github/jentuuh linkedin/jentevandersanden jente.vds.pro@gmail.com

About me

I am a third-year Ph.D. candidate at Max Planck Institute for Informatics, supervised by Gurprit Singh and Hans-Peter Seidel. In 2025, I spent a great summer at Adobe Research London, working on context compression for video generation with Yulia Gryaditskaya ! At the end of my Master studies, I also had the privilege of interning at Wenzel Jakob’s Realistic Graphics Lab at EPFL in Switzerland. Before that, I completed my undergraduate studies with a major in Visual Computing at Hasselt University in Belgium.

My research interests include, but are not limited to image and video generation, fundamentals behind generative modeling, physically-based (differentiable) rendering and real-time rendering. My PhD research so far has mainly explored the impact of anisotropy on a generative process, more recently I have been working on improving the subject consistency for long videos generated by diffusion models. I’m really excited about how generative models could reshape the entertainment industry, especially in real-time rendering, and I hope to continue exploring research at that intersection.

Publications

Edge-preserving noise for diffusion models

Jente Vandersanden, Sascha Holl, Xingchang Huang, Gurprit Singh

ICLR 2025 Workshop DeLTa

Project page | Paper

We introduce and analyze a structure-aware generative diffusion process, demonstrating notable improvements in structure-guided generative tasks.

Demystifying noise: The role of randomness in generative AI

Gurprit Singh, Xingchang Huang, Jente Vandersanden, Cengiz Öztireli, Niloy J Mitra

SIGGRAPH 2025 Course + Eurographics 2025 tutorial

Project page

This course offers a thorough exploration of the role of randomness in generative AI, leveraging foundational knowledge from statistical physics, stochastic differential equations, and computer graphics.

A Scalable and Coherent Approach to Monte Carlo Path Tracing for Multi-User Environments

Jente Vandersanden

Hasselt University (MSc Thesis, 2023)

Project page | Paper

An adaptation of the Monte-Carlo path tracing algorithm that makes it more modular and scalable for multi-user cloud-gaming environments.

Shared Rendering Computation in Cloud Gaming

Jente Vandersanden

Hasselt University (BSc Thesis, 2021)

Project page | Paper

A shared rendering system rendering diffuse interactive global illumination for dynamic scenes on a centralized server.

Other research projects

During my time at the Realistic Graphics Lab at EPFL in the summer of 2023, I worked together with the Mitsuba 3 team to implement physically-accurate differentiable rendering of heightfields. This technique can for example be used to reconstruct landscape geometry from satellite images. In addition, an adaptation of the projective Sampling method of [Zhang et al. 2023] was implemented for heightfields. A brief demonstration can be found here.

During the summer of 2022, I was an intern for Prof. dr. Philippe Bekaert's Visual Computing research group at Hasselt University. My main contribution there was the development of a synthetic data generation pipeline that renders photorealistic novel views of an object by querying a neural radiance field. The tool was created with the purpose to generate training data for an object detection model tailored to specular objects.

Extracurricular

When I am not tracing light paths, I like to travel and experience other cultures. The beauty of nature and our planet keeps surprising me (that is another reason why I enjoy simulating it on a computer). I enjoy a few sports, under which Muay Thai and snowboarding. I am also a great enthusiast of (visual) art, whether it be video games, photography or cinematography. One of my aspirations is to one day find and pursue a personal project that unites my technical knowledge with my passion for these arts. Below you can find a random selection of pictures from the past few years.