System, method, and computer program product for performing path space filtering

Patent metadata

Field	Value
Publication number	US-9953457-B2
Application number	US-201414166794-A
Country	US
Kind code	B2
Filing date	Jan 28, 2014
Priority date	Apr 22, 2013
Publication date	Apr 24, 2018
Grant date	Apr 24, 2018

Abstract

A system, method, and computer program product are provided for performing path space filtering. In use, a set of light transport paths associated with a scene is sampled. Additionally, a plurality of vertices associated with the sampled set of light transport paths is selected, where each selected vertex has an associated throughput and light contribution. Further, an averaged light contribution of each of the selected plurality of vertices is determined, utilizing one or more weights. Further still, the averaged light contribution of each of the selected plurality of vertices is combined after multiplying the averaged light contribution of each of the selected vertices by the associated throughput of the vertex.

First claim

What is claimed is:

1. A method, comprising:
sampling a set of light transport paths associated with a scene;
selecting a plurality of vertices associated with the sampled set of light transport paths, wherein each selected vertex in the plurality of vertices is associated with a throughput and a light contribution;
determining an averaged light contribution of each selected vertex in the plurality of vertices, utilizing one or more weights, wherein the averaged light contribution for a particular selected vertex is calculated based on the light contributions for vertices in proximity to the particular selected vertex;
combining the averaged light contribution c of each selected vertex in the plurality of vertices with the throughput α associated with the selected vertex to generate a product, α· c , for the selected vertex, wherein the throughput associated with each selected vertex includes an accumulated attenuation of light as transported from that vertex to at least one of a camera or a screen;
forming an image by accumulating the product for each selected vertex in the plurality of vertices; and
for the i-th light transport path within the sampled set of light transport paths, storing at least one vertex x i , an attenuation α i along an eye path segment to the vertex x i , and a radiance contribution c i , of a light path segment to the vertex x i .

2. The method of claim 1 , further comprising repeating the selecting, determining, and combining in an iterative manner.

3. The method of claim 1 , further comprising locally filtering the sampled set of light transport paths.

4. The method of claim 3 , wherein the filtering is performed before the determining of the averaged light contribution of each selected vertex in the plurality of vertices.

5. The method of claim 3 , wherein one or more of a random, pseudo-random, or quasi-random selection of the samples to be filtered is performed.

6. The method of claim 3 , wherein the local filtering improves an efficiency of a numerical integro-approximation.

7. The method of claim 3 , wherein combining the averaged light contribution of each selected vertex in the plurality of vertices is performed during a light transport simulation, and the local filtering improves an efficiency of the light transport simulation.

8. The method of claim 3 , wherein locally filtering the sampled set of light transport paths is performed linearly and is applied on top of a sampling based rendering engine.

9. The method of claim 1 , wherein the set of light transport paths is sampled by one or more of generating eye paths, generating light paths, and connecting one or more eye paths and light paths by one or more of shadow rays and proximity.

10. The method of claim 1 , wherein the method is implemented as a progressive method.

11. The method of claim 1 , wherein c i is a ratio of the sum of weighted contributions

w

i
,
j

·

c

⌊

i

b
m

⌋

⁢

b
m

+
j

and the sum of weights w i,j for all vertices

x

⌊

i

b
m

⌋

⁢

b
m

+
j

⁢

⁢

in a ball B of radius r(n) around the vertex x i , as selected by a characteristic function χ B(n) , for m contiguous blocks b m of n light transport paths.

12. The method of claim 11 , wherein the radius r(n) of the ball B has the property of

r
2

⁡

(
n
)

:=

r
0
2

n
α

⁢

⁢
for
⁢

⁢
α

∈

(

0
,
1

)

.

13. The method of claim 11 , wherein the radius r(n) is chosen by at least one of a distance along a path or a solid angle.

14. The method of claim 13 , wherein each of the one or more weights measure a similarity between one or more of the selected plurality of vertices, where such similarity depends on one or more of a geometric similarity, a transport similarity, and a visibility similarity between one or more of the selected plurality of vertices.

15. The method of claim 1 , wherein the plurality of vertices are selected utilizing a range search.

16. The method of claim 15 , wherein the range search is performed by at least one of a hash grid, a classic range search, and a divide-and-conquer method.

17. The method of claim 1 , wherein the method is applied across multiple views for one or more of rendering stereo image pairs, rendering multiple frames in an animation, and rendering a set of images required for one or more light field displays.

18. The method of claim 1 , further comprising simulating motion blur by averaging one or more images at one or more distinct points in time.

19. The method of claim 1 , wherein sampling the set of light transport paths is decoupled from one or more shading operations.

20. The method of claim 1 , wherein the sampling a set of light transport paths is performed by one or more of a random, pseudo-random, or quasi-random sampling.

Assignees

• Nvidia Corp

Inventors

• Keller Alexander
• Dahm Ken Patrik
• Binder Nikolaus

Classifications

• G06T15/06

  Ray-tracing · CPC title

• G06T15/506 Primary

  Illumination models · CPC title