Convolutional Neural Network-based Place Recognition
This work addresses place recognition for robotics or autonomous systems, presenting a novel application of CNNs in this domain.
The paper tackles place recognition by introducing a CNN-based technique combined with spatial and sequential filtering, achieving a 75% increase in recall at 100% precision on a 70 km benchmark dataset and outperforming previous state-of-the-art methods.
Recently Convolutional Neural Networks (CNNs) have been shown to achieve state-of-the-art performance on various classification tasks. In this paper, we present for the first time a place recognition technique based on CNN models, by combining the powerful features learnt by CNNs with a spatial and sequential filter. Applying the system to a 70 km benchmark place recognition dataset we achieve a 75% increase in recall at 100% precision, significantly outperforming all previous state of the art techniques. We also conduct a comprehensive performance comparison of the utility of features from all 21 layers for place recognition, both for the benchmark dataset and for a second dataset with more significant viewpoint changes.