Technical Papers

Global Illumination

Saturday, 01 December 11:00 - 13:00 |  Peridot 206

A Path Space Extension for Robust Light Transport Simulation - Picture

A Path Space Extension for Robust Light Transport Simulation

A novel light transport simulation framework that allows a sound combination of unbiased Monte Carlo path integration and photon density estimation.


Toshiya Hachisuka, Aarhus University
Jacopo Pantaleoni, NVIDIA Research
Henrik Jensen, UC San Diego


Light Transport Simulation with Vertex Connection and Merging - Picture

Light Transport Simulation with Vertex Connection and Merging

Our novel formulation of the photon mapping method in the path integral light transport framework allows us to build a new consistent rendering algorithm that is more robust than both bidirectional path tracing and photon mapping, allowing us to efficiently render scenes with diffuse and specular lighting.


Iliyan Georgiev, Saarland University, Intel Visual Computing Institute Saarbrücken
Jaroslav Křivánek, Charles University, Prague
Tomas Davidovic, Saarland University, Intel Visual Computing Institute Saarbrücken
Philipp Slusallek, DFKI Saarbrücken, Saarland University, Intel Visual Computing Institute Saarbrücken


Practical Hessian-Based Error Control for Irradiance Caching - Picture

Practical Hessian-Based Error Control for Irradiance Caching

We introduce a new error metric for irradiance caching based on a new derivation of occlusion-aware irradiance Hessian that significantly outperforms previous approaches such as the split-sphere and adaptive caching.


Jorge Schwarzhaupt, UC San Diego
Henrik Wann Jensen, UC San Diego
Wojciech Jarosz, Disney Research, Zürich


SURE-based Optimization for Adaptive Sampling and Reconstruction - Picture

SURE-based Optimization for Adaptive Sampling and Reconstruction

We propose a novel adaptive sampling and reconstruction method by introducing SURE for error estimation. The incorporation of SURE enables us to use more effective kernels, select optimal filter and sampling distribution. We demonstrate significant improvement over state-of-the-art methods.


Tzu-Mao Li, National Taiwan University
Yu-Ting Wu, National Taiwan University
Yung-Yu Chuang, National Taiwan University


Adaptive Rendering with Non-Local Means Filtering - Picture

Adaptive Rendering with Non-Local Means Filtering

We propose a novel image space adaptive sampling and filtering framework for Monte Carlo raytracing which can handle arbitrary light transport and lens effects. Our method leverages a state of the art image denoising technique, making it robust to high noise levels and complex image content.


Fabrice Rousselle, University of Bern
Claude Knaus, University of Bern
Matthias Zwicker, University of Bern