Jiewen Chan

I graduated with an M.S. degree in Computer Science from National Yang Ming Chiao Tung University, where I was advised by Prof. Yu-Lun Liu. My research focuses on computer vision, 3D scene representation, dynamic scene reconstruction, and video editing.

I previously graduated with a B.S. degree in Computer Science and Information Engineering from National Taiwan Normal University.

Email / GitHub

Research

My research interests include computer vision, 3D scene representation, dynamic scene reconstruction, video editing, and deep learning.

Jiewen Chan

Awards

IPPR Master's and Doctoral Thesis Award — Honorable Mention
AdaGaR: Adaptive Gabor Representation for Dynamic Scene Reconstruction
2026

Publications

2026
AdaGaR teaser
AdaGaR: Adaptive Gabor Representation for Dynamic Scene Reconstruction
Jiewen Chan, Zhenjun Zhao, Yu-Lun Liu
CVPR 2026 Findings
A dynamic scene representation using adaptive Gabor functions and temporal regularization to model complex motion and improve reconstruction quality over time.
[Paper] [Project]
Voxify3D teaser
Voxify3D: Pixel Art Meets Volumetric Rendering
Yi-Chuan Huang, Jiewen Chan, Hao-Jen Chien, Yu-Lun Liu
CVPR 2026
A 3D representation and rendering framework that transforms 3D scenes into stylized voxel art while preserving geometric structure and visual appearance.
[Paper] [Project]
Skyfall-GS: Synthesizing Immersive 3D Urban Scenes from Satellite Imagery
Jie-Ying Lee, Yi-Ruei Liu, Shr-Ruei Tsai, Wei-Cheng Chang, Chung-Ho Wu, Jiewen Chan, Zhenjun Zhao, Chieh Hubert Lin, Yu-Lun Liu
ECCV 2026
A method for synthesizing immersive large-scale 3D urban scenes from satellite imagery, enabling scene reconstruction from sparse overhead observations.
[Paper] [Project] [Code]
2024
NaRCan teaser
NaRCan: Natural Refined Canonical Image with Integration of Diffusion Prior for Video Editing
Ting-Hsuan Chen, Jiewen Chan, Hau-Shiang Shiu, Shih-Han Yen, Chang-Han Yeh, Yu-Lun Liu
NeurIPS 2024
A video editing framework that learns a refined canonical image and incorporates diffusion priors to produce natural and temporally consistent editing results.
[Paper] [Project]