Learning to Map Vehicles into Bird's Eye View
This addresses scene understanding for autonomous vehicles and driver assistance systems, but is incremental as it builds on existing detection and mapping techniques.
The paper tackles the problem of mapping vehicle detections from a dashboard camera view to a bird's eye occupancy map for road scene awareness, achieving generalization to real-world data using only synthetic training.
Awareness of the road scene is an essential component for both autonomous vehicles and Advances Driver Assistance Systems and is gaining importance both for the academia and car companies. This paper presents a way to learn a semantic-aware transformation which maps detections from a dashboard camera view onto a broader bird's eye occupancy map of the scene. To this end, a huge synthetic dataset featuring 1M couples of frames, taken from both car dashboard and bird's eye view, has been collected and automatically annotated. A deep-network is then trained to warp detections from the first to the second view. We demonstrate the effectiveness of our model against several baselines and observe that is able to generalize on real-world data despite having been trained solely on synthetic ones.