US 11,670,068 B2
Apparatus and method for processing image
Ye-hoon Kim, Seoul (KR); So-jung Yun, Seoul (KR); and Jun-ik Jang, Bucheon-si (KR)
Assigned to SAMSUNG ELECTRONICS CO., LTD., Suwon-si (KR)
Filed by SAMSUNG ELECTRONICS CO., LTD., Suwon-si (KR)
Filed on Jan. 25, 2021, as Appl. No. 17/156,930.
Application 17/156,930 is a continuation of application No. 15/850,690, filed on Dec. 21, 2017, granted, now 10,902,276.
Claims priority of application No. 10-2016-0176837 (KR), filed on Dec. 22, 2016; and application No. 10-2017-0135869 (KR), filed on Oct. 19, 2017.
Prior Publication US 2021/0142090 A1, May 13, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G06V 10/25 (2022.01); H04N 5/232 (2006.01); H04N 23/63 (2023.01); H04N 23/60 (2023.01); H04N 23/67 (2023.01)
CPC G06V 10/25 (2022.01) [H04N 23/63 (2023.01); H04N 23/635 (2023.01); H04N 23/64 (2023.01); H04N 23/67 (2023.01)] 15 Claims
OG exemplary drawing
 
1. An image processing apparatus comprising:
a communication interface;
a display;
a camera;
a memory configured to store computer-executable instructions; and
at least one processor, which, by executing the computer-executable instructions, is configured to:
obtain an image including at least one object via the camera,
select a data recognition model corresponding to a certain condition from among a plurality of data recognition models that are differently trained to estimate a region of interest including a salient region within the obtained image using training image data, the plurality of data recognition models comprising a first data recognition model being trained for estimating the region of interest based on information related to an object and a second data recognition model being trained for estimating the region of interest based on interest information of a user,
based on the selected data recognition model corresponding to the certain condition, estimate the region of interest within the obtained image according to criteria for the selected data recognition model to determine whether a region corresponds to learned information of interest,
the estimated region of interest including an estimate of a first region in the obtained image as the region of interest by applying the obtained image to the first data recognition model, based on the certain condition satisfying a first condition, and
the estimated region of interest including an estimate of a second region in the obtained image as the region of interest by applying the obtained image to the second data recognition model, based on the certain condition satisfying a second condition, and
control the display to display the image in which the region of interest is estimated.