VIPeR: Visual Incremental Place Recognition with Adaptive Mining and Continual Learning

TL;DR

VIPeR introduces an adaptive, memory-augmented lifelong learning approach for visual place recognition, significantly improving performance and generalization across unseen environments while mitigating catastrophic forgetting.

Contribution

It proposes a novel adaptive mining strategy and a memory bank inspired by human memory to enhance incremental place recognition in unseen environments.

Findings

VIPeR outperforms baseline and recent methods by up to 13.65% in average performance.

The memory bank effectively prevents catastrophic forgetting.

Adaptive mining balances environment-specific performance and generalizability.

Abstract

Visual place recognition (VPR) is an essential component of many autonomous and augmented/virtual reality systems. It enables the systems to robustly localize themselves in large-scale environments. Existing VPR methods demonstrate attractive performance at the cost of heavy pre-training and limited generalizability. When deployed in unseen environments, these methods exhibit significant performance drops. Targeting this issue, we present VIPeR, a novel approach for visual incremental place recognition with the ability to adapt to new environments while retaining the performance of previous environments. We first introduce an adaptive mining strategy that balances the performance within a single environment and the generalizability across multiple environments. Then, to prevent catastrophic forgetting in lifelong learning, we draw inspiration from human memory systems and design a novel memory bank for our VIPeR. Our memory bank contains a sensory memory, a working memory and a long-term memory, with the first two focusing on the current environment and the last one for all previously visited environments. Additionally, we propose a probabilistic knowledge distillation to explicitly safeguard the previously learned knowledge. We evaluate our proposed VIPeR on three large-scale datasets, namely Oxford Robotcar, Nordland, and TartanAir. For comparison, we first set a baseline performance with naive finetuning. Then, several more recent lifelong learning methods are compared. Our VIPeR achieves better performance in almost all aspects with the biggest improvement of 13.65% in average performance.