US 12,112,456 B2
Retouching digital images utilizing layer specific deep-learning neural networks
Federico Perazzi, San Jose, CA (US); and Jingwan Lu, Santa Clara, CA (US)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Oct. 11, 2022, as Appl. No. 18/045,730.
Application 18/045,730 is a continuation of application No. 17/072,372, filed on Oct. 16, 2020, granted, now 11,521,299.
Prior Publication US 2023/0058793 A1, Feb. 23, 2023
Int. Cl. G06K 9/00 (2022.01); G06F 18/21 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2023.01); G06T 3/4046 (2024.01); G06T 5/70 (2024.01); G06T 7/11 (2017.01); G06V 30/262 (2022.01)
CPC G06T 5/70 (2024.01) [G06F 18/21 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2013.01); G06T 3/4046 (2013.01); G06T 7/11 (2017.01); G06V 30/274 (2022.01); G06T 2207/20084 (2013.01); G06T 2207/30201 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
separating a digital image into a high-frequency layer and a mid-frequency layer;
generating a modified mid-frequency layer by utilizing a first neural network to correct skin blemishes in the mid-frequency layer; and
combining the high-frequency layer, the modified mid-frequency layer, and a low-frequency layer to generate a modified digital image.