The talk will focus on visual geo-localization, the task of estimating the
position depicted within a given photo.
The task will be introduced through practical examples of how geo-localization
is used in robotics, medicine, disaster management and climate change research,
to better understand its goals and challenges.
First we will have an overview of geo-localization to see the multiple methods
available to solve the problem, depending on the goals and requirements.
Then we will concentrate on recent methods for visual geo-localization through
image retrieval, see how they helped overcome long-standing problems, and how
they can be used to geolocalize challenging images within large cities, such as
frames from famous movies.
Finally, we will see how current techniques can be transferred to the domain of
astronaut photography localization, and how NASA is using them to localize
photos from the International Space Station.

BIO
Gabriele Berton is a Ph.D. student in Computer Vision at the Politecnico di
Torino, as a member of the Visual Learning and Multimodal Applications
Laboratory (VANDAL), supervised by Prof. Barbara Caputo and Prof. Carlo Masone.
Previously, he was a research scientist at the Italian Institute of Technology
and an Intern Applied Scientist at Amazon Berlin.
His current research is focused on Visual Geo-localization, Visual Place
Recognition and large scale Image Retrieval.