Visual Geo Localization

In robotics, visual place recognition (VPR) is a continuous process that receives as input a video stream to produce a hypothesis of the robot’s current position within a map of known places. This work proposes a taxonomy of the architectures used to learn sequential descriptors for VPR, highlighting different mechanisms to fuse the information from the individual images. This categorization is supported by a complete benchmark of experimental results that provides evidence of the strengths and weaknesses of these different architectural choices. The analysis is not limited to existing sequential descriptors, but we extend it further to investigate the viability of Transformers instead of CNN backbones. We further propose a new ad-hoc sequence-level aggregator called SeqVLAD, which outperforms prior state of the art on different datasets. The code is available at https://github.com/vandal-vpr/vg-transformers.

We tackle the task of cross-domain visual geo-localization, where the goal is to geo-localize a given query image against a database of geo-tagged images, in the case where the query and the database belong to different visual domains. In particular, at training time, we consider having access to only few unlabeled queries from the target domain. To adapt our deep neural network to the database distribution, we rely on a 2-fold domain adaptation technique, based on a hybrid generative-discriminative approach. To further enhance the architecture, and to ensure robustness across domains, we employ a novel attention layer that can easily be plugged into existing architectures. Through a large number of experiments, we show that this adaptive-attentive approach makes the model robust to large domain shifts, such as unseen cities or weather conditions. Finally, we propose a new large-scale dataset for cross-domain visual geo-localization, called SVOX.

In this paper we address the task of visual place recognition (VPR), where the goal is to retrieve the correct GPS coordinates of a given query image against a huge geotagged gallery. While recent works have shown that building descriptors incorporating semantic and appearance information is beneficial, current state-of-the-art methods opt for a top down definition of the significant semantic content. Here we present the first VPR algorithm that learns robust global embeddings from both visual appearance and semantic content of the data, with the segmentation process being dynamically guided by the recognition of places through a multi-scale attention module. Experiments on various scenarios validate this new approach and demonstrate its performance against state-of-the-art methods. Finally, we propose the first synthetic-world dataset suited for both place recognition and segmentation tasks.

Visual Geo-localization (VG) is the task of estimating the position where a given photo was taken by comparing it with a large database of images of known locations. To investigate how existing techniques would perform on a real-world city-wide VG application, we build San Francisco eXtra Large, a new dataset covering a whole city and providing a wide range of challenging cases, with a size 30x bigger than the previous largest dataset for visual geo-localization. We find that current methods fail to scale to such large datasets, therefore we design a new highly scalable training technique, called CosPlace, which casts the training as a classification problem avoiding the expensive mining needed by the commonly used contrastive learning. We achieve state-of-the-art performance on a wide range of datasets and find that CosPlace is robust to heavy domain changes. Moreover, we show that, compared to the previous state-of-the-art, CosPlace requires roughly 80% less GPU memory at train time, and it achieves better results with 8x smaller descriptors, paving the way for city-wide real-world visual geo-localization.

In this paper, we propose a new open-source benchmarking framework for Visual Geo-localization (VG) that allows to build, train, and test a wide range of commonly used architectures, with the flexibility to change individual components of a geo-localization pipeline. The purpose of this framework is twofold: i) gaining insights into how different components and design choices in a VG pipeline impact the final results, both in terms of performance (recall@N metric) and system requirements (such as execution time and memory consumption); ii) establish a systematic evaluation protocol for comparing different methods. Using the proposed framework, we perform a large suite of experiments which provide criteria for choosing backbone, aggregation and negative mining depending on the use-case and requirements. We also assess the impact of engineering techniques like pre/post-processing, data augmentation and image resizing, showing that better performance can be obtained through somewhat simple procedures: for example, downscaling the images’ resolution to 80% can lead to similar results with a 36% savings in extraction time and dataset storage requirement.

We address the task of cross-domain visual place recognition, where the goal is to geolocalize a given query image against a labeled gallery, in the case where the query and the gallery belong to different visual domains. To achieve this, we focus on building a domain robust deep network by leveraging over an attention mechanism combined with few-shot unsupervised domain adaptation techniques, where we use a small number of unlabeled target domain images to learn about the target distribution. With our method, we are able to outperform the current state of the art while using two orders of magnitude less target domain images. Finally we propose a new large-scale dataset for cross-domain visual place recognition, called SVOX.

In this paper we propose a novel method for image matching based on dense local features and tailored for visual geolocalization. Dense local features matching is robust against changes in illumination and occlusions, but not against viewpoint shifts which are a fundamental aspect of geolocalization. Our method, called GeoWarp, directly embeds invariance to viewpoint shifts in the process of extracting dense features. This is achieved via a trainable module which learns from the data an invariance that is meaningful for the task of recognizing places. We also devise a new self-supervised loss and two new weakly supervised losses to train this module using only unlabeled data and weak labels. GeoWarp is implemented efficiently as a re-ranking method that can be easily embedded into pre-existing visual geolocalization pipelines. Experimental validation on standard geolocalization benchmarks demonstrates that GeoWarp boosts the accuracy of state-of-the-art retrieval architectures.

In recent years visual place recognition (VPR), i.e., the problem of recognizing the location of images, has received considerable attention from multiple research communities, spanning from computer vision to robotics and even machine learning. This interest is fueled on one hand by the relevance that visual place recognition holds for many applications and on the other hand by the unsolved challenge of making these methods perform reliably in different conditions and environments. This paper presents a survey of the state-of-the-art of research on visual place recognition, focusing on how it has been shaped by the recent advances in deep learning. We start discussing the image representations used in this task and how they have evolved from using hand-crafted to deep-learned features. We further review how metric learning techniques are used to get more discriminative representations, as well as techniques for dealing with occlusions, distractors, and shifts in the visual domain of the images. The survey also provides an overview of the specific solutions that have been proposed for applications in robotics and with aerial imagery. Finally the survey provides a summary of datasets that are used in visual place recognition, highlighting their different characteristics.