IP Library Granted Patent US 10,140,732
Granted Patent B2
US 10,140,732 · App. 15/300,256 · Granted Nov 27, 2018

Method and apparatus for efficient texture compression

Inventors: Tomasz Maksymczuk (Gdansk, PL); Tomasz Poniecki (Gdynia, PL)
Assignee: Intel Corporation
G06T11/001G06T9/00G06T15/04
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 10,140,732
App. No.
15/300,256
Filed
Sep 28, 2016
Granted
Nov 27, 2018
Kind
B2
Art Unit
2618
USPC
345/582
Abstract

An apparatus and method are described for texture compression. For example, one embodiment of a method comprises: determining a distance between each of a plurality of texture block texels and each of a plurality of points; determining a set of texel color values sampled over the texture block; and generating a set of approximation coefficients to compress the texture block using the distance between each of the plurality of texture block texels and each of the plurality of points and the set of texel color values sampled over the texture block.

Claims (46)

1. A method comprising:

determining a distance between each of a plurality of texture block texels in a texture block and each of a plurality of points, wherein the plurality of points are selected from at least an edge of the texture block, and wherein the plurality of points comprise radial basis function (RBF) center points;

determining a set of texel color values sampled over the texture block;

generating a set of approximation coefficients to compress the texture block using the distance between each of the plurality of texture block texels and each of the plurality of points and the set of texel color values sampled over the texture block; and

applying texture data to an image using the set of approximation coefficients.

2. The method as in claim 1 further comprising:

determining a distance matrix [DM] using the distance between each of the plurality of texture block texels and each of the plurality of points;

determining a RBF matrix [RDM] of member-wise RBF values over the distance matrix [DM] using a specified type of RBF and specified shape parameter e p value.

3. The method as in claim 2 wherein the specified type of RBF is selected from a group consisting of Gaussian and Multiquadratic radial basis functions.

4. The method as in claim 2 wherein determining a set of texel color values comprises building a vector [T] of B texel color values sampled over the texture block where B is the number of texels in the texture block.

5. The method as in claim 4 wherein generating the set of approximation coefficients comprises:

determining a vector [A] comprising the approximation coefficients using the equation

[ RDM]*[A]=[T].

6. The method as in claim 5 further comprising:

determining a pseudoinverse matrix [iRDM] of the RBF matrix [RDM]; and

multiplying [iRDM] by [T] to determine [A].

7. The method as in claim 6 wherein the texture block comprises a 4×4 texture block, a 5×5 texture block, a 6×6 texture block, a 7×7 texture block, or an 8×8 texture block.

8. The method as in claim 7 wherein the plurality of points are selected from a group consisting of 4, 8, 12, 16, or 20 points.

9. The method as in claim 6 wherein the texture block comprises a rectangular texture block.

10. The method as in claim 5 further comprising:

decompressing the texture block by determining the vector [T] using the equation

[ RDM]*[A]=[T].

11. A processor comprising:

texture compression circuitry to:

determine a distance between each of a plurality of texture block texels in a texture block and each of a plurality of points, wherein the plurality of points are selected from at least an edge of the texture block, and wherein the plurality of points comprise radial basis function (RBF) center points;

determine a set of texel color values sampled over the texture block;

generate a set of approximation coefficients to compress the texture block using the distance between each of the plurality of texture block texels and each of the plurality of points and the set of texel color values sampled over the texture block; and

apply texture data to an image using the set of approximation coefficients.

12. The processor as in claim 11 wherein the texture compression circuitry is to determine a distance matrix [DM] using the distance between each of the plurality of texture block texels and each of the plurality of points and determine a RBF matrix [RDM] of member-wise RBF values over the distance matrix [DM] using a specified type of RBF and specified shape parameter e p value.

13. The processor as in claim 12 wherein the specified type of RBF is selected from a group consisting of Gaussian and Multiquadratic radial basis functions.

14. The processor as in claim 13 wherein determining a set of texel color values comprises building a vector [T] of B texel color values sampled over the texture block where B is the number of texels in the texture block.

15. The processor as in claim 14 wherein the texture compression circuitry is to generate the set of approximation coefficients by determining a vector [A] comprising the approximation coefficients using the equation [RDM]*[A]=[T].

16. The processor as in claim 15 wherein the texture compression circuitry is to determine a pseudoinverse matrix [iRDM] of the RBF matrix [RDM]; and multiply [iRDM] by [T] to determine [A].

17. The processor as in claim 16 wherein the texture block comprises a 4×4 texture block, a 5×5 texture block, a 6×6 texture block, a 7×7 texture block, or an 8×8 texture block.

18. The processor as in claim 17 wherein the plurality of points are selected from a group consisting of 4, 8, 12, 16, or 20 points.

19. The processor as in claim 16 wherein the texture block comprises a rectangular texture block.

20. The processor as in claim 15 further comprising:

texture decompression circuitry to decompress the texture block by determining the vector [T] using the equation [RDM]*[A]=[T].

21. A non-transitory machine-readable medium having program code stored thereon which, when executed by a machine, causes the machine to perform the operations of:

determining a distance between each of a plurality of texture block texels in a texture block and each of a plurality of points, wherein the plurality of points are selected from at least an edge of the texture block, and wherein the plurality of points comprise radial basis function (RBF) center points;

determining a set of texel color values sampled over the texture block; and

generating a set of approximation coefficients to compress the texture block using the distance between each of the plurality of texture block texels and each of the plurality of points and the set of texel color values sampled over the texture block; and

applying texture data to an image using the set of approximation coefficients.

22. The non-transitory machine-readable medium as in claim 21 comprising additional program code to cause the machine to perform the operations of:

determining a distance matrix [DM] using the distance between each of the plurality of texture block texels and each of the plurality of points;

determining a RBF matrix [RDM] of member-wise RBF values over the distance matrix [DM] using a specified type of RBF and specified shape parameter e p value.

Continuity (1)
Related Publication 20170154443A1 · Jun 1, 2017
Cited By (1)
US 12,376,210