US 12,073,536 B2
Dirty lens image correction
Sagi Katz, Yokneam Ilit (IL); and David Ben Ezra, Los Angeles, CA (US)
Assigned to Snap Inc., Santa Monica, CA (US)
Filed by Snap Inc., Santa Monica, CA (US)
Filed on Aug. 9, 2022, as Appl. No. 17/883,602.
Application 17/883,602 is a continuation of application No. 17/145,963, filed on Jan. 11, 2021, granted, now 11,461,883.
Application 17/145,963 is a continuation of application No. 16/567,005, filed on Sep. 11, 2019, granted, now 10,896,494, issued on Jan. 19, 2021.
Claims priority of provisional application 62/737,442, filed on Sep. 27, 2018.
Prior Publication US 2022/0383467 A1, Dec. 1, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 13/239 (2018.01); G06T 5/10 (2006.01); G06T 5/50 (2006.01); G06T 7/00 (2017.01); G06T 7/593 (2017.01); G06T 7/90 (2017.01); H04N 13/00 (2018.01); H04N 13/128 (2018.01); H04N 13/332 (2018.01)
CPC G06T 5/50 (2013.01) [G06T 5/10 (2013.01); G06T 7/0002 (2013.01); G06T 7/593 (2017.01); G06T 7/90 (2017.01); G06T 7/97 (2017.01); H04N 13/128 (2018.05); H04N 13/239 (2018.05); H04N 13/332 (2018.05); G06T 2207/30168 (2013.01); H04N 2013/0077 (2013.01); H04N 2013/0081 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A dirty lens correction system comprising:
a memory storing instructions; and
a processor configured to execute the instructions to configure the dirty lens correction system to:
obtain a first image of a scene captured from a first viewpoint;
obtain a second image of the scene captured from a second viewpoint separated from the first viewpoint;
rectify the first and second images to create respective first and second rectified images, the first and second rectified images including pixel data;
determine disparity correspondence between corresponding image pixel pairs of the first and second rectified images;
map the first and second rectified images into the same domain using the determined disparity to produce a first domain mapped image and a second domain mapped image;
detect an image artifact region within the first domain mapped image if at least one of a color attribute value, a gradient value, or an intensity value in a region of the first domain mapped image is different than the color attribute value, the gradient value, or the intensity value in a corresponding region of the second domain mapped image;
determine correction factors for the detected image artifact region;
correct at least one of the rectified first or second images by applying the determined correction factors; and
create a three-dimensional (3D) scene for presentation using the corrected rectified images and the determined disparity.