US 12,235,891 B2
Generating embeddings for text and image queries within a common embedding space for visual-text image searches
Zhifei Zhang, San Jose, CA (US); and Zhe Lin, Fremont, CA (US)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Jun. 28, 2022, as Appl. No. 17/809,503.
Prior Publication US 2023/0418861 A1, Dec. 28, 2023
Int. Cl. G06F 16/33 (2019.01); G06F 16/3332 (2025.01); G06F 16/532 (2019.01); G06F 16/535 (2019.01)
CPC G06F 16/535 (2019.01) [G06F 16/3334 (2019.01); G06F 16/532 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving, from a client device, a text query and an image query for conducting an image search;
detecting a user interaction with an interactive element displayed within a graphical user interface of the client device, the interactive element corresponding to combining the text query and the image query in conducting the image search;
determining, from the user interaction with the interactive element, both a first weight for the text query and a second weight for the image query;
generating a weighted combination of the text query and the image query using the first weight for the text query and the second weight for the image query determined from the user interaction with the interactive element; and
retrieving one or more digital images for provision to the client device utilizing the weighted combination of the text query and the image query.