Cathrin Elich

Max Planck Institute for Intelligent Systems

Papers

Total Citations

H-Index

About

Cathrin Elich is a researcher advancing the frontier of robotic perception through object-level scene understanding. Her work centers on enabling intelligent robots to reason about their environment not just through pixels or keypoints, but through meaningful relational structures among objects. Her most-cited paper, "Learning-based Relational Object Matching Across Views" (2023, 4 citations), tackles the fundamental challenge of matching objects across different camera views by learning their spatial and semantic relationships. This contribution is pivotal for tasks like scene reconstruction, image retrieval, and place recognition, where object-level reasoning offers richer, more robust cues than traditional keypoint-based methods. By bridging the gap between low-level features and high-level task reasoning, Elich’s research empowers robots to understand scenes as collections of interacting objects—a crucial step toward autonomous systems that can manipulate, navigate, and collaborate in complex, unstructured environments. Her work is laying the groundwork for more intuitive human-robot interaction and smarter, context-aware AI.