Online Visual Place Recognition via Saliency Re-identification

TL;DR

This paper introduces a fast, saliency-based approach for visual place recognition that mimics human attention, performing saliency detection and re-identification in the frequency domain for improved efficiency.

Contribution

It proposes a novel saliency re-identification method in the frequency domain, achieving competitive accuracy with significantly higher speed compared to existing feature-matching methods.

Findings

Achieves competitive accuracy in place recognition.

Runs much faster than state-of-the-art feature-based methods.

Operates efficiently in resource-constrained robotic applications.

Abstract

As an essential component of visual simultaneous localization and mapping (SLAM), place recognition is crucial for robot navigation and autonomous driving. Existing methods often formulate visual place recognition as feature matching, which is computationally expensive for many robotic applications with limited computing power, e.g., autonomous driving and cleaning robot. Inspired by the fact that human beings always recognize a place by remembering salient regions or landmarks that are more attractive or interesting than others, we formulate visual place recognition as saliency re-identification. In the meanwhile, we propose to perform both saliency detection and re-identification in frequency domain, in which all operations become element-wise. The experiments show that our proposed method achieves competitive accuracy and much higher speed than the state-of-the-art feature-based methods. The proposed method is open-sourced and available at https://github.com/wh200720041/SRLCD.git.