US 12,033,670 B2
Editing portrait videos
Seoungwug Oh, San Jose, CA (US); Joon-Young Lee, San Jose, CA (US); Jingwan Lu, Sunnyvale, CA (US); and Kwanggyoon Seo, Seoul (KR)
Assigned to ADOBE INC., San Jose, CA (US)
Filed by ADOBE INC., San Jose, CA (US)
Filed on Jul. 5, 2022, as Appl. No. 17/810,651.
Prior Publication US 2024/0013809 A1, Jan. 11, 2024
Int. Cl. G11B 27/036 (2006.01); G06T 7/30 (2017.01); G06T 11/60 (2006.01); G06V 10/776 (2022.01); G06V 20/40 (2022.01); G06V 40/16 (2022.01)
CPC G11B 27/036 (2013.01) [G06T 7/30 (2017.01); G06T 11/60 (2013.01); G06V 10/776 (2022.01); G06V 20/46 (2022.01); G06V 40/171 (2022.01); G06V 40/176 (2022.01); G06T 2207/10016 (2013.01); G06T 2207/20212 (2013.01); G06T 2207/30201 (2013.01); G06T 2210/22 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for video processing, comprising:
identifying an image that depicts an expression of a face;
encoding the image to obtain a latent code representing the image;
editing the latent code by adding an edit vector corresponding to a target attribute to obtain an edited latent code in a same latent space as the latent code, wherein the edited latent code represents the face with the target attribute that is different from an original attribute of the face and with an edited expression that is different from the expression of the face;
modifying the edited latent code using a landmark loss corresponding to a facial landmark to obtain a modified latent code in the same latent space as the latent code and the edited latent code, wherein the modified latent code represents the face with the target attribute and with a modified expression that is different from the expression and the edited expression; and
generating a modified image based on the modified latent code, wherein the modified image depicts the face with the target attribute and the modified expression.