US 12,450,747 B2
Performing multiple segmentation tasks
Jason Wen Yong Kuen, San Jose, CA (US); Zhe Lin, Clyde Hill, WA (US); Sukjun Hwang, Seoul (KR); Jianming Zhang, Fremont, CA (US); and Brian Lynn Price, Pleasant Grove, UT (US)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Jan. 23, 2023, as Appl. No. 18/100,419.
Prior Publication US 2024/0249413 A1, Jul. 25, 2024
Int. Cl. G06T 7/11 (2017.01)
CPC G06T 7/11 (2017.01) [G06T 2207/20081 (2013.01); G06T 2207/20104 (2013.01)] 20 Claims
OG exemplary drawing
 
16. A non-transitory computer-readable storage medium storing executable instructions, which when executed by a processing device, cause the processing device to perform operations comprising:
receiving input data describing a digital image depicting an object;
computing per-pixel embeddings for the digital image using a pixel decoder of a single machine learning model;
generating output embeddings using a transformer decoder of the single machine learning model based on the per-pixel embeddings for the digital image, input embeddings for a first segmentation task, and input embeddings for a second segmentation task; and
outputting a first digital image and a second digital image based on the output embeddings, the first digital image depicting the object segmented based on the first segmentation task and the second digital image depicting the object segmented based on the second segmentation task.