Lingdong Kong

Papers

Total Citations

H-Index

About

Lingdong Kong is an emerging researcher at the forefront of 3D scene understanding and embodied AI, with work spanning 3D visual grounding, augmented reality, and robotics perception. His research tackles one of the most pressing challenges in computer vision: enabling machines to interpret and interact with three-dimensional environments using natural language, without being constrained by expensive annotated datasets or rigid object taxonomies. His most notable recent contribution, "SeeGround" (2025), addresses zero-shot open-vocabulary 3D visual grounding — a technically demanding problem where systems must localize objects in 3D scenes from textual descriptions alone, even for unseen categories. By decoupling the dependency on annotated 3D training data and predefined object classes, Kong's approach represents a significant step toward scalable, adaptable perception systems critical for real-world robotics and augmented reality deployment. Already accumulating citations early in his career, Kong demonstrates a sharp focus on bridging the gap between language understanding and spatial reasoning in complex 3D environments. His work is particularly relevant to researchers and students interested in multimodal learning, open-world perception, and the next generation of intelligent embodied agents capable of understanding their surroundings through human-like description.