Presentations · 2026.09.16
Closing perception-data gaps through monitoring and generation
Closed-Loop Data Gap Mitigation for Safe AI Perception
Overview
The poster describes selecting frames and objects with ASAM OpenLABEL and JSON-configured requirements to examine ODD coverage gaps. Generated weather/lighting variations feed an iterative process whose usefulness is assessed through comparative training with synthetic and real data.
Publication and version information
One-page research poster in the materials for the 16 September 2026 IAA Transportation midterm event.