INFORMED CHOICES IN PRIMARY SAMPLE SPACE FOR LIGHT TRANSPORT SIMULATION

Number: 20190139295
Published: 2019-05-09
Filed: 2017-11-09
Assignee: Disney Enterprises, Inc.
Inventors: NOVÁK; Jan, JAKOB; Wenzel A., JAROSZ; Wojciech, BITTERLI; Benedikt Martin
CPC: G06T15/55; G06T15/06; G06F30/00
Verdict: Medium Notable software
Source: Google Patents · FreePatentsOnline

The keeper's note

Informed choices primary sample space light transport simulation (rendering).

Abstract

Systems, methods and articles of manufacture for rendering three-dimensional virtual environments using reversible jumps are disclosed herein. In one embodiment, mappings from random numbers to light paths are modeled as an explicit iterative random walk. Inverses of path construction techniques are employed to turn light transport paths back into the random numbers that produced them. In particular, such inverses may be used to extend the Multiplexed Metropolis Light Transport (MMLT) technique to perform path-invariant perturbations that produce a new path sample using a different path construction technique but preserve the path's geometry. To render an image, a rendering application in one embodiment may trace light paths through a virtual scene, with some path samples being generated by probabilistically selecting one or more techniques through technique perturbation and using inverses of the selected technique(s) to invert existing path(s), and with new paths being obtained by mutating or perturbing existing paths.

Background

BACKGROUNDField

This disclosure provides techniques for rendering images of scenes including three-dimensional virtual geometry.Description of the Related Art

Rendering is the automated process of generating photorealistic or nonphotorealistic images from two or three-dimensional models. Monte Carlo rendering techniques have been used to simulate light propagation by sampling random paths connecting a light source and a virtual sensor. In scenes with complex materials, geometry, or lighting, the space of light paths can be large and high-dimensional, while the subset of paths contributing significantly to the rendered image may occupy only a narrow subspace. This can make Monte Carlo rendering a difficult sampling problem.

In contrast to traditional Monte Carlo rendering techniques, Markov Chain Monte Carlo (MCMC) rendering techniques, such as Metropolis Light Transport (MLT), generate a statistically dependent sequence of samples (light paths) with a density proportional to a non-negative function, typically using the Metropolis-Hastings steps, drawing a proposal state from a proposal distribution which is accepted with a certain probability to become the next state of the Markov Chain (with the process repeating anew if the proposal is not accepted). Primary Sample Space MLT (PSSMLT) retrofits such Metropolis sampling to traditional Monte Carlo techniques by treating them as abstract path samplers and perturbing the random numbers they consume. Multiplexed

Claims

1. A computer-implemented method for rendering an image of a virtual scene, the method comprising: generating a first path sample using a first path construction technique and a first sample-space vector; probabilistically selecting a second path construction technique from a plurality of path construction techniques; generating a second path sample, wherein generating the second path sample includes determining, using an inverse of the second path construction technique, a second sample-space vector corresponding to the first path sample; and adding a contribution of the second path sample to the image of the virtual scene. 12. A non-transitory computer-readable storage medium storing a program, which, when executed by a processor performs operations for rendering an image of a virtual scene, the operations comprising a plurality of steps, and at least one of the plurality of steps includes: generating a first path sample using a first path construction technique and a first sample-space vector; probabilistically selecting a second path construction technique from a plurality of path construction techniques; generating a second path sample, wherein generating the second path sample includes determining, using an inverse of the second path construction technique, a second sample-space vector corresponding to the first path sample; and adding a contribution of the second path sample to the image of the virtual scene. 20. A system, comprising: a processor; and a memory, wherein the memory includes an application program configured to perform operations for rendering an image of a virtual scene, the operations comprising a plurality of steps, and at least one of the plurality of steps includes: generating a first path sample using a first path construction technique and a first sample-space vector, probabilistically selecting a second path construction technique from a plurality of path construction techniques, generating a second path sample, wherein generating the second path sample includes determining, using an inverse of the second path construction technique, a second sample-space vector corresponding to the first path sample, and adding a contribution of the second path sample to the image of the virtual scene.