Curated datasets for autonomous driving computer vision research and development
Large-scale object detection, segmentation, and captioning dataset with 80 object categories. Widely used for benchmarking computer vision models.
Autonomous driving dataset with stereo cameras, LiDAR, and GPS/IMU data. Focused on urban driving scenarios with precise 3D annotations.
Large-scale dataset with pixel-level annotations of urban street scenes. Includes 30 classes with fine-grained semantic segmentation.
Large-scale lane detection dataset with high-quality lane annotations. Contains highway and urban driving scenarios with various weather conditions.
Comprehensive lane detection dataset with challenging scenarios including crowded roads, shadows, and various lighting conditions.
German Traffic Sign Recognition Benchmark with 43 different traffic sign classes. Includes real-world images with various lighting and weather conditions.
Large-scale traffic sign dataset with US traffic signs. Includes stop signs, speed limits, and various regulatory signs.
Comprehensive dataset combining object detection, lane detection, and traffic sign recognition. Annotated for multi-task learning in autonomous driving applications.
Specialized dataset for challenging weather conditions including rain, snow, fog, and low-light scenarios. Essential for robust autonomous driving systems.
When using these datasets in your research, please cite the original papers and acknowledge the dataset creators: