Kavisha Vidanapathirana

Papers

Total Citations

101

H-Index

About

Kavisha Vidanapathirana is a leading researcher in robotics and computer vision, specializing in 3D place recognition and LiDAR-based localization for autonomous systems. Her work addresses a critical challenge in Simultaneous Localization and Mapping (SLAM): enabling robots to reliably re-localize within large-scale environments using only 3D point cloud data. Her most influential contribution, **LoGG3D-Net** (97 citations), introduces a locally guided global descriptor learning framework that significantly improves retrieval-based place recognition accuracy, directly impacting the robustness of long-term autonomous navigation. She also developed **Locus**, a novel method employing spatiotemporal higher-order pooling to enhance LiDAR-based place recognition in complex, large-scale settings. Vidanapathirana’s research bridges the gap between efficient global data association and precise localization, with her methods being foundational for modern SLAM systems. Her work is highly cited and widely adopted by both academic labs and industry teams working on self-driving cars, mobile robotics, and mapping. Through her innovative descriptor learning and pooling techniques, she continues to push the boundaries of how robots understand and navigate their environments, making her a key figure in the advancement of 3D perception and autonomous navigation.