US 12,254,498 B2
Generation of recommendations for visual product details
Sanjeev Tagra, Redmond, WA (US); Sachin Soni, New Delhi (IN); Ryan Rozich, Austin, TX (US); Nitish Maurya, Noida (IN); Jonathan Roeder, Round Rock, TX (US); Ajay Jain, Ghaziabad (IN); and Ajay Bedi, Hamirpur (IN)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Aug. 5, 2022, as Appl. No. 17/817,699.
Application 17/817,699 is a continuation of application No. 17/186,495, filed on Feb. 26, 2021, granted, now 11,430,030.
Prior Publication US 2022/0383369 A1, Dec. 1, 2022
Int. Cl. G06Q 30/0601 (2023.01); G06F 3/0482 (2013.01); G06F 3/0484 (2022.01); G06F 3/04842 (2022.01); G06F 18/214 (2023.01); G06F 18/22 (2023.01); G06F 18/23 (2023.01); G06N 3/045 (2023.01); G06N 3/0464 (2023.01); G06N 3/08 (2023.01); G06N 3/09 (2023.01)
CPC G06Q 30/0603 (2013.01) [G06F 3/0484 (2013.01); G06F 18/214 (2023.01); G06F 18/22 (2023.01); G06F 18/23 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for generating image recommendations, the method comprising:
identifying, by a first neural network, a category associated with an object appearing in an input image, wherein the input image depicts the object as viewed from an initial perspective viewpoint, and wherein the input image is included in an image set associated with the object;
retrieving, by an image retrieval module, a plurality of sample images, each of which depicts a corresponding sample object that is associated with the identified category;
clustering, by a second neural network trained to recognize similarity of perspective viewpoint, the retrieved sample images into a plurality of groups based on perspective viewpoint, wherein a first group includes a first sample image depicting a first sample object as viewed from a first alternative perspective viewpoint that is different from the initial perspective viewpoint, and wherein a second group includes a second sample image depicting a second sample object as viewed from a second alternative perspective viewpoint that is different from the initial perspective viewpoint and the first alternative perspective viewpoint;
selecting, by an image selection module, a first representative sample image from the first group;
selecting, by the image selection module, a second representative sample image from the second group;
causing display of one or more of the first or second representative sample images in a user interface; and
causing display of a progress indicator in the user interface, wherein the progress indicator includes a completeness level of the image set relative to a target quantity of perspective viewpoints.