Case Study: Bosch achieves high-precision annotated sensor data for training autonomous-driving neural networks with Understand.ai

Bosch - Customer Case Study

Bosch faced the challenge of training neural networks for highly automated driving (multimodal perception from video, radar, and lidar) and needed highly accurate, large‑scale annotated sensor data. To meet this need Bosch partnered with Understand.ai, using their labeling automation service and the web‑based UAI tool to produce precise 3‑D lidar bounding‑box annotations and plausibility checks with camera data.

Understand.ai delivered iterative, web‑ and AI‑based annotation workflows, expert handling of boundary cases (e.g., differentiating cars/vans or vehicles with roof boxes), and continuous feedback loops with Bosch to refine requirements. Understand.ai achieved the agreed quality targets for reference data, enabling Bosch to train and evaluate neural networks effectively for perception tasks; the collaboration improved annotation efficiency and readiness for further campaigns (including planned 360° surround‑view data collection).

Bosch