Learning Sequence Descriptor based on Spatio-Temporal Attention for Visual Place Recognition

TL;DR

This paper introduces a novel spatio-temporal attention-based sequence descriptor for visual place recognition, improving robustness by capturing intrinsic dynamics in frame sequences and outperforming existing methods on benchmark datasets.

Contribution

It proposes a new sequence descriptor that integrates spatial and temporal attention with relative positional encoding for enhanced VPR performance.

Findings

Outperforms recent state-of-the-art methods on benchmark datasets

Effectively captures spatio-temporal dynamics in frame sequences

Utilizes a sliding window and relative positional encoding for sequence modeling

Abstract

Visual Place Recognition (VPR) aims to retrieve frames from a geotagged database that are located at the same place as the query frame. To improve the robustness of VPR in perceptually aliasing scenarios, sequence-based VPR methods are proposed. These methods are either based on matching between frame sequences or extracting sequence descriptors for direct retrieval. However, the former is usually based on the assumption of constant velocity, which is difficult to hold in practice, and is computationally expensive and subject to sequence length. Although the latter overcomes these problems, existing sequence descriptors are constructed by aggregating features of multiple frames only, without interaction on temporal information, and thus cannot obtain descriptors with spatio-temporal discrimination.In this paper, we propose a sequence descriptor that effectively incorporates spatio-temporal information. Specifically, spatial attention within the same frame is utilized to learn spatial feature patterns, while attention in corresponding local regions of different frames is utilized to learn the persistence or change of features over time. We use a sliding window to control the temporal range of attention and use relative positional encoding to construct sequential relationships between different features. This allows our descriptors to capture the intrinsic dynamics in a sequence of frames.Comprehensive experiments on challenging benchmark datasets show that the proposed approach outperforms recent state-of-the-art methods.The code is available at https://github.com/tiev-tongji/Spatio-Temporal-SeqVPR.