HCPM: Hierarchical Candidates Pruning for Efficient Detector-Free Matching

TL;DR

HCPM introduces a hierarchical pruning approach for detector-free image matching that significantly improves efficiency by focusing on informative candidates, achieving faster matching without sacrificing accuracy.

Contribution

This paper proposes a novel hierarchical pruning method for detector-free image matching, reducing computational load while maintaining high accuracy, which is a new approach in the field.

Findings

HCPM outperforms existing methods in speed.

HCPM maintains high matching accuracy.

The method reduces the number of candidates needed for matching.

Abstract

Deep learning-based image matching methods play a crucial role in computer vision, yet they often suffer from substantial computational demands. To tackle this challenge, we present HCPM, an efficient and detector-free local feature-matching method that employs hierarchical pruning to optimize the matching pipeline. In contrast to recent detector-free methods that depend on an exhaustive set of coarse-level candidates for matching, HCPM selectively concentrates on a concise subset of informative candidates, resulting in fewer computational candidates and enhanced matching efficiency. The method comprises a self-pruning stage for selecting reliable candidates and an interactive-pruning stage that identifies correlated patches at the coarse level. Our results reveal that HCPM significantly surpasses existing methods in terms of speed while maintaining high accuracy. The source code will be made available upon publication.