👋 Hi there!
I’m a joint PhD student at the Computer Graphics Lab of ETH Zürich and DisneyResearch|Studios, supervised by Prof. Dr. Markus Gross and Dr. Vinicius Azevedo. I earned my Master’s degree of Computer Science in 2022 from ETH Zürich. Before that, I completed my engineering degree at Ecole polytechnique in Paris, France (equivalent to BSc & MSc). My research interests include machine learning, generative AI, and their intersection with computer graphics and artistic control.
🔥 News
- 2026.06: 🎉 I will be presenteing What Is It Like to Be a Noise? at CVPR 2026 in Denver!
- 2025.12: 🎉 Strands Neural Style Transfer has been published at SIGGRAPH Asia 2025!
- 2025.08: 🎖 LookingGlass has received the DCAJ Award and the Laval Virtual Award at SIGGRAPH 2025 Emerging Technologies!
- 2025.04: 🎉 LookingGlass has been accepted to CVPR 2025 (Oral)!
- 2025.04: 🎉 How I Warped Your Noise is now available on ArXiv!
- 2024.05: 🎉 How I Warped Your Noise was presented at ICLR 2024 in Vienna, Austria!
📝 Publications
Pascal Chang, Kai Lascheit, Jingwei Tang, Markus Gross, Vinicius C. Azevedo
We propose a novel entropy-based regularization method for Gaussian noise in diffusion models.
Shaping Strands with Neural Style Transfer (SIGGRAPH Asia 2025 Journal)
Beyzanur Coban, Pascal Chang, Guilherme G. Haetinger, Jingwei Tang, Vinicius C. Azevedo
We propose the first neural style transfer method for hair and fur strands stylization.
LookingGlass: Generative Anamorphoses via Laplacian Pyramid Warping (CVPR 2025 Oral)
Pascal Chang, Sergio Sancho, Jingwei Tang, Markus Gross, Vinicius C. Azevedo
We repurpose latent diffusion models to generate high-quality, complex, multi-view optical illusions such as anamorphoses.
How I Warped Your Noise: a Temporally-Correlated Noise Prior for Diffusion Models (ICLR 2024 Oral)
Pascal Chang, Jingwei Tang, Markus Gross, Vinicius C. Azevedo
ArXiv / Webpage / OpenReview / Video
We propose a method to warp a Gaussian noise sample while keeping it Gaussian and apply it to diffusion models to help temporal coherency.
Urban Brush: Intuitive and Controllable Urban Layout Editing (UIST 2021)
Xiaochen Zhou*, Pascal Chang*, Marie-Paule Cani, Bedrich Benes
Paper / Webpage / Video (short) / Video (full)
We design intuitive brushes for controllable editing and authoring of procedural urban landscapes.
📖 Education
- 2023.01 - Present, Joint PhD at ETH / Disney Research supervised by Prof. Dr. Markus Gross and Dr. Vinicius C. Azevedo.
- 2020.09 - 2022.11, MSc ETH in Computer Science (Zürich, Switzerland).
- 2017.09 - 2022.03, Engineering Degree from Ecole polytechnique (Paris, France).
💻 Services
Conference Reviewer
- Computer Vision and Pattern Recognition (CVPR)
- International Conference on Learning Representations (ICLR)
- European Association for Computer Graphics (Eurographics)
- ACM SIGGRAPH