SEVAR: a stereo event camera dataset for virtual and augmented reality

Abstract

In this paper, we present a precisely synchronized event-based dataset, designed especially for multi-sensor fusion in SLAM applications, with a particular emphasis on VR and AR scenarios. Alongside setting up commonly used stereo regular cameras and an IMU, we have integrated stereo event cameras. We specialize in recording sequences to imitate real-life scenarios, while adding challenging sequences such as low light and fast motion. Consequently, it is our aspiration that this dataset will serve as a valuable resource for the advancement of research in the domain of event-based multi-sensor fusion algorithms.