1 paper
Anil Bayram Gogebakan, Filippo Marostica, Alessio Caviglia +2
Existing automotive event datasets rely on appearance-based annotations from frame pipelines, making them poorly suited for motion-aware event perception. We present a geometry-dri…