wilkei

Alexander Wilkie

Associate Professor

Head of the computer graphics branch of CGG
At Charles University since 2008.

Core research interest: computer graphics so accurate that it can be used to predict what things would look like, if they were real.
Predictive Rendering, in other words.

Of the many potential research areas within Predictive Rendering, I have over the years specialised in the following:

  1. Spectral rendering, in particular
    • Inclusion of fluorescence into rendering pipelines
    • Polarisation effects
    • Spectral uplifting
  2. Skylight models
  3. Colour science
  4. Contrast enhancement for full colour 3D printing

The last point might seem like an odd addition to this research portfolio, until one realises that optimisation of 3D print appearance critically depends on being able to predict what a 3D print-out would look like, if it were created with a specific voxel configuration. So moving the state of the art in this area forward really belongs to the rendering people, as prediction of print-out appearance is the hardest part of any such feedback loop.

A very long term project of mine is ART: The Advanced Rendering Toolkit. This is rendering research software we have used over the years to prototype various effects that do not fit well into conventional rendering software. It is still being updated, although a public release of the current version is overdue (planned for June/July 2021, as we want to publicly demonstrate an integration of our new sky dome model into a path tracer).

 

Teaching

 

  • Introduction to Colour Science
  • Predictive Image Synthesis Technologies
  • Seminar on Scientific Soft Skills
  • Introduction to Computer Graphics
  • Special Seminar in Computer Graphics

 

Publications & Research

 

  • A Fitted Radiance and Attenuation Model for Realistic Atmospheres
  • A Gradient-Based Framework for 3D Print Appearance Optimization
  • An Open EXR Layout for Spectral Images
  • Moment-based Constrained Spectral Uplifting
  • A Compact Representation for Fluorescent Spectral Data
  • A Multiscale Microfacet Model Based on Inverse Bin Mapping
  • Neural Acceleration of Scattering-Aware Color 3D Printing
  • Robust and Practical Measurement of Volume Transport Parameters in Solid Photo-polymer Materials for 3D Printing
  • Efficient Multiscale Rendering of Specular Microstructure
  • Quantum Yield Bias in Materials With Lower Absorptance
  • Wide Gamut Spectral Upsampling with Fluorescence
  • Geometry-Aware Scattering Compensation for 3D Printing
  • Image-based Remapping of Spatially-varying Material Appearance
  • Handling Fluorescence in a Uni-directional Spectral Path Tracer
  • Scattering-aware Texture Reproduction for 3D Printing
  • Image-based Remapping of Material Appearance
  • Virtual Ellipsometry on Layered Micro-facet Surfaces
  • Bi-directional Polarised Light Transport
  • Hero Wavelength Spectral Sampling
  • Procedural Modelling of Urban Road Networks
  • Adding a Solar-Radiance Function to the Hošek-Wilkie Skylight Model
  • Predicting Sky Dome Appearance on Earth-like Extrasolar Worlds
  • An Analytic Model for Full Spectral Sky-Dome Radiance
  • Interactive Cloud Rendering Using Temporally-Coherent Photon Mapping
  • Modeling and Verifying the Polarizing Reflectance of Real-World Metallic Surfaces
  • A Physically Plausible Model for Light Emission from Glowing Solid Objects
  • A Standardised Polarisation Visualisation for Images
  • Anomalous Dispersion in Predictive Rendering
  • A Robust Illumination Estimate for Chromatic Adaptation in Rendered Images
  • Rendering the Effect of Labradorescence
  • Realistic Rendering of Birefringency in Uniaxial Crystals