Zhiyao Zhang

Papers

Total Citations

H-Index

About

Zhiyao Zhang is a leading researcher at the intersection of computer vision and robotics, with a primary focus on dense simultaneous localization and mapping (SLAM) and neural implicit scene representations. His most cited work, "VPE-SLAM: Neural Implicit Voxel-permutohedral Encoding for SLAM," tackles a critical limitation in applying Neural Radiance Fields (NeRF) to robotic perception. While NeRF enables highly realistic environmental mapping, it frequently suffers from geometric distortions in indoor settings, compromising the accuracy of robot navigation. Zhang’s major contribution lies in developing a novel voxel-permutohedral encoding scheme that corrects these distortions, allowing for more precise and reliable 3D reconstructions. This innovation directly enhances the quality of dense SLAM, providing robots with richer, more geometrically accurate scene maps. Although recently published in 2024 and already garnering 4 citations, the work signals a significant step forward in bridging the gap between photorealistic rendering and robust geometric mapping. Zhang’s research is poised to influence future autonomous systems, from service robots to augmented reality, by making neural implicit representations more practical for real-world spatial understanding.