Uniform density cube map rendering for spherical projections

Patent metadata

Field	Value
Publication number	US-11145108-B2
Application number	US-201916353178-A
Country	US
Kind code	B2
Filing date	Mar 14, 2019
Priority date	Mar 21, 2018
Publication date	Oct 12, 2021
Grant date	Oct 12, 2021

Abstract

A cube map is used for determining the appearance of a surface by means of a precomputed texture image. Embodiments of the present invention are drawn computer systems and methods for rendering a spherical projection as a cube map that mitigates non-uniform pixel density near the edges of the cube map to avoid artifacts and increase rendering performance.

First claim

What is claimed is:

1. A computer implemented method comprising:
dividing at least one cube face of a cube into a plurality of surfaces;
assigning viewports to the plurality of surfaces;
determining coordinates of vertexes of a spherical shape in world-space projected onto the plurality of surfaces;
generating rendered views of the spherical shape projected onto the plurality of surfaces from perspectives of the viewports based at least on rendering the views using the coordinates, wherein the views comprise the vertexes; and
generating a cube map based at least on expanding the rendered views into the at least one cube face, wherein the expanding comprises applying a transformation function to the rendered views of the spherical shape thereby increasing a uniformity of pixel density of the rendered views by adjusting world-space pixel sizes of pixels of the rendered views.

2. A method as described in claim 1 , wherein the transformation function is a first transformation function and the determining the coordinates of the vertexes is performed using a second transformation function:

{

u
,
v

}

=

{

x

x
2

+

y
2

+
1

,

y

x
2

+

y
2

+
1

}

,

wherein {u, v} represents the coordinates, and
wherein further {x, y} represents Cartesian coordinates of vertexes within a respective surface of the plurality of surfaces.

3. A method as described in claim 1 , further comprising:
performing a first rendering pass that performs the determining the coordinates of the vertexes and the rendering of the views; and
performing a second rendering pass that performs the expanding.

4. A method as described in claim 1 , wherein the expanding the rendered views into the at least one cube face is performed using the transformation function:

{

u
,
v

}

=

{

x

w
+

x
⁢

⁢
α

+

y
⁢

⁢
β

,

y

w
+

x
⁢

⁢
α

+

y
⁢

⁢
β

}

,

wherein

α
=

x
-
u

u
⁡

(

x
+
y

)

,

wherein |ß|==|α|, wherein {u, v} represents the vertexes projected onto the plurality of surfaces, wherein w represents a perspective value, and wherein further {x, y} represents Cartesian coordinates of vertexes within a respective surface of the plurality of surfaces.

5. A method as described in claim 1 , wherein the spherical shape represented in the rendered views is transformed into a plurality of planes, each plane corresponding to a respective viewport of the viewports.

6. A method as described in claim 1 , wherein the expanding includes shearing the spherical shape represented in the rendered views based at least on a portion of the spherical shape that is between screen coordinates.

7. A method as described in claim 1 , wherein the expanding unwraps regions of the spherical shape represented in the rendered views resulting in unwrapped versions of the rendered views, each of the regions corresponds to a respective viewport of the viewports, and the unwrapped versions of the rendered views are stored in a common face of the cube map.

8. A computer system comprising:
a memory operable for storing data and instructions of an application;
a processor configured to execute the instructions for controlling rendering of three-dimensional (3D) environments; and
a graphics processing unit communicatively coupled to the processor and the memory, wherein the graphics processing unit is configured to perform, responsive to the instructions, a method comprising:
dividing at least one cube face of a cube into a plurality of surfaces;
assigning viewports to the plurality of surfaces;
determining coordinates of vertexes of a spherical shape in world-space projected onto the plurality of surfaces;
generating rendered views of the spherical shape projected onto the plurality of surfaces from perspectives of the viewports based at least on rendering the views using the coordinates, wherein the views comprise the vertexes; and
generating a cube map based at least on expanding the rendered

Assignees

• Nvidia Corp

Inventors

• Dunn Alexander
• Schott Mathias
• Spitzer John

Classifications

• G06T15/005

  General purpose rendering architectures · CPC title

• G06T15/04 Primary

  Texture mapping · CPC title

• G09G5/363

  Graphics controllers · CPC title

• G09B29/005

  Map projections or methods associated specifically therewith · CPC title

• G06T2215/06

  Curved planar reformation of 3D line structures · CPC title