Geolocation Estimation of Photos Using a Hierarchical Model and Scene Classification

Müller-Budack, Eric; Pustu-Iren, Kader; Ewerth, Ralph

doi:10.1007/978-3-030-01258-8_35

Cited by 60 publications

(106 citation statements)

References 40 publications

Supporting

Mentioning

106

Contrasting

Order By: Relevance

“…This can be explained not only by the limitation of the training dataset, but also because the discretization used to create the classes is naturally lossy. The performance of the classification-based place recognition can be boosted by combining it with scene recognition [239]. In this approach, a first network classifies the category of the scene and, based on the label, forwards the image to a classifier specialized for that category.…”

Section: Visual Geo-localization As Classificationmentioning

confidence: 99%

A Survey on Deep Visual Place Recognition

2021

View full text Add to dashboard Cite

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. INDEX TERMS Visual place recognition, image representation learning, deep learning.

show abstract

Section: Visual Geo-localization As Classificationmentioning

confidence: 99%

A Survey on Deep Visual Place Recognition

2021

View full text Add to dashboard Cite

show abstract

“…One of the descriptors is taken from a ResNet [21] model pre-trained on the Places365 dataset [57], where the task is to recognize 365 distinct places -examples include beach, stadium, street. The second descriptor is taken from the model [38] based on ResNet101 [21] aimed at predicting the geolocation information of an image. This model is pretrained on a subset of the Yahoo Flickr Creative Commons 100 Million dataset (YFCC100M) [47].…”

Section: Systemmentioning

confidence: 99%

“…Location estimation methods related to this research breakdown into three main groups [10]: natural [4,56], city-scale [14,24] and global [26,48]. Several methods in the published research rely on a single modality [36,38,52] and we focus here on multimodal approaches. MediaEval benchmark placing datasets [15,29] include more than five million instances with images, videos and metadata used to estimate capturing locations represented in multimedia.…”

Section: Related Researchmentioning

confidence: 99%

MLM

Armitage

Kacupaj

Tahmasebzadeh

et al. 2020

Proceedings of the 29th ACM International Conference on Information &Amp; Knowledge Management

Self Cite

View full text Add to dashboard Cite

In this paper, we introduce the MLM (Multiple Languages and Modalities) dataset-a new resource to train and evaluate multitask systems on samples in multiple modalities and three languages. The generation process and inclusion of semantic data provide a resource that further tests the ability for multitask systems to learn relationships between entities. The dataset is designed for researchers and developers who build applications that perform multiple tasks on data encountered on the web and in digital archives. A second version of MLM provides a geo-representative subset of the data with weighted samples for countries of the European Union. We demonstrate the value of the resource in developing novel applications in the digital humanities with a motivating use case and specify a benchmark set of tasks to retrieve modalities and locate entities in the dataset. Evaluation of baseline multitask and single task systems on the full and geo-representative versions of MLM demonstrate the challenges of generalising on diverse data. In addition to the digital humanities, we expect the resource to contribute to research in multimodal representation learning, location estimation, and scene understanding. CCS CONCEPTS • Machine learning; • Multitask learning; • Multimodal data; • Multilingual data; * Denotes equal contribution to this research.

show abstract

“…The described system should use the methods of machine learning and, in particular, deep learning of its neural networks. To solve the problem of classifying an object in an image, it is proposed to use multilayer convolutional neural networks (Müller-Budack, E., Pustu-Iren, K. and Ewerth, R., 2018;Weyand, T., Kostrikov, I. and Philbin, J., 2016;Fedorenko, Y. S. and Gapanyuk, Y. E., 2018), and the training and tuning of the neural network parameters should be carried out taking into account the specifics of the arrays of the processed photos.…”

Section: Scene Classification Modulementioning

confidence: 99%

“…In (Nam, N. Vo, Jacobs, N. and Hays, J., 2017), the Im2GPS project is presented, in which the geolocation estimation based on a comparison of the original photo with 6 million prepared Flickr images for which GPS coordinates are given. Comparative results of the assessment of the geo-location of the projects PlaNet and Im2GPS are described in (Müller-Budack, E., Pustu-Iren, K. and Ewerth, R., 2018;Nam, N. Vo, Jacobs, N. and Hays, J., 2017).…”

Section: Introductionmentioning

confidence: 99%

Application of the Methods of Hybrid Artificial Intelligence to Determine the Location of Objects by Their Photography

Berezkin¹,

Proletarsky²,

Sidorov³

et al. 2019

Proceedings of the 16th International Conference on Applied Computing 2019

View full text Add to dashboard Cite

Successful assessment of the geolocation of photography can be used in many interesting projects, although this is an extremely difficult task. The article is devoted to the development of an approach to the creation of a hybrid intellectual system for determining the location of objects from their photographs. The analysis of existing approaches to the assessment of geolocation of photographs was carried out; a generalized structure of the hybrid intelligent system for determining the location of objects from their photographs was proposed. To assess the location, it is proposed to analyze both the data on the GPS coordinates of the original photo and those similar to it, and the classification of objects in the photo: city, nature. Geolocation is considered as a classification problem when the earth's surface is divided into disjoint adjacent cells.

show abstract

Geolocation Estimation of Photos Using a Hierarchical Model and Scene Classification

Cited by 60 publications

References 40 publications

A Survey on Deep Visual Place Recognition

A Survey on Deep Visual Place Recognition

MLM

Application of the Methods of Hybrid Artificial Intelligence to Determine the Location of Objects by Their Photography

Contact Info

Product

Resources

About