US 12,081,627 B2
Profiling based on personal region of interest
Boosoon Jung, Seoul (KR); Sangyeob Yoon, Seoul (KR); and Mingyoung Kam, Seoul (KR)
Assigned to LG ELECTRONICS INC., Seoul (KR)
Filed by LG ELECTRONICS INC., Seoul (KR)
Filed on Nov. 24, 2020, as Appl. No. 17/103,049.
Claims priority of application No. 10-2020-0006795 (KR), filed on Jan. 17, 2020.
Prior Publication US 2021/0224542 A1, Jul. 22, 2021
Int. Cl. G06K 9/00 (2022.01); G06F 16/9535 (2019.01); G06F 16/9537 (2019.01); G06K 9/32 (2006.01); G06N 3/04 (2023.01); G06V 10/20 (2022.01); G06V 20/30 (2022.01); H04L 67/306 (2022.01); G06V 40/16 (2022.01); H04L 67/1095 (2022.01)
CPC H04L 67/306 (2013.01) [G06F 16/9535 (2019.01); G06F 16/9537 (2019.01); G06N 3/04 (2013.01); G06V 10/255 (2022.01); G06V 20/30 (2022.01); G06V 40/161 (2022.01); H04L 67/1095 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A profiling method comprising:
obtaining, by a processor, source data for generating a profile;
performing, by the processor, clustering based on the source data and map information related to the source data;
generating, by the processor, a profile of a person using a cluster generated through the clustering, wherein the profile of the person includes personal information of the person related to a geographical area;
generating, by the processor, region of interest (ROI) data indicative of an interest region of a user based on the profile of the person and collecting a plurality of images obtained from the ROI data; and
determining, by the processor, a degree of intimacy between the user and a specific person included in the plurality of images by analyzing the plurality of images and generating relationship information indicative of a relationship between the specific person and the user based on the degree of intimacy and the ROI data,
wherein the generating the ROI data indicative of an interest region of a user based on the profile of the person and collecting the plurality of images obtained from the ROI data includes:
setting, by the processor, a geographical area interested by the user based on the profile of the person; and
generating, by the processor, the ROI data of the user by performing a labeling task on the geographical area using the profile of the person,
wherein the degree of intimacy is determined by a pre-trained first neural network model, and
wherein the first model is a neural network model pre-trained by setting, as an input, at least one of (i) a distance between the specific person and the user and (ii) a number of images in which the specific person and the user simultaneously appear and setting the degree of intimacy as an output.