Weisheng Dong

Papers

Total Citations

H-Index

About

Weisheng Dong is a leading researcher in computer vision and robotics, with a primary focus on 3D scene understanding and autonomous navigation. His major contributions center on advancing 3D scene graph prediction from sequential RGB-D data—a critical capability for long-term robotic operations. Dong pioneered methods to bridge the semantic gap between raw sensor inputs and high-level environmental representations, introducing history-enhanced reasoning frameworks that leverage temporal context to build richer, more accurate scene graphs. His most cited work, "Hyperrectangle Embedding for Debiased 3D Scene Graph Prediction From RGB Sequences" (2025, 41 citations), tackles the challenge of biased predictions by embedding scene elements into hyperrectangular spaces, enabling more robust and unbiased graph construction. This work, along with his "History-Enhanced 3D Scene Graph Reasoning From RGB-D Sequences" (2025, 28 citations), has established him as a key innovator in making 3D scene graphs practical for real-world autonomous systems. Dong’s research directly addresses the fundamental problem of enabling robots to build and maintain rich, dynamic representations of their environments over time—a prerequisite for truly autonomous long-term operation.