Xianzu Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xianzu Wu has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
Harry Yang
Division of Arts and Machine Creativity, The Hong Kong University of Science and Technology, Hong Kong, China
1 shared publication- Huan Wang
Westlake University
1 shared publication - Jiahao Chang
SSE, CUHKSZ, Shenzhen, China
1 shared publication - Jinjian Zhao
SSE, The Chinese University of Hong Kong, Shenzhen, Shenzhen, China and VAST, Beijing, China
1 shared publication - Jun Liu
Lancaster University
1 shared publication - Ser-Nam Lim
University of Central Florida, Orlando, USA
1 shared publication
Latest Publications
- Open Access
- 30 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-11
TwinRecon: Unposed Generative Reconstruction via Coupled Shape-Pointmap Diffusion
- 30 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Open AccessTwinRecon: Unposed Generative Reconstruction via Coupled Shape-Pointmap Diffusion
- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (5) , 6866-6876
Niagara : Normal-Integrated Geometric Affine Field for Scene Reconstruction From a Single View
Recent advances in single-view 3D scene reconstruction have highlighted the challenges in capturing fine geometric details and ensuring structural consistency, particularly in high-fidelity outdoor scene modeling. This paper presents Niagara, a new single-view 3D scene reconstruction framework that can faithfully reconstruct challenging outdoor scenes from a single input image for the first time. Our approach integrates monocular depth and normal estimation as input, which substantially improves its ability to capture fine details, mitigating common issues like geometric detail loss and deformation. Additionally, we introduce a geometric affine field (GAF) and 3D self-attention as geometry-constraint, which combines the structural properties of explicit geometry with the adaptability of implicit feature fields, striking a balance between efficient rendering and high-fidelity reconstruction. Our framework finally proposes a specialized encoder-decoder architecture, where a depth-based 3D Gaussian decoder is proposed to predict 3D Gaussian parameters, which can be used for novel view synthesis. Extensive results and analyses suggest that our Niagara surpasses prior SoTA approaches such as Flash3D in both single-view and dual-view settings, significantly enhancing the geometric accuracy and visual fidelity, especially in outdoor scenes. Webpage: https://ai-kunkun.github.io/Niagara page/.
- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (5) , 6866-6876
Niagara : Normal-Integrated Geometric Affine Field for Scene Reconstruction From a Single View
Recent advances in single-view 3D scene reconstruction have highlighted the challenges in capturing fine geometric details and ensuring structural consistency, particularly in high-fidelity outdoor scene modeling. This paper presents Niagara, a new single-view 3D scene reconstruction framework that can faithfully reconstruct challenging outdoor scenes from a single input image for the first time. Our approach integrates monocular depth and normal estimation as input, which substantially improves its ability to capture fine details, mitigating common issues like geometric detail loss and deformation. Additionally, we introduce a geometric affine field (GAF) and 3D self-attention as geometry-constraint, which combines the structural properties of explicit geometry with the adaptability of implicit feature fields, striking a balance between efficient rendering and high-fidelity reconstruction. Our framework finally proposes a specialized encoder-decoder architecture, where a depth-based 3D Gaussian decoder is proposed to predict 3D Gaussian parameters, which can be used for novel view synthesis. Extensive results and analyses suggest that our Niagara surpasses prior SoTA approaches such as Flash3D in both single-view and dual-view settings, significantly enhancing the geometric accuracy and visual fidelity, especially in outdoor scenes. Webpage: https://ai-kunkun.github.io/Niagara page/.