CrossLocate: Cross-Modal Large-Scale Visual Geo-Localization in Natural Environments using Rendered Modalities

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F22%3APU143327" target="_blank" >RIV/00216305:26230/22:PU143327 - isvavai.cz</a>

Výsledek na webu

<a href="https://openaccess.thecvf.com/content/WACV2022/html/Tomesek_CrossLocate_Cross-Modal_Large-Scale_Visual_Geo-Localization_in_Natural_Environments_Using_Rendered_WACV_2022_paper.html" target="_blank" >https://openaccess.thecvf.com/content/WACV2022/html/Tomesek_CrossLocate_Cross-Modal_Large-Scale_Visual_Geo-Localization_in_Natural_Environments_Using_Rendered_WACV_2022_paper.html</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/WACV51458.2022.00225" target="_blank" >10.1109/WACV51458.2022.00225</a>

Jazyk výsledku

angličtina

Název v původním jazyce

CrossLocate: Cross-Modal Large-Scale Visual Geo-Localization in Natural Environments using Rendered Modalities

Popis výsledku v původním jazyce

We propose a novel approach to visual geo-localization in natural environments. This is a challenging problem due to vast localization areas, the variable appearance of outdoor environments and the scarcity of available data. In order to make the research of new approaches possible, we first create two databases containing "synthetic" images of various modalities. These image modalities are rendered from a 3D terrain model and include semantic segmentations, silhouette maps and depth maps. By combining the rendered database views with existing datasets of photographs (used as "queries" to be localized), we create a unique benchmark for visual geo-localization in natural environments, which contains correspondences between query photographs and rendered database imagery. The distinct ability to match photographs to synthetically rendered databases defines our task as "cross-modal". On top of this benchmark, we provide thorough ablation studies analysing the localization potential of the database image modalities. We reveal the depth information as the best choice for outdoor localization. Finally, based on our observations, we carefully develop a fully-automatic method for large-scale cross-modal localization using image retrieval. We demonstrate its localization performance outdoors in the entire state of Switzerland. Our method reveals a large gap between operating within a single image domain (e.g. photographs) and working across domains (e.g. photographs matched to rendered images), as gained knowledge is not transferable between the two. Moreover, we show that modern localization methods fail when applied to such a cross-modal task and that our method achieves significantly better results than state-of-the-art approaches. The datasets, code and trained models are available on the project website: http://cphoto.fit.vutbr.cz/crosslocate/.

Název v anglickém jazyce

CrossLocate: Cross-Modal Large-Scale Visual Geo-Localization in Natural Environments using Rendered Modalities

Popis výsledku anglicky

We propose a novel approach to visual geo-localization in natural environments. This is a challenging problem due to vast localization areas, the variable appearance of outdoor environments and the scarcity of available data. In order to make the research of new approaches possible, we first create two databases containing "synthetic" images of various modalities. These image modalities are rendered from a 3D terrain model and include semantic segmentations, silhouette maps and depth maps. By combining the rendered database views with existing datasets of photographs (used as "queries" to be localized), we create a unique benchmark for visual geo-localization in natural environments, which contains correspondences between query photographs and rendered database imagery. The distinct ability to match photographs to synthetically rendered databases defines our task as "cross-modal". On top of this benchmark, we provide thorough ablation studies analysing the localization potential of the database image modalities. We reveal the depth information as the best choice for outdoor localization. Finally, based on our observations, we carefully develop a fully-automatic method for large-scale cross-modal localization using image retrieval. We demonstrate its localization performance outdoors in the entire state of Switzerland. Our method reveals a large gap between operating within a single image domain (e.g. photographs) and working across domains (e.g. photographs matched to rendered images), as gained knowledge is not transferable between the two. Moreover, we show that modern localization methods fail when applied to such a cross-modal task and that our method achieves significantly better results than state-of-the-art approaches. The datasets, code and trained models are available on the project website: http://cphoto.fit.vutbr.cz/crosslocate/.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/LTAIZ19004" target="_blank" >LTAIZ19004: Topografická analýza obrazu s využitím metod hlubokého učení</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)

ISBN

978-1-6654-0477-8

ISSN

—

e-ISSN

—

Počet stran výsledku

10

Strana od-do

3174-3183

Název nakladatele

Institute of Electrical and Electronics Engineers

Místo vydání

Waikoloa

Místo konání akce

Waikoloa, Hawaii

Datum konání akce

4. 1. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—