Sampling & Variance Analysis for Mone Carlo Integration in the Spherical domain

Gurprit Singh presents a theoretical framework to study different sampling patterns in the spherical domain and their effects in the evaluation of global illumination integrals.

October 13, 2015
4:15 pm - 5:30 pm
Location
Moore B03
Sponsored by
Computer Science Department
Audience
Public
More information
Sandra Hall

Abstract: 

We present a theoretical framework to study different sampling patterns in the spherical domain and their effects in the evaluation of global illumination integrals. Evaluating illumination (light transport) integrals is one of the essential aspect in image synthesis to achieve realism which involves solving multi-dimensional space integrals. Monte Carlo based numerical integration schemes are heavily employed to solve these high dimensional integrals. One of the most important aspect of any numerical integration method is sampling. The way samples are distributed on an integration domain can greatly affect the final result. For example, in images, the effects of various sampling patterns appears in the form of either structural artifacts or completely unstructured noise. In many cases, we may get completely false (biased) results due to the sampling pattern used in integration.

We study the Fourier content of various sampling patterns in the spherical domain and develop a theoretical framework that provides a direct closed-form relation between the variance in Monte Carlo integration and the frequency content of both the integrand and the sampling pattern involved.

Bio:

Gurprit Singh is a Postdoc in the Visual Computing Lab at Dartmouth College. He recently graduated (2015) from the Université de Lyon 1, France, with his PhD focus on sampling analysis for Monte Carlo integration in global illumination related problems.  Before his PhD, he did his Masters (2012) in Graphics, Vision and Robotics from INP Grenoble, France  and his B.Tech (2010) from Indian Institute of Technology Delhi, India. 

Location
Moore B03
Sponsored by
Computer Science Department
Audience
Public
More information
Sandra Hall