PedSynth++ Demo Subset: One Clip per Weather Condition

This dataset contains a representative subset of the PedSynth++ dataset generated with the ARCANE-PedSynth framework, a CARLA-based synthetic data generation pipeline for pedestrian crossing prediction in autonomous driving. The subset was prepared as supplementary material for the associated paper and contains 12 video clips in total: one video from each weather/category condition. The included categories are: clear_noon, clear_sunset, cloudy_noon, dawn, foggy_noon, for_paper, heavy_rain_noon, night_clear, night_foggywet_noon, night_rainy, rainy_sunset, and soft_rain_noon. Each video clip contains synchronized multi-modal sensor data, including RGB image frames (.png), DVS/event-camera data (.npz), LiDAR point clouds (.bin), and accompanying metadata/annotation files. The dataset also includes the combined annotation file ALL_WEATHER_combined_labels.csv. The data are intended to support research on pedestrian crossing prediction, multi-modal perception, synthetic-to-real transfer, and autonomous-driving safety scenarios. The full PedSynth++ dataset described in the associated paper contains 533 multi-pedestrian clips across 12 weather conditions with RGB, LiDAR, and DVS streams.