Shuhong Liu
Papers
1
Total Citations
4
H-Index
1
About
Shuhong Liu is making significant strides in the field of 3D computer vision and robotic perception, with a primary focus on advancing simultaneous localization and mapping (SLAM) systems. Their most notable contribution, "DenseSplat: Densifying Gaussian Splatting SLAM With Neural Radiance Prior" (2025), addresses a critical limitation in Gaussian-based SLAM: the impractical reliance on extensive keyframes for real-world robotic deployment. By introducing a neural radiance prior to densify sparse-view inputs, Liu’s work enables high-fidelity reconstruction and real-time rendering under the constrained conditions typical of actual robotic systems. This innovation bridges the gap between the superior rendering quality of Gaussian splatting and the operational demands of sparse-view environments, earning 4 citations since its publication. Liu’s research directly tackles the trade-off between reconstruction accuracy and practical deployability, offering a pathway for more robust robotic navigation and mapping. Their work is particularly relevant for students and researchers exploring efficient SLAM solutions, as it demonstrates how neural priors can compensate for data sparsity without sacrificing performance.
Research Focus
Key Achievements
Top Papers
- 1