Soohyun Bae

Papers

Total Citations

H-Index

About

Soohyun Bae is a researcher advancing the field of robotics and computer vision, with a primary focus on the robustness and reliability of visual perception systems. Her key research area centers on developing methods for learning and predicting the repeatability of interest points—a critical challenge for applications like visual odometry, SLAM, and long-term autonomous navigation. Bae's major contribution, detailed in her most-cited work "Learning to Predict Repeatability of Interest Points" (2021), addresses the fundamental problem of how interest points change appearance over time due to varying viewpoints and lighting conditions. By enabling systems to anticipate which features will remain stable, her work directly improves the performance of robots operating in dynamic, real-world environments. While her citation count is currently modest, the foundational nature of this research positions it as a building block for future advancements in lifelong robotic perception. Bae's work is particularly notable for tackling the "continuous and indefinite" nature of environmental change, a problem that is both practically urgent and theoretically challenging, making her a promising voice in the quest for more resilient autonomous systems.