Potential-driven Monte Carlo Particle Tracing for Diffuse Environments with Adaptive Probability Functions

Abstract

A possible method for solving the global illumination problem is to use a Monte Carlo model, where particles are shot from the light sources and perform a random walk through the scene. The proposed algorithm tries to optimise the sampling process by constructing probability functions that closely match the visual potential function. Importance sampling ensures us that, within the given resolution and accuracy of the probability functions, particles are used in an optimal way, thereby lowering the overall variance of the picture. Sampling based on the local potential functions is done at light sources and surface patches, and thus influences every step of the random walk of a particle.