Long-Term Ensemble Learning of Visual Place Classifiers

TL;DR

This paper proposes a long-term ensemble learning framework for visual place classification that efficiently transfers knowledge across seasons, optimizes retraining schedules, and improves performance through unsupervised workspace partitioning.

Contribution

It introduces a unified framework for retraining scheduling of CNN classifiers and an unsupervised workspace partitioning method to enhance long-term visual place classification.

Findings

Retraining scheduling significantly improves classification performance.

Planned scheduling outperforms random retraining strategies.

Unsupervised workspace partitioning enhances place classification accuracy.

Abstract

This paper addresses the problem of cross-season visual place classification (VPC) from a novel perspective of long-term map learning. Our goal is to enable transfer learning efficiently from one season to the next, at a small constant cost, and without wasting the robot's available long-term-memory by memorizing very large amounts of training data. To realize a good tradeoff between generalization and specialization abilities, we employ an ensemble of convolutional neural network (DCN) classifiers and consider the task of scheduling (when and which classifiers to retrain), given a previous season's DCN classifiers as the sole prior knowledge. We present a unified framework for retraining scheduling and discuss practical implementation strategies. Furthermore, we address the task of partitioning a robot's workspace into places to define place classes in an unsupervised manner, rather than using uniform partitioning, so as to maximize VPC performance. Experiments using the publicly available NCLT dataset revealed that retraining scheduling of a DCN classifier ensemble is crucial and performance is significantly increased by using planned scheduling.