Xufan He is a master student at the School of Mathematics and Statistics, Nanjing University of Science and Technology, supervised by Prof. Dong Du. His research interests lie in 3D Vision, particularly 3D reconstruction and 3D generation.
Education
-
Nanjing University of Science and TechnologyM.S. in Mathematics (Expected)Sep. 2025 - present -
Nanjing University of Science and TechnologyB.S. in MathematicsSep. 2021 - Jun. 2025
Experience
-
ByteDanceResearch Intern @ Game AIGC
Supervised by Yushuang WuMay 2025 - present
Selected Publications (view all )
UniPart: Part-Level 3D Generation with Unified 3D Geom-Seg Latents
Xufan He*, Yushuang Wu*, Xiaoyang Guo, Chongjie Ye, Jiaqing Zhou, Tianlei Hu, Xiaoguang Han, Dong Du# (* equal contribution, # corresponding author)
CVPR 2026 CCF A
UniPart proposes a unified geometry-segmentation latent representation that jointly encodes object geometry and part-level structure, enabling high-fidelity part-level 3D generation.
UniPart: Part-Level 3D Generation with Unified 3D Geom-Seg Latents
Xufan He*, Yushuang Wu*, Xiaoyang Guo, Chongjie Ye, Jiaqing Zhou, Tianlei Hu, Xiaoguang Han, Dong Du# (* equal contribution, # corresponding author)
CVPR 2026 CCF A
UniPart proposes a unified geometry-segmentation latent representation that jointly encodes object geometry and part-level structure, enabling high-fidelity part-level 3D generation.
NGR: Neural Gradient Rendering for High-Quality 3D Reconstruction from Multi-View Images
Xufan He, Dong Du#, Yushuang Wu, Yunbi Liu (# corresponding author)
CVM 2026 CCF C
NGR leverages gradient information for volume rendering, enabling more accurate surface reconstruction from mutil-view images.
NGR: Neural Gradient Rendering for High-Quality 3D Reconstruction from Multi-View Images
Xufan He, Dong Du#, Yushuang Wu, Yunbi Liu (# corresponding author)
CVM 2026 CCF C
NGR leverages gradient information for volume rendering, enabling more accurate surface reconstruction from mutil-view images.