ÑÇÖÞÉ«°É

Abstract: Visual SLAM (Simultaneous Localization and Mapping) research is often hindered by fragmented toolchains, inconsistent evaluation protocols, and complex system setups. In this seminar, Dr. Alejandro Fontan will present VSLAM-LAB, a unified and extensible framework designed to streamline the development, benchmarking, and deployment of Visual SLAM systems. VSLAM-LAB simplifies the end-to-end SLAM pipeline—from automated dataset downloading and formatting to standardized experiment execution and trajectory evaluation—using a single command-line interface. The framework supports a wide array of state-of-the-art SLAM methods and datasets, enabling reproducible research and facilitating fair, comprehensive comparisons.

Find VSLAM-LAB here:  

Bio: Alejandro Fontan is a Research Fellow at the QUT Centre for Robotics since 2022, leading research on Visual SLAM systems. His recent work includes AnyFeature-VSLAM, a flexible framework that automates the integration of any feature type into SLAM pipelines, and a novel ground-truth-free tuning method that enables robust SLAM and SfM evaluation without requiring reference trajectories. Previously, he was an Associate Researcher at the German Aerospace Center (DLR), developing visual odometry for aerial robots in planetary exploration. He holds a PhD in System Engineering from the University of Zaragoza in collaboration with DLR, focused on robustifying Visual SLAM using Information Theory.