Abstract: Local Feature Matching, a pivotal component of numerous computer vision tasks (e.g., structure from motion and visual localization), has been effectively addressed by Transformer-based ...