US 12,013,918 B2
Method and apparatus for clustering images
Joo Yeon Chung, Seoul (KR); Min Sik Chu, Seoul (KR); Seong Mi Park, Seoul (KR); and Kyong Hee Joo, Seoul (KR)
Assigned to SAMSUNG SDS CO., LTD., Seoul (KR)
Filed by SAMSUNG SDS CO., LTD., Seoul (KR)
Filed on Sep. 16, 2021, as Appl. No. 17/476,667.
Claims priority of application No. 10-2021-0018051 (KR), filed on Feb. 9, 2021.
Prior Publication US 2022/0253641 A1, Aug. 11, 2022
Int. Cl. G06F 18/2413 (2023.01); G06T 7/00 (2017.01); G06V 10/32 (2022.01); G06V 10/40 (2022.01)
CPC G06F 18/2413 (2023.01) [G06T 7/001 (2013.01); G06V 10/32 (2022.01); G06V 10/40 (2022.01); G06T 2207/30148 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method performed by a computing device for clustering a plurality of images, the method comprising:
converting the plurality of images into feature vectors;
first-clustering the feature vectors of the plurality of images based on a first clustering parameter; and
second-clustering feature vectors belonging to at least a part of clusters, which do not satisfy a reference score among clusters formed as a result of the first-clustering, based on a second clustering parameter different from the first clustering parameter,
wherein at least one of the first clustering parameter or the second clustering parameter is determined by optimizing a value of an objective function while a number of clusters to be formed is in a reference range, the objective function being a function related to a resolution of a clustering target image, a cluster radius, and a minimum number of vectors for cluster formation.