Junchen Liu 刘俊辰

I'm a research intern of BAIR in UC Berkeley, working closely with Ruilong Li under the guidance of Prof. Angjoo Kanazawa. Previously, I received my B.Eng. degree at Beihang University, supervised by Prof. Miao Wang. I've been interned at Institute for AI Industry Research (AIR), Tsinghua University supervised by Prof. Hao Zhao.

My research interest lies in general 3D representation for reconstruction and generation. I believe the real world can be modeled as a probability distribution of clusters on a high-dimensional manifold, suggesting the existence of a general 3D representation that is both compact and non-scene-specific.

Email  /  CV  /  Scholar  /  Twitter  /  Github

profile photo

Publications

(* indicates joint first authorship)

clean-usnob Rip-NeRF: Anti-aliasing Radiance Fields with Ripmap-Encoded Platonic Solids
Junchen Liu*, Wenbo Hu*, Zhuo Yang*, Jianteng Chen, Guoliang Wang, Xiaoxue Chen, Yantong Cai, Huan-ang Gao, Hao Zhao
SIGGRAPH 2024
project page / arXiv / video / code

Efficiently anti-alias NeRF using anisotropic pre-filtering low rank structure, which achieves higher PSNR within less training time compared with Zip-NeRF in the bounded scenes.
clean-usnob NeRRF: 3D Reconstruction and View Synthesis for Transparent and Specular Objects with Neural Refractive-Reflective Fields
Xiaoxue Chen*, Junchen Liu*, Hao Zhao, Guyue Zhou, Yaqin Zhang
arXiv 2023, under review
arXiv / code

Reconstruct transparent and specular objects using a uniform neural representation.

Projects
clean-usnob Tri-MipRF nerfstudio integration
Junchen Liu
code

Nerfstudio version of Tri-MipRF.

Selected Awards

  • National Scholarship (Top 1%), Ministry of Education of China, 2022
  • National Level First Prize (Top 0.65%), The 13th Contemporary Undergraduate Mathematical Contest in Modeling
  • First Prize, The Chinese Mathematics Competitions

Thanks to Jon Barron for sharing the code of his personal webpage.