Yuuto Chokushi

Papers

Total Citations

H-Index

About

Yuuto Chokushi is a researcher whose work sits at the intersection of computer vision and robotics, with a particular focus on enabling robots to navigate and understand their environments over the long term. His primary research areas include visual place recognition, long-term visual SLAM (Simultaneous Localization and Mapping), and scene understanding in dynamic environments. Chokushi’s most notable contribution is his work on "Mining Visual Phrases for Visual Robot Localization" (2016), where he proposed a novel discriminative and compact scene descriptor designed to overcome the limitations of traditional bag-of-words approaches. Unlike standard methods that rely on vector-quantized visual libraries, his technique is explicitly built for robustness in semi-dynamic and partially changing environments—a critical challenge for robots operating in real-world settings over extended periods. While his citation count is modest, the conceptual innovation of his work offers a promising direction for more reliable and efficient long-term autonomy. Chokushi’s research is particularly valuable for students and engineers interested in the practical challenges of deploying robots in familiar but ever-changing spaces, such as homes, offices, or warehouses.