Attention-based Pyramid Aggregation Network for Visual Place Recognition

TL;DR

This paper introduces APANet, an end-to-end trainable neural network that enhances visual place recognition by encoding multi-scale features and suppressing confusing regions, outperforming existing methods on benchmark datasets.

Contribution

The paper proposes a novel Attention-based Pyramid Aggregation Network with spatial pyramid pooling and attention blocks for improved place recognition, along with a PCA power whitening strategy.

Findings

APANet outperforms state-of-the-art methods on place recognition benchmarks.

The spatial pyramid pooling effectively encodes multi-scale urban features.

The PCA power whitening enhances retrieval performance.

Abstract

Visual place recognition is challenging in the urban environment and is usually viewed as a large scale image retrieval task. The intrinsic challenges in place recognition exist that the confusing objects such as cars and trees frequently occur in the complex urban scene, and buildings with repetitive structures may cause over-counting and the burstiness problem degrading the image representations. To address these problems, we present an Attention-based Pyramid Aggregation Network (APANet), which is trained in an end-to-end manner for place recognition. One main component of APANet, the spatial pyramid pooling, can effectively encode the multi-size buildings containing geo-information. The other one, the attention block, is adopted as a region evaluator for suppressing the confusing regional features while highlighting the discriminative ones. When testing, we further propose a simple yet effective PCA power whitening strategy, which significantly improves the widely used PCA whitening by reasonably limiting the impact of over-counting. Experimental evaluations demonstrate that the proposed APANet outperforms the state-of-the-art methods on two place recognition benchmarks, and generalizes well on standard image retrieval datasets.