Xinghan Niu

Papers

Total Citations

H-Index

About

Xinghan Niu is a leading researcher in the intersection of computer vision, robotics, and semantic mapping. His primary contributions lie in advancing visual simultaneous localization and mapping (SLAM) by integrating high-level semantic information directly into the geometric estimation pipeline. Niu’s most influential work, "TextSLAM: Visual SLAM With Semantic Planar Text Features" (2023), has already garnered 38 citations, reflecting its immediate impact on the field. In this seminal paper, he proposed a novel framework that treats text objects not merely as visual landmarks but as texture-rich planar patches with extractable semantic meaning. By tightly coupling geometric priors with on-the-fly semantic updates, Niu demonstrated how everyday textual elements—such as signs, labels, and logos—can be leveraged to improve both mapping accuracy and loop closure robustness. This work bridges the gap between low-level feature tracking and high-level scene understanding, offering a practical path toward more intelligent autonomous systems. Niu’s research is particularly notable for its elegant fusion of geometric and semantic cues, a direction that promises to redefine how robots perceive and navigate human-centric environments.