Miriam Ackermann

Papers

Total Citations

H-Index

About

Miriam Ackermann is a leading researcher in computer vision and autonomous systems, with a primary focus on anomaly detection and robust perception for self-driving cars and robotics. Her seminal work, "OoDIS: Anomaly Instance Segmentation and Detection Benchmark," introduces a critical benchmark for evaluating how well perception systems identify and segment unknown objects—such as wild animals or untypical debris—that fall outside their training data. This contribution directly addresses a fundamental safety challenge: ensuring that autonomous vehicles can reliably navigate unpredictable real-world environments. Though her benchmark paper has garnered 2 citations since its 2025 publication, its foundational nature positions it as a key resource for future research in open-world perception. Ackermann’s work bridges the gap between controlled training datasets and the messy reality of deployment, offering both a rigorous evaluation framework and a call to action for safer AI systems. Her research is essential reading for students and engineers working on anomaly detection, instance segmentation, and the practical deployment of autonomous agents in dynamic, unconstrained settings.