3D Lidar / 2D Bounding Box
The project involved large-scale 3D LiDAR and 2D image annotation to support AI and automation systems, including autonomous and object-detection models. My responsibilities included accurate point cloud labeling, 3D cuboid placement, object classification, semantic segmentation, and 2D bounding box annotation, as well as data validation and error correction to ensure consistency with client guidelines. The project covered thousands of images and LiDAR frames, requiring efficient dataset management while meeting strict productivity targets and deadlines. Quality measures included multi-level review processes, adherence to detailed annotation guidelines, regular feedback implementation, accuracy checks, and continuous performance monitoring to maintain high precision and consistency across all labeled data.