Profile

Short bio

Hengyuan Chang is a researcher at the Japan Advanced Institute of Science and Technology JAIST. He received Ph.D. in information science in Human-Centered AI Laboratory, JAIST, Ishikawa, Japan, supervised by Prof. Haoran Xie. He received his bachelor’s degree in computer science at Donghua University, Shanghai, China, and his master degree in information science at JAIST, Ishikawa, JAPAN. His research lies at the intersection of image&video generation and fluid dynamics.

Research Interests

Image and video generation with hand-drawn sketch guidance

Physics-informed generation

Fluid data analysis and simulation

Selected Publications

DiffSmoke: Two-Stage Sketch-Based Smoke Illustration Design using Diffusion Models

Hengyuan Chang, Tianyu Zhang, Syuhei Sato, Haoran Xie.

IEEE Access, 2025 Paper

Sketch-Guided Stylized Landscape Cinemagraph Synthesis

Hao Jin, Hengyuan Chang, Xiaoxuan Xie, Zhengyang Wang, Xusheng Du, Shaojun Hu, Haoran Xie.

Computers & Graphics, 2026 Paper

Two-Stage Sketch-Based Smoke Illustration Generation Using Stream Function

Hengyuan Chang, Xiaoxuan Xie, Syuhei Sato, Haoran Xie

ACM SIGGRAPH 2025, poster, 2025 Paper

Physics-Aware Fluid Field Generation from User Sketches Using Helmholtz-Hodge Decomposition

Ryuichi Miyauchi, Hengyuan Chang, Tsukasa Fukusato, Kazunori Miyata, Haoran Xie

NICOGRAPH International 2025, full paper, 2025 Paper

Sketch-Guided Flow Field Generation with Diffusion Model

Hengyuan Chang, Yichen Peng, Syuhei Sato, Haoran Xie

International Workshop on Advanced Image Technology 2024 (IWAIT), full paper, full paper, 2024 Paper

Experiences

2025.05 - 2026.02, Research Collaboration, ANIMINS, OLM, Digital, Inc, Japan (Remote).

2018.11 - 2020.02, research student, Tokyo Metropolitan University, Japan.

Educations

Ph.D. in Information Science, JAIST, Japan (2022.04 – 2026.03)

M.S. in Information Science, JAIST, Japan (2020.04 – 2022.03)

B.E. in Computer Science, Donghua University, China (2014.09 – 2018.06)