Cross-View Urban Traffic Dataset: Drone-Supervised Ground Truth for Monocular Bird's-Eye View Localization
Researchers have introduced a new dataset and benchmark designed to improve urban traffic perception by aligning street-level and aerial drone views. This benchmark focuses on two key tasks: matching object tracks across these different viewpoints and predicting a bird's-eye view from monocular street-level imagery using aerial supervision. The dataset aims to advance research in cross-view perception and urban scene understanding, providing standardized evaluation tools and baseline implementations for these challenging tasks. AI
IMPACT Enables more robust urban traffic analysis by improving perception across diverse camera viewpoints.