Abstract:Accurately matching local features between a pair of images is a challenging computer vision task. Previous studies typically use attention based graph neural networks (GNNs) with fully-connected graphs over keypoints within/across images for visual and geometric information reasoning. However, in the context of feature matching, considerable keypoints are non-repeatable due to occlusion and failure of the detector, and thus irrelevant for message passing. The connectivity with non-repeatable keypoints not only introduces redundancy, resulting in limited efficiency, but also interferes with the representation aggregation process, leading to limited accuracy. Targeting towards high accuracy and efficiency, we propose MaKeGNN, a sparse attention-based GNN architecture which bypasses non-repeatable keypoints and leverages matchable ones to guide compact and meaningful message passing. More specifically, our Bilateral Context-Aware Sampling Module first dynamically samples two small sets of well-distributed keypoints with high matchability scores from the image pair. Then, our Matchable Keypoint-Assisted Context Aggregation Module regards sampled informative keypoints as message bottlenecks and thus constrains each keypoint only to retrieve favorable contextual information from intra- and inter- matchable keypoints, evading the interference of irrelevant and redundant connectivity with non-repeatable ones. Furthermore, considering the potential noise in initial keypoints and sampled matchable ones, the MKACA module adopts a matchability-guided attentional aggregation operation for purer data-dependent context propagation. By these means, we achieve the state-of-the-art performance on relative camera estimation, fundamental matrix estimation, and visual localization, while significantly reducing computational and memory complexity compared to typical attentional GNNs.

Edge Feature Based Learning Complete Graph Matching Model

GNCCP Learning Graph Matching

GLMNet: Graph Learning-Matching Networks for Feature Matching

Partial Gromov-Wasserstein Learning for Partial Graph Matching

Learning Graph Matching with GNCCP

Deep Reinforcement Learning of Graph Matching.

Learning Combinatorial Embedding Networks for Deep Graph Matching

Deep Graphical Feature Learning for the Feature Matching Problem

Combinatorial Learning of Robust Deep Graph Matching: an Embedding Based Approach.

Learning Bipartite Graph Matching for Robust Visual Localization.

Learnable Graph Matching: A Practical Paradigm for Data Association

Adaptive Edge Attention for Graph Matching with Outliers.

Deep Latent Graph Matching.

GLMNet: Graph Learning-Matching Convolutional Networks for Feature Matching.

Learnable Graph Matching: Incorporating Graph Partitioning with Deep Feature Learning for Multiple Object Tracking

Contrastive Learning Network for Unsupervised Graph Matching

An Edge-Based Matching Kernel for Graphs Through the Directed Line Graphs

Learning for Feature Matching Via Graph Context Attention

Learning Combinatorial Solver for Graph Matching

Learning Deep Graph Matching with Channel-Independent Embedding and Hungarian Attention.

Learning Feature Matching via Matchable Keypoint-Assisted Graph Neural Network