US 11,854,165 B2
Debanding using a novel banding metric
Yilin Wang, Sunnyvale, CA (US); Balineedu Adsumilli, Sunnyvale, CA (US); and Feng Yang, Sunnyvale, CA (US)
Assigned to GOOGLE LLC, Mountain View, CA (US)
Appl. No. 17/922,531
Filed by Google LLC, Mountain View, CA (US)
PCT Filed May 19, 2020, PCT No. PCT/US2020/033545
§ 371(c)(1), (2) Date Oct. 31, 2022,
PCT Pub. No. WO2021/236061, PCT Pub. Date Nov. 25, 2021.
Prior Publication US 2023/0131228 A1, Apr. 27, 2023
Int. Cl. G06T 5/00 (2006.01); G06T 5/20 (2006.01); G06V 10/56 (2022.01); G06V 10/74 (2022.01); G06T 7/13 (2017.01)
CPC G06T 5/002 (2013.01) [G06T 5/20 (2013.01); G06T 7/13 (2017.01); G06V 10/56 (2022.01); G06V 10/761 (2022.01); G06T 2207/20081 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer apparatus, comprising:
a memory storing processor readable instructions; and
a processor arranged to read and execute instructions stored in the memory, wherein the processor readable instructions comprise instructions arranged to control the computer to carry out a method of training a model to deband an image, comprising:
for a training image of a second set of second training images:
receiving a training debanded image, the training debanded image comprising image data obtained by removing banding artefacts from the training image;
generating a banding score for the training debanded image;
generating an image difference between the training image of the second set of training images and the training debanded image; and
using a weighted combination of the image difference and the banding score in a loss function that is used to train a second model.