Alejandro Sztrajman

I'm a postdoctoral research associate at the University of Cambridge. I hold a PhD in Computer Science from University College London (UCL), where I worked at the Digital Reality Lab, under the supervision of Profs. Tim Weyrich and Tobias Ritschel. My research is in the intersection between machine learning and visual computing, combining methods from neural fields, hypernetworks, physically-based rendering and generative models. My PhD project was funded by a Marie Curie Fellowship, granted by the European Commission. During my PhD I also worked as research intern at Microsoft and Adobe Substance 3D. Before joining UCL, I obtained a BSc in Physics from the University of Buenos Aires, and I was a visiting student at the Columbia Computer Graphics Group.


Contact
Room SE18
Computer Laboratory
William Gates Building
15 JJ Thomson Avenue
University of Cambridge, UK

Publications

FrePolad: Frequency-Rectified Point Latent Diffusion for Point Cloud Generation
Chenliang Zhou, Fangcheng Zhong, Param Hanji, Zhilin Guo, Kyle Thomas Fogarty, Alejandro Sztrajman, Hongyun Gao, Cengiz Oztireli
In Review (2024).
Keywords: Diffusion Models, 3D Point Clouds.

Hypernetworks for Generalizable BRDF Estimation
Fazilet Gozbudak, Alejandro Sztrajman, Chenliang Zhou, Fangcheng Zhong, Rafal Mantiuk, Cengiz Öztireli
In Review (2024).
Keywords: Neural Fields, Hypernetworks, Generalizable Models, Material Appearance.

iHyperTime: Interpretable Time Series Generation with Implicit Neural Representations
Elizabeth Fons, Alejandro Sztrajman, Yousef El-Laham, Andrea Coletta, Alexandros Iosifidis, Svitlana Vyetrenko
In Review (2024).
Keywords: Time-series, Neural Fields, Generative Models, Interpretability, Hypernetworks.

Neural Fields with Hard Constraints of Arbitrary Differential Order
Fangcheng Zhong, Kyle Fogarty, Param Hanji, Tianhao Wu, Alejandro Sztrajman, Andrew Spielberg, Andrea Tagliasacchi, Petra Bosilj, Cengiz Oztireli
NeurIPS (2023).
Keywords: Neural Fields, Hypernetworks, Constrained Learning.

Color Calibration Methods for OLED Displays
Maliha Ashraf, Alejandro Sztrajman, Dounia Hammou, Rafał Mantiuk
Color Imaging (2023).
Keywords: Color Correction, Implicit Neural Representations.

HyperTime: Implicit Neural Representations for Time-Series
Elizabeth Fons, Alejandro Sztrajman, Yousef El-Laham, Alexandros Iosifidis, Svitlana Vyetrenko
NeurIPS SyntheticData4ML (2022).
Keywords: Time-series, Neural Fields, Signal Processing, Hypernetworks.

Neural BRDF Representation and Importance Sampling (Wiley Top Cited Paper)
Alejandro Sztrajman, Gilles Rainer, Tobias Ritschel, Tim Weyrich
Computer Graphics Forum (CGF), 40(6), pp. 332-346, 2021 (Oral Presentation at EGSR 2022).
Keywords: Material Appearance, Implicit Neural Representations, Meta-learning, Hypernetworks, Differentiable Rendering, Monte Carlo Sampling.

Fast Blue Noise Generation via Unsupervised Learning
*Daniele Giunchi, *Alejandro Sztrajman, Anthony Steed
International Joint Conference on Neural Networks (IJCNN), Padua, Italy, July 2022 (Oral Presentation).
Keywords: Blue-Noise, Unsupervised Learning, Dithering, Monte Carlo Integration.

Machine Learning Applications in Appearance Modelling
Alejandro Sztrajman
University College London, 2022 (PhD Thesis).
Reviewers: Profs. Lourdes Agapito and Diego Gutierrez.

Mixing Modalities of 3D Sketching and Speech for Interactive Model Retrieval in Virtual Reality
Daniele Giunchi, Alejandro Sztrajman, Stuart James, Anthony Steed
ACM International Conference on Interactive Media Experiences (IMX), New York, USA, Jun 2021 (Oral Presentation).
Keywords: Virtual Reality, 3D Sketching, CNNs, Human-in-the-loop.

High-Dynamic-Range Lighting Estimation from Face Portraits
Alejandro Sztrajman, Alexandros Neophytou, Tim Weyrich, Eric Sommerlade
International Conference on 3D Vision (3DV), Fukoka, Japan, Nov 2020 (Oral Presentation).
Keywords: HDR Lighting, CNNs, Inverse Rendering, Light Estimation.

Image-based remapping of spatially-varying material appearance
Alejandro Sztrajman, Jaroslav Krivanek, Alexander Wilkie, Tim Weyrich
Journal of Computer Graphics Techniques (JCGT), 8(4), pp. 1-30, 2019.
Keywords: Material Appearance, SVBRDF, Non-linear Optimization, Neural Networks.

Image-based remapping of material appearance
Alejandro Sztrajman, Jaroslav Krivanek, Alexander Wilkie, Tim Weyrich
Eurographics Workshop on Material Appearance Modeling (MAM), Helsinki, Finland, Jun 2017 (Oral Presentation).
Keywords: Material Appearance, BRDF, Non-linear Optimization.

Elementary Electromagnetism
Juan G. Roederer
Buenos Aires University Press, Buenos Aires (2015).

I coordinated the editorial team behind the undergraduate textbook by Prof. Juan Roederer. The book was first published in 2015. A second edition was released in 2018, and a digital edition followed in 2020.

Patents

Method and System for Generation of Interpretable Time Series with Implicit Neural Representations
Elizabeth Fons, Svitlana Vyetrenko, Yousef El-Laham, Alejandro Sztrajman, Alexandros Iosifidis
US Patent App. 18/215, 499 (Pending, 2024)
Estimating Illumination in an Environment Based on an Image of a Reference Object
Alexandros Neophytou, Eric Sommerlade, Alejandro Sztrajman, Sunando Sengupta
US Patent 0116549 A1 (2022).

Some Old Codes

Feature Saliency HMM
Implementation of the Feature Saliency HMM algorithm, for feature selection on Hidden Markov Models.
[Code]
  • Python
  • Hidden Markov Model
  • Path tracing
    A path tracer with support for environment map illumination.
    [Code]
  • C++
  • Global Illumination
  • 2D Lattice-Boltzmann Simulation
    Interactive simulation of incompressible viscous fluid with Lattice-Boltzmann D2Q9.
    [Code] [Video]
  • C++
  • Qt
  • Physics-based Animation
  • UTIA BRDF Reader
    Script to read measured BRDF data from the UTIA BRDF database.
    [Code]
  • Python
  • BRDF
  • WebGL 3D Engine
    An interactive WebGL 3D engine with phong shading and support for Bezier surfaces.
    [Video] [Code] [Try it!]
  • Javascript
  • WebGL
  • 2D FLIP Simulation
    Fluid Implicit Particle (FLIP) simulation of a 2D incompressible inviscid fluid.
    [Code] [Video]
  • C++
  • Qt
  • Physics-based Animation
  • Teaching

    2023 Advanced Graphics and Image Processing Cambridge
    2023 Introduction to Graphics Cambridge
    2019 Advanced Deep Learning and Reinforcement Learning UCL/DeepMind
    2018 Research Methods and Reading UCL
    2017 Scientific Programming in Python UCL
    2016 Robotics Programming UCL
    2016 Principles of Programming UCL