Abstract:Feature matching plays a key role in computer vision. However, due to the limitations of the descriptors, the putative matches are inevitably contaminated by massive outliers. This paper attempts to tackle the outlier filtering problem from two aspects. First, a robust and efficient graph interaction model, is proposed, with the assumption that matches are correlated with each other rather than independently distributed. To this end, we construct a graph based on the local relationships of matches and formulate the outlier filtering task as a binary labeling energy minimization problem, where the pairwise term encodes the interaction between matches. We further show that this formulation can be solved globally by graph cut algorithm. Our new formulation always improves the performance of previous locality-based method without noticeable deterioration in processing time, adding a few milliseconds. Second, to construct a better graph structure, a robust and geometrically meaningful topology-aware relationship is developed to capture the topology relationship between matches. The two components in sum lead to topology interaction matching (TIM), an effective and efficient method for outlier filtering. Extensive experiments on several large and diverse datasets for multiple vision tasks including general feature matching, as well as relative pose estimation, homography and fundamental matrix estimation, loop-closure detection, and multi-modal image matching, demonstrate that our TIM is more competitive than current state-of-the-art methods, in terms of generality, efficiency, and effectiveness. The source code is publicly available at http://github.com/YifanLu2000/TIM.

An Evaluation of Feature Matchers Forfundamental Matrix Estimation.

MatchBench: an Evaluation of Feature Matchers.

Robust and Accurate Fundamental Matrix Estimation with Propagated Matches

Image Matching: An Application-oriented Benchmark

A New Method Based on Spatial Relations for Feature Matching

A Robust Algorithm For Feature Point Matching

Performance Comparison and Analysis of Fundamental Matrix Estimating Methods for Computer Vision Applications

Efficient Image Features Selection and Weighting for Fundamental Matrix Estimation

Image Feature Correspondence Selection: A Comparative Study and a New Contribution

A Comprehensive Performance Evaluation for 3D Transformation Estimation Techniques.

Comparative evaluation of 2D feature correspondence selection algorithms

A Brief Survey of Feature Based Image Matching

The Estimation Algorithm of Epipolar Geometry Based on Sub-Pixel Match Strategy

Mixed Feature Extraction and Matching for Large Affine Scene

Mismatched: Evaluating the Limits of Image Matching Approaches and Benchmarks

Feature Matching via Topology-Aware Graph Interaction Model

Image Matching Via Ranking and Neighborhood Manifold Embedding

Robust Feature Matching Via Support-Line Voting and Affine-Invariant Ratios

An Image Matching Algorithm Based On Singular Value Decomposition and Belief Propagation

Application of Feature Point Detection and Matching in 3D Objects Reconstruction

Evaluation of Image Feature Detection and Matching Algorithms