Guangfu Che

Papers

Total Citations

H-Index

About

Guangfu Che is a researcher specializing in visual SLAM (Simultaneous Localization and Mapping) and efficient image retrieval for robotics. His key contributions center on advancing loop closure detection—a critical problem that allows robots to recognize previously visited locations and correct accumulated mapping drift. Che’s most notable work, "Fast and Incremental Loop Closure Detection Using Proximity Graphs" (2019), introduces a novel approach that moves beyond traditional bag-of-words (BoW) models. While BoW methods offer high precision, they often suffer from high computational costs and limited recall. Che’s solution leverages proximity graphs to enable faster, incremental querying, significantly reducing time complexity without sacrificing accuracy. This work, which has garnered over 43 citations, addresses a fundamental bottleneck in real-time SLAM systems, making it highly influential for researchers developing autonomous navigation for drones, self-driving cars, and mobile robots. By tackling the trade-off between speed and reliability, Che has provided a practical framework that enhances the robustness of visual SLAM in dynamic environments. His research continues to impact the fields of computer vision and robotics, offering scalable solutions for long-term autonomous operation.