US 11,921,777 B2
Machine learning for digital image selection across object variations
Ajay Jain, Ghaziabad (IN); Sanjeev Tagra, Panipat (IN); Sachin Soni, New Delhi (IN); Ryan Timothy Rozich, Austin, TX (US); Nikaash Puri, New Delhi (IN); and Jonathan Stephen Roeder, Round Rock, TX (US)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Apr. 26, 2022, as Appl. No. 17/729,515.
Application 17/729,515 is a continuation of application No. 16/774,681, filed on Jan. 28, 2020, granted, now 11,397,764.
Prior Publication US 2022/0253478 A1, Aug. 11, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 16/583 (2019.01); G06F 16/535 (2019.01); G06F 16/9535 (2019.01); G06F 18/214 (2023.01); G06F 18/22 (2023.01); G06F 18/24 (2023.01); G06N 3/045 (2023.01); G06N 3/047 (2023.01); G06N 3/08 (2023.01); G06N 20/00 (2019.01); G06Q 30/0251 (2023.01); G06Q 30/0601 (2023.01); G06F 16/957 (2019.01)
CPC G06F 16/583 (2019.01) [G06F 16/535 (2019.01); G06F 16/9535 (2019.01); G06F 18/214 (2023.01); G06F 18/22 (2023.01); G06F 18/24 (2023.01); G06N 3/045 (2023.01); G06N 3/047 (2023.01); G06N 3/08 (2013.01); G06N 20/00 (2019.01); G06Q 30/0254 (2013.01); G06Q 30/0255 (2013.01); G06Q 30/0261 (2013.01); G06Q 30/0269 (2013.01); G06Q 30/0621 (2013.01); G06Q 30/0641 (2013.01); G06F 16/9577 (2019.01); G06V 2201/10 (2022.01)] 20 Claims
OG exemplary drawing
 
16. A method comprising:
determining, by a processing system, whether to explore or exploit user behavior associated with a user ID in response to a request for digital content;
randomly selecting, by the processing system, a digital image from a plurality of digital images depicting variations of an object, one to another, responsive to the determining to explore;
selecting, by the processing system, a digital image from the plurality of digital images depicting the variations of the object based on a machine-learning model responsive to the determining to exploit;
generating, by the processing system, training data including a user profile associated with the user ID, outcome data describing an outcome of including the selected digital image as part of digital content, and image metadata having features extracted from the selected digital image using machine learning; and
training, by the processing system, a machine-learning model using the training data and a loss function, the training configuring the machine-learning model to generate a prediction score based on image metadata and features extracted from a respective digital image.