Pixel blending for neural network-based image generation

Patent metadata

Field	Value
Publication number	US-12394113-B2
Application number	US-202117351303-A
Country	US
Kind code	B2
Filing date	Jun 18, 2021
Priority date	Jun 18, 2021
Publication date	Aug 19, 2025
Grant date	Aug 19, 2025

Abstract

Apparatuses, systems, and techniques are presented to generate one or more images. In at least one embodiment, two or more pixels from two or more images are blended based, at least in part, on a distance of the two or more pixels from a region of the two or more images, in which pixel colors are substantially similar.

First claim

What is claimed is:

1. A processor, comprising:
one or more circuits to blend two or more pixels from two or more images based, at least in part, on a distance of the two or more pixels from a region of the two or more images, in which pixel colors are substantially similar.

2. The processor of claim 1 , wherein the two or more images are adjacent images in an image sequence, and wherein the two or more pixels correspond to correlated pixel locations in the two or more images.

3. The processor of claim 1 , wherein the region includes correlated pixels with color values that agree within a determined agreement threshold or agreement range.

4. The processor of claim 1 , wherein the two or more pixels are to be blended using a blending algorithm that is to produce one or more blending factors as a function of the distance.

5. The processor of claim 4 , wherein the blending algorithm includes one or more user-adjustable parameters to control a value of the one or more blending factors.

6. The processor of claim 1 , wherein the one or more circuits are further to blend the two or more pixels after error correction is applied to the two or more images based on at least one of optical flow or geometric motion data.

7. A system comprising:
one or more processors to blend two or more pixels from two or more images based, at least in part, on a distance of the two or more pixels from a region of the two or more images, in which pixel colors are substantially similar.

8. The system of claim 7 , wherein the two or more images are adjacent images in an image sequence, and wherein the two or more pixels correspond to correlated pixel locations in the two or more images.

9. The system of claim 7 , wherein the region includes correlated pixels with color values that agree within a determined agreement threshold or agreement range.

10. The system of claim 7 , wherein the two or more pixels are to be blended using a blending algorithm that is to produce one or more blending factors as a function of the distance.

11. The system of claim 10 , wherein the blending algorithm includes one or more user-adjustable parameters to control a value of the one or more blending factors.

12. The system of claim 7 , wherein the one or more circuits are further to blend the two or more pixels after error correction is applied to the two or more images based on at least one of optical flow or geometric motion data.

13. A method comprising:
blending two or more pixels from two or more images based, at least in part, on a distance of the two or more pixels from a region of the two or more images, in which pixel colors are substantially similar.

14. The method of claim 13 , wherein the two or more images are adjacent images in an image sequence, and wherein the two or more pixels correspond to correlated pixel locations in the two or more images.

15. The method of claim 13 , wherein the region includes correlated pixels with color values that agree within a determined agreement threshold or agreement range.

16. The method of claim 13 , wherein the two or more pixels are to be blended using a blending algorithm that is to produce one or more blending factors as a function of the distance.

17. The method of claim 16 , wherein the blending algorithm includes one or more user-adjustable parameters to control a value of the one or more blending factors.

18. The method of claim 13 , wherein the one or more circuits are further to blend the two or more pixels after error correction is applied to the two or more images based on at least one of optical flow or geometric motion data.

19. A machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least:
blend two or more pixels from two or more images based, at least in part, on a distance of the two or more pixels from a region of the two or more images, in which pixel colors are substantially similar.

20. The machine-readable medium of claim 19 , wherein the two or more images are adjacent images in an image sequence, and wherein the two or more pixels correspond to correlated pixel locations in the two or more images.

21. The machine-readable medium of claim 19 , wherein the region includes correlated pixels with color values that agree within a determined agreement threshold or agreement range.

22. The machine-readable medium of claim 19 , wherein the two or more pixels are to be blended using a blending algorithm that is to produce one or more blending factors as a function of the distance.

23. The machine-readable medium of claim 22 , wherein the blending algorithm includes one or more user-adjustable parameters to control a value of the one or more blending factors.

24. The machine-readable medium of claim 19 , wherein the one or more circuits are further to blend the two or more pixels after error correction is applied to the two or more images based on at least one of optical flow or geometric motion data.

25. An image blending system, comprising:
one or more processors to blend two or more pixels from two or more images based, at least in part, on a distance of the two or more pixels from a region of the two or more images, in which pixel colors are substantially similar; and
memory for storing image data comprising the pixel colors for the blended pixels.

26. The image blending system of claim 25 , wherein the two or more images are adjacent images in an image sequence, and wherein the two or more pixels correspond to correlated pixel locations in the two or more images.

27. The image blending system of claim 25 , wherein the region includes correlated pixels with color values that agree within a determined agreement threshold or agreement range.

28. The image blending system of claim 25 , wherein the two or more pixels are to be blended using a blending algorithm that is to produce one or more blending factors as a function of the distance.

29. The image blending system of claim 28 , wherein the blending algorithm includes one or more user-adjustable parameters to control a value of the one or more blending factors.

30. The image blending system of claim 25 , wherein the one or more circuits are further to blend the two or more pixels after error correction is applied to the two or more images based on at least one of optical flow or geometric motion data.

Assignees

• Nvidia Corp

Inventors

• Pottorff Robert
• Sapra Karan
• Tao Andrew
• Catanzaro Bryan
• Lunden Jarmo

Classifications

• G06T11/10 Primary

  Texturing; Colouring; Generation of textures or colours (retouching, inpainting or scratch removal G06T5/77) · CPC title

• G06T7/90

  Determination of colour characteristics · CPC title

• G06T2207/10024

  Color image · CPC title

• G06T2200/24

  involving graphical user interfaces [GUIs] · CPC title

• G06T2207/20084

  Artificial neural networks [ANN] · CPC title