Detail of the student project

Topic:Visual localization across Seasons Challenge
Department:Katedra kybernetiky
Supervisor:Assia Benbihi
Announce as:Diplomová práce, Bakalářská práce, Semestrální projekt
Description:[](The visual localization server) hosts an
international localization competition that all researchers use to evaluate their
methods. There, datasets typically depict urban scenes across various
conditions of light and seasons. We are contributing
[](a novel dataset) to this challenge that exhibits unseen difficulties: it depicts images of a lakeshore collected by an autonomous boat over 4 years. Most images show vegetation under different lights, weather, and seasons. To enable fair comparison, we want to propose a baseline evaluation for other researchers. These baselines will consist of existing localization methods, mostly 3D structured-based localization or SLAM-based methods.
Responsible person: Petr Pošík