US 11,854,167 B2
Photographic underexposure correction using a neural network
Kevin Gordon, Edmonton (CA); Darcy Daugela, Edmonton (CA); and Martin Humphreys, Sherwood Park (CA)
Assigned to SPECTRUM OPTIX INC., Vancouver (CA)
Filed by Spectrum Optix Inc., Vancouver (CA)
Filed on Jun. 22, 2021, as Appl. No. 17/354,926.
Application 17/354,926 is a continuation of application No. 16/570,537, filed on Sep. 13, 2019, granted, now 11,076,103.
Claims priority of provisional application 62/844,496, filed on May 7, 2019.
Claims priority of provisional application 62/730,799, filed on Sep. 13, 2018.
Prior Publication US 2021/0392258 A1, Dec. 16, 2021
Int. Cl. G06T 5/00 (2006.01); G06T 5/50 (2006.01); H04N 23/68 (2023.01); H04N 23/72 (2023.01); G06F 18/214 (2023.01); G06V 10/82 (2022.01); H04N 23/741 (2023.01)
CPC G06T 5/003 (2013.01) [G06F 18/214 (2023.01); G06T 5/002 (2013.01); G06T 5/50 (2013.01); G06V 10/82 (2022.01); H04N 23/683 (2023.01); H04N 23/72 (2023.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/20182 (2013.01); H04N 23/741 (2023.01)] 6 Claims
OG exemplary drawing
 
1. A method for image noise reduction, comprising:
capturing with a camera an image having noise;
processing the image using a first neural network to produce denoised tensor data and reduce noise captured within the image; and
passing the denoised tensor data and noise reduced image to a second neural network that provides at least one of object recognition, pattern recognition, face identification, image stabilization, robot or vehicle odometry and positioning, or tracking or targeting applications.