Jaroslav Křivánek

Adaptive sampling and Markov chain Monte Carlo methods in light transport simulation

Research project funded by the Czech Science Foundation (GA ČR).

Abstract

An important problem in computer graphics is realistic image synthesis, with light transport simulation being its fundamental component. The objective of this project is to develop new methods that will be able to efficiently simulate light in a much wider range of environments than the existing ones. To this end, we propose novel research on adaptive sampling in Monte Carlo methods as well as the use of Markov chain Monte Carlo. The main novelty of our research is a rigorous statistical approach to exploiting different sources of information – often suffering from statistical error – in adaptive simulation. Apart from making a theoretical contribution, our project will benefit practical applications such as the film industry or architectural visualization.

Publications

2018

Oskar Elek and Jaroslav Křivánek.
Towards a Principled Kernel Prediction for Spatially Varying BSSRDFs
Eurographics Workshop on Material Appearance Modeling, June 2018.

Jaroslav Křivánek, Ondřej Karlík, Vladimir Koylazov, Henrik Wann Jensen, Thomas Ludwig, and Christophe Chevallier.
Realistic Rendering in Architecture and Product Visualization
SIGGRAPH 2018 Course.

Jan Novák, Iliyan Georgiev, Johannes Hanika, Jaroslav Křivánek, and Wojciech Jarosz.
Monte Carlo Methods for Physically Based Volume Rendering
SIGGRAPH 2018 Course.

Alexander Keller, Jaroslav Křivánek, Jan Novák, Anton Kaplanyan, and Marco Salvi.
Machine Learning and Rendering
SIGGRAPH 2018 Course.

Pascal Grittmann, Arsene Pérard-Gayot, Philipp Slusallek, and Jaroslav Křivánek
Efficient Caustic Rendering with Lightweight Photon Mapping
Computer Graphics Forum, EGSR 2018.

Sebastian Herholz, Oskar Elek, Jens Schindel, Jaroslav Křivánek, and Hendrik Lensch
A Unified Manifold Framework for Efficient BRDF Sampling based on Parametric Mixture Model
Proceedings of EGSR (Experimental Ideas & Implementations), June 2018.

Petr Vévoda*, Ivo Kondapaneni*, and Jaroslav Křivánek
Bayesian online regression for adaptive direct illumination sampling
ACM Transactions on Graphics (Proceedings of SIGGRAPH 2018), 37(4), 2018

2017

Oskar Elek*, Denis Sumin*, Ran Zhang, Tim Weyrich, Karol Myszkowski, Bernd Bickel, Alexander Wilkie, and Jaroslav Křivánek
Scattering-aware Texture Reproduction for 3D Printing
ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia), 36(6), 2017
* Oskar Elek and Denis Sumin share the first authorship of this work.

Toshiya Hachisuka, Iliyan Georgiev, Wojciech Jarosz, Jaroslav Křivánek, and Derek Nowrouzezahrai.
Extended Path Integral Formulation for Volumetric Transport
Proceedings of EGSR (Experimental Ideas & Implementations), June 2017.

Jan Beneš, Tom Kelly, Filip Děchtěrenko, Jaroslav Křivánek, and Pascal Müller.
On Realism of Architectural Procedural Models
Computer Graphics Forum (Eurographics 2017)

2016

Sebastian Herholz, Oskar Elek, Jiří Vorba, Hendrik Lensch, and Jaroslav Křivánek.
Product Importance Sampling for Light Transport Path Guiding
Computer Graphics Forum, EGSR 2016.

Martin Šik, Hisanari Otsu, Toshiya Hachisuka, Jaroslav Křivánek
Robust Light Transport Simulation via Metropolized Bidirectional Estimators
ACM Trans. Graph., ACM SIGGRAPH Asia 2016

Martin Šik, Jaroslav Křivánek
Improving global exploration of MCMC light transport simulation
ACM SIGGRAPH 2016 posters.

Petr Vévoda, and Jaroslav Křivánek
Adaptive Direct Illumination Sampling
ACM SIGGRAPH Asia 2016 posters.

Adrien Gruson, Mickael Ribardiere, Martin Šik, Jiří Vorba, Rémy Cozot, Kadi Bouatouch, Jaroslav Křivánek
A Spatial Target Function for Metropolis Photon Tracing
ACM Trans. Graph. 2016.

Jiří Vorba, Jaroslav Křivánek
Adjoint-Driven Russian Roulette and Splitting in Light Transport Simulation
ACM Trans. Graph., ACM SIGGRAPH 2016.