A Detector-oblivious Multi-arm Network for Keypoint Matching

TL;DR

This paper introduces a Multi-Arm Network that enhances keypoint matching robustness across different detectors with minimal computational overhead, outperforming existing methods in diverse datasets.

Contribution

The proposed MAN framework learns region overlap and depth, enabling detector-oblivious keypoint matching without re-training for different detectors.

Findings

Outperforms state-of-the-art methods on outdoor datasets

Works seamlessly with various keypoint detectors without re-training

Improves robustness of keypoint matching with minimal inference cost

Abstract

This paper presents a matching network to establish point correspondence between images. We propose a Multi-Arm Network (MAN) to learn region overlap and depth, which can greatly improve the keypoint matching robustness while bringing little computational cost during the inference stage. Another design that makes this framework different from many existing learning based pipelines that require re-training when a different keypoint detector is adopted, our network can directly work with different keypoint detectors without such a time-consuming re-training process. Comprehensive experiments conducted on outdoor and indoor datasets demonstrated that our proposed MAN outperforms state-of-the-art methods.