Nimet Kaygusuz

Papers

Total Citations

H-Index

About

Nimet Kaygusuz is a researcher advancing the field of visual odometry (VO) for robotics and autonomous systems. Her work addresses a critical challenge: making VO both computationally efficient and reliable in real-world settings. In her most-cited paper, "MDN-VO: Estimating Visual Odometry with Confidence" (2021, 10 citations), Kaygusuz introduces a novel approach that replaces traditional, computationally expensive feature-matching methods with a more efficient framework. Crucially, her work goes beyond simple estimation by incorporating a confidence measure, directly tackling the problem of failure detection—a task previously reliant on heuristic methods. This contribution offers a principled way for autonomous systems to assess the reliability of their own motion estimates, a key step toward safer and more robust navigation. By focusing on both efficiency and uncertainty quantification, Kaygusuz’s research has the potential to impact applications from drone navigation to autonomous driving, where accurate and trustworthy odometry is essential. Her work represents a thoughtful integration of deep learning with practical robotics challenges.