Autonomous Driving LiDAR & Image Multi-Modal Annotation Project
This project provides multi-modal annotation services for autonomous driving perception. Tasks include 2D bounding box, 3D cuboid, lane polyline and semantic segmentation for LiDAR point cloud and camera image data. Scenarios cover urban roads, highways and mountain roads. We adopt a three-stage quality control process: annotation, preliminary review and final audit to ensure labeling accuracy above 97%. We have delivered over 500,000 annotated frames.