Learning series · October 2026
Video Generation from Scratch
A hands-on path from generative modeling fundamentals to video. The first chapter covers DDPM, DDIM, Flow Matching, and Reflow with code, real MNIST experiments, and sampling animations.
I am a PhD researcher in Applied Mathematics and Computer Science at the Technical University of Denmark (DTU). My research is conducted at the Visual Computing section of DTU Compute, where I am supervised by Prof. Morten Rieger Hannemose, Prof. Anders Nymark Christensen, and Prof. Anders Bjorholm Dahl.
From June to September 2026, I was a Visiting PhD Researcher at the University of Hong Kong (HKU) in Prof. Lequan Yu’s lab, working on Agentic AI for Medical Image Generation.
I also serve as a Part-time Research Scientist at Retinalyze, researching deep learning methods for retinal fundus and OCT image analysis.
Previously, I earned my Master’s degree in Computer Engineering on a full scholarship at the Budapest University of Technology and Economics (BME) and worked as a Research Assistant (June 2019–August 2020) in the Neural Information Processing Group (NIPG) at Eötvös Loránd University (ELTE).
1,155
Citations
7
H-index
6
i10-index
Google Scholar metrics as of 17 October 2026.
Research
I develop efficient and controllable visual generative models, with a focus on diffusion and flow-based modeling. My work spans controllable image generation, adaptive sampling, visual counterfactual explanations, and image editing.
- Efficient Visual Generation: Diffusion Transformers, latent diffusion, and flow-based models for controllable generation with reduced computational cost.
- Visual Counterfactual Explanations: Precise and localized image edits that explain model predictions while preserving unrelated content.
- Medical Image Analysis: Robust visual understanding and segmentation of retinal fundus, OCT, and other medical images.
I am also interested in compact image and video generation, generative model distillation, and predictive world models.
News
- [10/2026]: Started Video Generation from Scratch, a hands-on tutorial series. The first chapter on diffusion and flow matching is now available.
- [10/2026]: My current work focuses on histology image generation. The manuscript is under review at ICLR 2027.
- [10/2026]: CLEAR has been accepted to WACV 2027.
- [10/2026]: Recognized as a Top Reviewer at NeurIPS 2026.
- [09/2026]: MedDiME, our work on efficient latent diffusion for medical counterfactual generation, is available on arXiv.
- [09/2026]: Completed my research visit (June–September 2026) with Prof. Lequan Yu at the University of Hong Kong (HKU), working on Agentic AI for Medical Image Generation.
- [08/2026]: FiRe has been accepted to BMVC 2026.
- [04/2026]: Oral Presentation of SA-UNetv2 at ISBI 2026, London.
- [02/2026]: MaskDiME has been accepted to CVPR 2026!
- [12/2025]: Joined Retinalyze as a Part-time Research Scientist.
- [09/2025]: Oral Presentation for Med-Art at DGM4MICCAI 2025, Daejeon, South Korea.
- [09/2025]: Awarded the Otto Mønsted Travel Grant for MICCAI 2025.
- [08/2024]: Started my PhD studies at the Technical University of Denmark.
Tutorials
Selected Publications
Awards
2026
CVPR Compute Efficiency Award for MaskDiMEOne of just 26 papers out of 4,000+.
2026
Research Visit Grant from Otto Mønsteds Fond
2026
Research Visit Grant from William Demant Fonden
2025
Otto Mønsted Travel Grant for MICCAI 2025
2025
Oxford Machine Learning Summer School Scholarship, England
2024
PhD Student Position at Technical University of Denmark (2024-2027)
2019
EFOP Fellowship (Hungary, EU-funded)
2019
Stipendium Hungaricum Scholarship (Hungary, Government-funded)
2016
First-class Academic Scholarship & National Endeavor Scholarship, China
Academic Service
Reviewer Recognition
Top Reviewer, NeurIPS 2026
Program Committee
Technical Program Committee for ISBI 2026 and 2027; Program Committee for ICONIP 2023
Reviewer
Conferences: ICLR 2027, ISBI 2027, CVPR 2026, NeurIPS 2026, MICCAI 2026, BMVC 2026, MICCAI 2025, ISBI 2025, ICPR 2024
Journals: IEEE Transactions on Medical Imaging, Expert Systems With Applications, Biomedical Signal Processing and Control, IEEE Signal Processing Letters
Teaching Assistant
Introductory and advanced courses on computer vision with deep learning at Technical University of Denmark.
Visitors
All recorded sources since 18 July 2026.
Loading historical visitor locations…