Matching Via Graph Context Attention

A New Method Based on Spatial Relations for Feature Matching

Feature Matching of Remote-Sensing Images Based on Bilateral Local-Global Structure Consistency.

Feature Matching for Remote-Sensing Image Registration via Neighborhood Topological and Affine Consistency

Robust Feature Matching Via Advanced Neighborhood Topology Consensus