US 12,073,540 B2
Use motion data to generate higher resolution images
Raymond Kirk Price, Carnation, WA (US); Michael Bleyer, Seattle, WA (US); and Christopher Douglas Edmonds, Carnation, WA (US)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on May 2, 2023, as Appl. No. 18/142,418.
Application 18/142,418 is a continuation of application No. 17/567,570, filed on Jan. 3, 2022, granted, now 11,676,250.
Prior Publication US 2023/0267577 A1, Aug. 24, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06T 5/73 (2024.01); G06T 3/4053 (2024.01); G06T 5/50 (2006.01); H04N 5/265 (2006.01); H04N 23/84 (2023.01); H04N 23/951 (2023.01)
CPC G06T 5/73 (2024.01) [G06T 3/4053 (2013.01); G06T 5/50 (2013.01); H04N 5/265 (2013.01); H04N 23/843 (2023.01); H04N 23/951 (2023.01); G06T 2207/10024 (2013.01); G06T 2207/20201 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for generating an output color image that has a higher resolution than resolutions of input images that were used to generate the output color image, said method comprising:
accessing a plurality of images that were generated by an image sensor having a sparse Bayer pattern comprising a red pixel, a green pixel, and a blue pixel;
acquiring pose data for each image included in the plurality of images;
feeding the plurality of images and the pose data into a motion model, wherein the motion model performs temporal filtering on the plurality of images and uses the pose data to generate a red-only image, a green-only image, and a blue-only image; and
generating an output color image by combining the red-only image, the green-only image, and the blue-only image, wherein a resolution of the output color image is higher than resolutions of the plurality of images.