cv
Contact Information
| Name | Zhaorong Wang |
| zhaorong.wang1997@gmail.com | |
| Phone | +81-80-6380-8804 |
| Website | https://larsph.github.io |
Experience
-
2024 - present Tsukuba, Japan
Doctoral Research, Computer Graphics and Vision
University of Tsukuba, CGG Lab
Research digital human editing with 3D Gaussian Splatting, focusing on outfit personalization, cross-avatar editing, and animatable avatar control.
- Design and implement a multi-stage 3D avatar editing system in PyTorch, combining mesh-based geometric optimization, self-supervised data generation, and diffusion-based refinement for high-fidelity free-viewpoint editing.
- Prototype feed-forward, animatable, and editable Gaussian avatars for body-shape control and outfit personalization.
- Build supporting tooling for GPU-cluster experimentation and data processing on Linux servers, improving reproducibility and day-to-day research workflow.
-
2022 - 2024 Tsukuba, Japan
Master's Research, Computer Graphics and Vision
University of Tsukuba, CGG Lab
Researched sparse-view digital human reconstruction and rendering, with a focus on generalizable human NeRF models that balance rendering quality and inference efficiency.
- Designed and implemented a human-prior-guided rendering pipeline in PyTorch, integrating geometry-aware guidance and occlusion-aware refinement for high-quality novel-view synthesis from only 3 to 4 input images.
-
2025 - present Japan
Rendering Research Contributor
Industry-Academia Research Collaboration
Collaborative research project with an industry partner.
- Contributed to a product-oriented face rendering and inverse rendering research project, focusing on practical modeling and analysis for facial geometry and appearance.
Education
-
2024 - 2027 Tsukuba, Japan
Ph.D.
University of Tsukuba
Computer Science
-
2022 - 2024 Tsukuba, Japan
M.S.
University of Tsukuba
Computer Science
-
2018 - 2019 Ottawa, Canada
Non-degree exchange (with transcript)
Carleton University
Computer Science
-
2016 - 2020 Wuxi, China
B.S.
Jiangnan University
Information and Computing Science
Publications
-
2026 -
2026 EG-HumanNeRF: Efficient Generalizable Human NeRF Utilizing Human Prior for Sparse View
Computational Visual Media
Skills
Programming: Python, C++, Shell scripting
ML / Graphics: PyTorch, PyTorch3D, 3D Gaussian Splatting, NeRF, Differentiable rendering, Digital human reconstruction/editing pipelines
Tools: Git, Linux, LaTeX, Experiment management (W&B, Hydra), AI-assisted prototyping/debugging
Languages
English : TOEFL iBT 106
Japanese : JLPT N1
Chinese : Native
Awards and Activities
Scholarship: MEXT Scholarship, Apr. 2024 - Mar. 2027 (Japan)
Presentations: CVPR 2026 Findings poster; Visual Computing 2025 poster (Japan); The 192nd Computer Graphics and Visual Informatics Research Symposium (Japan)