Shuting Dong

Papers

Total Citations

H-Index

About

Shuting Dong is a researcher advancing the field of visual place recognition (VPR), a critical capability for robot localization, autonomous navigation, and augmented reality. Her work focuses on developing hierarchical VPR methods that balance high accuracy with computational efficiency—a key challenge for real-world deployment. Dong’s most cited paper, "Deep Homography Estimation for Visual Place Recognition" (2024, 17 citations), introduces a novel approach that leverages deep homography estimation to improve spatial understanding in VPR systems. Her subsequent work, "AANet: Aggregation and Alignment Network with Semi-hard Positive Sample Mining for Hierarchical Place Recognition" (2023, 10 citations), further refines this paradigm by proposing a network that aggregates and aligns visual features while employing a semi-hard positive sample mining strategy to enhance retrieval performance. These contributions address the limitations of traditional two-stage VPR pipelines, which first retrieve candidate locations using global features before performing geometric verification. By integrating deep learning into homography estimation and feature alignment, Dong’s research pushes the boundaries of how robots and autonomous systems perceive and localize within their environments, offering more robust and scalable solutions for real-world applications.