Presentations · 2026.09.16
Effects of data and knowledge gaps on pedestrian perception
Effect of Data or Knowledge Gaps on AI Model Performance
Overview
The demonstrator compares a baseline with models trained using deliberately introduced data gaps, visualizing coverage alongside perception performance. Its vehicle setup and output-consistency monitoring examine missed/false detections and illustrate how test-data gaps can hide model insufficiencies.
Publication and version information
One-page research poster in the materials for the 16 September 2026 IAA Transportation midterm event.