Spatio-Semantic ConvNet-Based Visual Place Recognition
This work addresses the problem of accurate place recognition for robotics and autonomous systems, though it is incremental as it builds on existing image retrieval pipelines.
The authors tackled visual place recognition by using a two-stage pipeline with VGG16 CNN features for encoding semantic and spatial information, achieving over twofold improvement on challenging datasets and outperforming state-of-the-art methods on five benchmarks.
We present a Visual Place Recognition system that follows the two-stage format common to image retrieval pipelines. The system encodes images of places by employing the activations of different layers of a pre-trained, off-the-shelf, VGG16 Convolutional Neural Network (CNN) architecture. In the first stage of our method and given a query image of a place, a number of top candidate images is retrieved from a previously stored database of places. In the second stage, we propose an exhaustive comparison of the query image against these candidates by encoding semantic and spatial information in the form of CNN features. Results from our approach outperform by a large margin state-of-the-art visual place recognition methods on five of the most commonly used benchmark datasets. The performance gain is especially remarkable on the most challenging datasets, with more than a twofold recognition improvement with respect to the latest published work.